摘要:

THE ANALYSTTHE ANALYTICAL JOURNAL OF THE ROYAL SOCIETY OF CHEMISTRYADVISORY BOARD*Chairman: J. M. Ottaway (Glasgow, U.K.)'L. S. Bark (Salford, U.K.)E. Bishop (Exeter, U.K.)W. L. Budde (U.S.A.)D. T. Burns (Belfast, U.K.)L. R. P. Butler (South Africa)H. J. Cluley (Wembley, U.K.)E. A. M. F. Dahmen (The Netherlands)L. de Galan (The Netherlands)A. C. Docherty (Billingham, U.K.)D. Dyrssen (Sweden)G. Ghersini (ltaly)J. Hoste (Belgium)A. Hulanicki (Poland)'G. W. Kirby (Glasgow, U.K.)W. S. Lyon (U.S.A.)H. V. Malmstadt (U.S.A.)G. W. C. Milner (Harwell, U.K.)'A. C. Moffat (Aldermaston, U.K.)E. J. Newman (Poole, U.K.)H. W. Nurnberg (West Germany)'T. B. Pierce (Harwell, U.K.)E. Pungor (Hungary)P. H. Scholes (Middlesbrough, U.K.)D. Simpson (Thorpe-le-Soken, U.K.)"J.M. Skinner (Billingham, U.K.)"J. D. R. Thomas (Cardiff, U.K.)K. C. Thompson (Sheffield, U.K.)'A. M. Ure (Aberdeen, U.K.)A. Walsh, K.B. (Australia)G. Werner (German Democratic Republic)T. S. West (Aberdeen, U.K.)"P. C. Weston (London, U.K.)J. White head (Stockton -on- Tees, U. K.)J. D. Winefordner (U.S.A.)P. Zuman (U.S.A.)"G. J. Dickes (Bristol, U.K.)'Members of the Board serving on the Analytical Editorial BoardEditor: P. C. WestonSenior Assistant Editor: R. A. YoungAssistant Editors: Mrs. J. Brew, Miss D. ChevinREG I ONAL ADVISO RY EDIT0 RSDr. J. Aggett, Department of Chemistry, University of Auckland, Private Bag, Auckland, NEW ZEALAND.Professor L. Gierst, Universite Libre de Bruxelles, Facult6 des Sciences, Avenue F.- D.Roosevelt 50,Professor H. M. N. H. Irving, Department of Theoretical Chemistry, University of Cape Town, Ronde-Professor W. A. E. McBryde, Faculty of Science, University of Waterloo, Waterloo, Ontario, CANADA.Dr. 0. Osibanjo, Department of Chemistry, University of Ibadan, Ibadan, NIGERIA.Dr. G. Rossi, Chemistry Division, Spectroscopy Sector, CEC Joint Research Centre, EURATOM, lepraDr. 1. Rubeika, Geological Survey of Czechoslovakia, Malostranske 19, 118 21 Prague 1, CZECHO-Professor J. R&iGka, Chemistry Department A, Technical University of Denmark, 2800 Lyngby,Professor K. Saito, Department of Chemistry, Tohoku University, Sendai, JAPAN.Professor L. E. Smythe, Department of Chemistry, University of New South Wales, P.O. Box 1,Professor P.C. Uden, Department of Chemistry, University of Massachusetts, Amherst, MA 01 003,Editorial: Editor, The Analyst, The Royal Society of Chemistry, Burlington House,Piccadilly, London, W1 V OBN. Telephone 01 -734 9864. Telex No. 268001Advertisements: Advertisement Department, The Royal Society of Chemistry, Burlington House,Piccadilly, London, W1V OBN. Telephone 01 -734 9864. Telex No. 268001The Analyst (ISSN 0003-2654) is published monthly by The Royal Society of Chemistry, BurlingtonHouse, London WIV OBN, England. All orders accompanied with payment should be sent directly toThe Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letchworth, Herts. SG6 1 HN.England. 1983 Annual subscription rate UK f93.50, Rest of World f99.00, USA $201 .OO. Purchased withAnalytical Abstracts UK f226.50, Rest of World f238.50, USA $487.00. Purchased with AnalyticalAbstracts plus Analytical Proceedings UK f251 .OO, Rest of World f265.00, USA $539.00. Purchasedwith Analytical Proceedings UK f 1 1 7.50, Rest of World f 124.50, USA $253.00. Air freight and mailingin the USA by Publications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003.USA Postmaster: Send address changes to: The Analyst, Publications Expediting Inc., 200 MeachamAvenue, Elmont, NY 11003. Second class postage paid at Jamaica, NY 11431. All otherdespatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Post outside Europe.PRINTED IN THE UK.Volume 108 No 1285 Q The Royal Society of Chemistry 1983 April 1983B ruxelles, BELGIUM.bosch 7700, SOUTH AFRICA.Establishment, 21 020 lspra (Varese), ITALY.SLOVAKIA.DENMARK.Kensington, N.S.W. 2033, AUSTRALIA.U.S.A

ISSN:0003-2654    

DOI:10.1039/AN98308FX013

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

ANALAO 108 (1 285) 425-552 (1 983) April 1983THE ANALYSTTHE ANALYTICAL JOURNAL OF THE ROYAL SOCIETY OF CHEMISTRY42544345245746447047648148549250551 051 5CONTENTSKinetic, Static and Stirring Errors of Liquid Junction Reference Electrodes-Donald P. BrezinskiAnalytical Use of the Kinetics of Complex Formation : Simultaneous Determination of Iron andCobalt by Differential Kinetic Methods-L. Ballesteros and D. PBrez-BenditoImprovements t o the Oxygen Flask Combustion Procedure for Assay of Halogenated OrganicCompounds-Duncan Thorburn Burns and Binod K. MaitinEvaluation of Equivalence Points in the Potentiometric Titration of Mixtures of Halides-DuncanThorburn Burns, Binod K. Maitin and Gyula SvehlaFlow Injection Voltammetric Determination of Nitrate After Reduction t o Nitrite-Arnbld G.Fogg,Antoine Y. Chamsi and Mohamed A. AbdallaFully Automatic Flow Injection System for the Determination o f Uranium at Trace Levels in OreLeachates-Thomas P. Lynch, Arthur F. Taylor and John N. WilsonVolatilisation of Zirconium, Vanadium, Uranium and Chromium Using Electrothermal Carbon CupSample Vaporisation into an Inductively Coupled Plasma-Kin C. Ng and Joseph A. CarusoElectrothermal Atomisation Atomic-absorption Spectrophotometric Determination of Chrom-ium(VI) in Urine by Solvent Extraction Separation w i t h Liquid Anion Exchangers-ClaudiaMinoia, Ambrogio Mazzucotelli, Alessandro Cavalleri and Vincenzo MingantiUse of Laser Raman Spectrometry for a Quantitative Study of the Urea Synthesis under ProcessConditions.A Feasibility Study-Martin van Eck, Johannes P. J. van Dalen and Leo de GalanDistribution of Zinc Amongst Human Serum Proteins Determined by Affinity Chromatography andAtomic-absorption Spectrophotometry-John W. Foote and H. Trevor DelvesSpectrofluorimetric Determination and Thin-layer Chromatographic Identification of Selenium inFoods-Teresa Moreno-Dominguez, Concepcion Garcia-Moreno and Abel MarinB-FontChromatographic Seperation of Chlorinated Hydrocarbons Using Columns of Silica Gel o f VaryingDegrees o f Porosity and Activation-Vittorio Contardi, Renzo Capelli, Gilda Zanicchi and Marco DragoSolid - Liquid Separation after Liquid - Liquid Extraction : Spectrophotometric Determination ofCopper by Extraction of i t s 1 -Phenyl-4,4,6-trimethyl-(l H,4H)-pyrimidine-2-thiol into MoltenNaphthalene-Abdul Wasey, Raj Kurnar Bansal, Bal Krishan Puri and Masatada SatakePart 1.SHORT PAPERS521 Gas Chromatographic - Mass Spectrometric Identification of 9,lO-Epoxystearate in Human Blood-Gunnar A.Ulsaker and Gerd Teien524 Spectrophotometric Determination ofw Small Amounts of Niobium in Steels Using Sulphochloro-phenol S-ZdenBk &iek and Vlasta Studlarova528 Spectrophotometric Determination o f Cobalt after Coprecipitation o f i t s Morpholine-4-carbodi-thioate with Microcrystalline Naphtha!ene-Charnan La1 Sethi, Ashok Kumar, Bal Krishan Puri andMasatada Satake531534537540543546549552Spectrophotometric Determination o f Some Lanthanides as TetraethylenepentamineheptaaceticAcid Chelates-M.Tarek M. Zaki, Abdel F. Shoukry and Mohamed B. HafezSpectrophotometric Determination of Cerium w i t h Methylthymol Blue in the Presence o f Oxalateand Cyanide as Masking Agents-Amalia Cabrera-Martin, Roberto Izquierdo-Hornillos, Albert0 J.Quejido-Cabezas and Jose L. Peral-FernandezPotentiometric Titration Method for the Similtaneous Determination of Two Monofunctional WeakAcids of Similar Strength : Titration t o the Initial Potential-Hilda Szalai and Tames L. PaelRapid Routine Procedure for the Determination of Anhydrous and Hydrated Tripolyphosphate bymeans of X-ray Powder Diffraction-Alessandro MangiaCatalimetric Determination of Iodine in Common Salt-Kambharnpati Sriramam, Brahmandam S. R.Sarma, Agnihotram R.K. Vara Prasad and Konidena KalidasSome Comments on Calibration Procedures-Allan G. C. MorrisBOOK REVIEWSERRATASummaries of Papers in this hue-Pages iii, iv, v, vi, vii, viii, ixPrinted by Heffers Printers Ltd Cambridge EnglandEntered as Second Class at New York, USA, Post OfficNEWSTAINLESS STEELCERTIFIED REFERENCEMATERIALSavailable fromBUREAU OF ANALYSEDSAMPLES LTDFor full details of these, and for copiesof new catalogue of all CRMs suppliedby BAS, write, telephone or telex to:BAS Ltd., Newham Hall, Newby,Middlesbrough, Cleveland, TS8 9EATelephone: Middlesbrough 31 721 6Telex: 587765 BASRIDA202 for further information. See page xCOMPUTERS IN AUTOMATION ANDLABORATORY MANAGEMENTThe 5th Summer School ofAutomatic Chemical AnalysisA residential course at the University of Sussex,Falmer, Brighton, England.10th to 15th July 1983For the past four years this highly acclaimed coursehas been held at the University College of Swansea.The new venue will provide an opportunity to2xpand the scope of the course while retaining theproven format which includes a balanced mixtureof theory, practice and personal tuition.The programme will include lectures, tutorials andpractical sessions on the following broad topics:AutomationComputingData handlingManagementMicro-electronicsDetailed information and registration forms from:The 5th Summer School ofAutomatic Chemical Ana I ysis,176A North View Road,London, N8 7NB.England.A201 for further information.See page xSITUATIONS VACANTAppllcatlons are lnvlted for appolntment to the abovepost wlth effect from 1 February 1984.The University wishes to appoint a sclentlst wlth anestablished research record to take responslblllty forthe further development of research and teachlng Inthis recently created department. Candidates currentlyDccupylng either academic or non-academlc positlonsshould have a wide knowledge of modern Instrumentalmalysls and thls requlrement wlll be of primeimportance In maklng the selectlon for the post.Appointment, dependlng on quallflcatlons and:3xperlence, wlll be made on the salary scalei?23 109-R24 045 x 1 035-R30 255 per annum, todependants at UCT, generous study leave prlvlle es, aqouslng subsldy scheme subject to State regula#ons,pension fund, medlcal ald and group llfe assurance..4ppllcants should submlt a currlculum vitae statlngpresent salary, research Interests and publlcatlons, thedate duty could be assumed and the names andaddresses of three referees whom the Unlverslty mayIon should be obtalned from Mlss Jsltles Office, Chlchester House, 278ondon WClV 7HE, or from the ReglstrarThe Unlverslty's policy Is not torace or religion. FurtherInformation on thedlscrlmlnate on the grounds of sex,lmplernentatlon of thlspollcy Is obtainableFOR SALEX-RAY FLUORESCENCE SYSTEMFOR SALEPhilips PW 1410 X-ray FluorescenceSystem, 10 years old, for sale. Includes inaddition to normal specification a fivecrystal changer, Goniometer, Control Unitand Helium Flush.Apply to the Chief Executive, BritishCeramic Research Association Limited.Telephone No. 0782 45431

ISSN:0003-2654    

DOI:10.1039/AN98308BX015

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

April, 1983 SUMMARIES OF PAPERS IN THIS ISSUESummaries of Papers in this IssueKinetic, Static and Stirring Errors of Liquid JunctionReference ElectrodesResponse characteristics of reference electrodes were determined by subjectingthem to large changes in ionic strength and transference. Commercial elec-trodes often gave surprisingly poor performance, exhibiting slow, inaccurateand stirring-dependent responses. Slow response is caused primarily bydiffusional entrapment of previously measured solutions within the junction.Large offsets at pH extremes or a t low ionic strength are caused by storage ofno-flow electrodes in standard buffers, junction charge and improper junctiongeometry. Stirring potentials are caused by shifts in offset error with localchanges in concentration at the junction surface.Adequate outward flow ordiffusion of junction electrolyte serves to suppress these anomalies. Referenceproblems are often undetectable in the standard pH buffers typically used forcalibration.Keywords : Reference electrode ; liquid junction ; offset error; kinetics ; stirringpotentialDONALD P. BREZINSKICorning Glass Works, Sullivan Research Park, Corning, NY 14831, USA.Analyst, 1983, 108, 425-442.Analytical Use of the Kinetics of Complex Formation :Simultaneous Determination of Iron and Cobalt byDifferential Kinetic MethodsThe kinetic determination of iron and cobalt mixtures without a prior separa-tion is described. The methods are based on the differential reaction ratebetween pyridoxal thiosemicarbazone and these metallic ions.Variousratios a t the M level of the two ions can be determined photometricallyby using either the “logarithmic extrapolation” or the “single-point” methods.These two methods are compared.Keywords : Iron and cobalt simultaneous determination ; pyridoxal thio-semicarbazone ; kinetics of complex formation ; differential kinetic methodsL. BALLESTEROS and D. PEREZ-BENDITODepartment of Analytical Chemistry, Faculty of Sciences, University of Cbrdoba,Cbrdoba, Spain.Analyst, 1983, 108, 443-451.Improvements to the Oxygen Flask Combustion Procedure forAssay of Halogenated Organic CompoundsAn improved oxygen flask procedure is described for the accurate, preciseand rapid determination of chlorine, bromine and iodine in halogenatedorganic compounds.Quantitative recoveries were achieved by using neutralhydrazine hydrate in the absorbing solution, which may be titrated potentio-metrically directly with silver nitrate solution. Hydrazine hydrate does notinterfere under acidic conditions. Accurate and precise results were obtainedwhen the equivalence points were determined by Gran’s method.Keywords ; Halogen determination ; oxygen flask combustion ; hydrazinehydrate ; potentiometric titration ; Gran’s methodDUNCAN THORBURN BURNS and BINOD K. MAITINDepartment of Analytical Chemistry, The Queen’s University of Belfast, Belfast,BT9 5AG.Analyst, 1983, 108, 452-456.iiApril, 1983 SUMMARIES OF PAPERS I N THIS ISSUEEvaluation of Equivalence Points in the Potentiometric Titrationof Mixtures of HalidesThe accura,te and precise location of the equivalence points in the potentio-metric titration of halide ions can be achieved numerically by using Gran’smethod.The results have been shown to be better by this method thanthose obtained on the same titration data by the differential methods ofKolthoff, of Fortuin and of Hahn.Keywords ; Equivalence points ; potentiometric titration ; halide ions ; Gran’sDUNCAN THORBURN BURNS, BINOD K. MAITIN and GYULA SVEHLAmethodDepartment of Analytical Chemistry, Queen’s University of Belfast, Belfast,BT9 5AG.Analyst, 1983, 108, 457-463.Flow Injection Voltammetric Determination of NitrateAfter Reduction to NitriteNitrate can be determined conveniently by chemical reduction to nitrite, whichis then injected directly into an acidic bromide eluent in a flow injection systemand monitored using its reduction signal a t a glassy carbon electrode held a t+0.3 V versus a saturated calomel electrode.Chemical reduction was bestcarried out on a batch basis but partial success was experienced in using amethod in which a nitrate sample solution was passed continuously through acadmium sponge column and then through the injection loop of the flowinjection valve from which aliquots were injected into the flow injectionsystem. Determinations were also made by injection of nitrate sample solu-tion directly into an acidic bromide eluent and reducing the nitrate on-line withcadmium wire.Keywords : Flow injection analysis ; voltammetry ; nitrate determination ;nitriteARNOLD G.FOGG, ANTOINE Y. CHAMS1 and MOHAMED A. ABDALLAChemistry Department, Loughborough University of Technology, Loughborough,Leicestershire, LE11 3TU.Analyst, 1983, 108, 464-469.Fully Automatic Flow Injection System for the Determination ofUranium at Trace Levels in Ore LeachatesAn automatic flow injection system is described for the determination ofuranium in ore leachates. Following injection from an autosampler, theleachate is extracted with a solution of tributyl phosphate in heptane, whichremoves uranium, and the organic phase is separated. The extract is reactedwith an ethanolic solution of 2-( 5-bromo-2-pyridylazo)-S-diethylami1iophenol(BrPADAP) and benzyldimethyltetradecylammonium chloride (zephiramine)and the resulting ternary complex with U(V1) is measured spectrophoto-metrically a t 579 nm.The lower limit of determination is 0.1 p.p.m. ofuranium and up to 50 samples per hour can be analysed. In terms of speedand sensitivity this improves significantly on published procedures usingsegmented flow systems. The technique is ideal for process control and canbe applied to the analysis of ores following mineralisation.Keywords : Flow injection ; ore leaclzates ; spectrophotometry ; uraniumdeterminationTHOMAS P. LYNCH, ARTHUR F. TAYLOR and JOHN N. WILSONChemical Analysis Branch, BP Research Centre, Chertsey Road, Sunbury-on-Thames,Middlesex, TWl6 7LN.Analyst, 1983, 108, 470-475Aj5riL, 1983 SUMMARIES OF PAPERS I N THIS ISSUEVolatilisation of Zirconium, Vanadium, Uranium andChromium Using Electrothermal Carbon Cup SampleVaporisation into an Inductively Coupled PlasmaZirconium, vanadium, uranium and chromium react with ammoniumchloride ( 7 9 ; iii/I’) when heated in an electrothermal carbon cup to form theircorresponding chlorides.These nictal chlorides are subsequently vaporisedinto an inductively coupled plasma for optical emission spectroscopy. Thepreferential lialide formation of these refractory elements in the electrothermalcarbon cup has alloxved their determinations to proceed with sub-nanogramdetection limits and adequate precision of about GO/, relative standarddeviation for 5-pl samples. Linear dynamic ranges span about three ordersof magnitude.Keywords : Metal clzloride formation ; electrothermal carbon cup vaporisation ;inductively coufiled plasizza ; optical emission spectroscopyKIN C.NG and JOSEPH A. CARUSODepartment of Chemistry University of Cincinnati, Cincinnati, OH 4522 1, USA.Analyst, 1983, 108, 476-480.Electrothermal Atomisation Atomic-absorption SpectrophotometricDetermination of Chromium(V1) in Urine by SolventExtraction Separation with Liquid Anion ExchangersAn electrothermal atomic-absorption spectrophotonietric determination ofchromium (VI) in urine samples is described. Tlic separation of these ionsfrom the biological matrix by using high relative molccular mass amines(such as Ambcrlite LA-I or LA-2 liquid anion exchangers) is also reported.Keywords : lien-avalent cJironii.unz ; liquid anion exchangers ; electvothermalnlor?iisntion atoiiaic-absovptio.tz spectrometry ; wine analysisCLAUD10 MINOIACentro Ricerchc di Fisiopatologia e Sicurczza del Lavoro, Fondazione Clinica delLavoro, Universitd di Pavia, Pavia, Italy.AMBROGIO MAZZUCOTELLIIstituto di Pctrografia, UnivcrsitA di Genova, Genoa, Italy.ALESSANDRO CAVALLERICattedra di Riedicina dcl Lavoro, Universitd di Modena, Modena, Italy.and VINCENZO MINGANTIIstituto di Chimica Generale ed Inorganica, UniversitA di Genova, Genoa, Italy.Analyst, 1983, 108, 481-484.Use of Laser Raman Spectrometry for a QuantitativeStudy of the Urea Synthesis under Process ConditionsPart I.A Feasibility StudyThe feasibility of laser Raman spectrometry for the in situ determination ofthe chemical cornposi tion during the synthesis of urea under process conditionshas been studied. Ranian bands suitable for quantitative analysis were foundfor all components. The use of an internal standard appears t o be essential.The influence of temperature and pressure upon niolar intensities has beenstudied for model Components and, as a result, for most components of theurea synthcsis, no effect of pressure is expected.By contrast, the Ramanspectra of all componcnts will bc affected by a change of temperature.Keywords : Laser Rainan spectroiiietry ; in situ analysis ; w e a syntlaesisMARTIN VAN ECK, JOHANNES P. J. VAN DALEN and LEO DE GALANLaboratory for Analytical Chemistry, University of Technology, Jaffalaan 9, 2628 BXDelft, The Netherlands.Analyst, 1983, 108, 485-491.vi SUMMARIES OF PAPERS IN THIS ISSUEDistribution of Zinc Amongst Human Serum Proteins Determinedby Affinity Chromatography and Atomic-absorptionSpectrophotometryApril, 1983Affinity chromatography for albumin has been coupled with electrothermalatomic-absorption spectrophotometry to determine the distribution of zincbetween albumin and globulin ligands in normal human serum.The pro-cedure is both simple and rapid and requires only 400 pl of serum for duplicateanalyses. There is no alteration in the distribution of zinc betwccn albuminand the globulins during the separation process and the total recovery ofzinc from the column is quantitative, 98.604. Albumin-bound zinc andglobulin-bound zinc are determined with relative standard deviations of 4.5and 5.9%, respectively.The distribution of zinc obtained is in very goodagreement with that found using more complex techniques.Keywords ; Zinc determination ; serum proteins ; afinity cliromatography ;atomic-absorption spectrophotometry ; kinetic immunoturbidimetryJOHN W. FOOTE and H. TREVOR DELVESChemical Pathology and Human Metabolism, Medical Faculty of the University ofSouthampton, South Laboratory and Pathology Block, Level D, SouthamptonGeneral Hospital, Tremona Road, Southampton, SO9 4XY.Analyst, 1983, 108, 492-504.Spectrofluorimetric Determination and Thin-layer ChromatographicIdentification of Selenium in FoodsA method is described for the determination of selenium in foods.Digestionof the samples and fluorimetric determination are based on the method ofMichie et al., with minor modifications. To confirm the results from aqualitative point of view, a new thin-layer chromatographic procedure isproposed.Keywords : Selenium determination ; food analysis ; spectrofluoviinetry ; thin-TERESA MORENO-DOMfNGUEZ, CONCEPCION GARCfA-MORENOand ABEL MARINfi-FONTDepartment of Bromatology, Toxicology and Chemical Analysis, Faculty ofPharmacy, University of Salamanca, Salamanca, Spain.Analyst, 1983, 108, 505-509.layer chromatography ; 2,3-diaminonaphthaleneChromatographic Separation of Chlorinated Hydrocarbons UsingColumns of Silica Gel of Varying Degrees of Porosityand ActivationA rapid metliocl for the separation of chlorinated pesticicles such as the di-chlorodiphenyltricliloroetliane ( m T ) group and dieldrin from polychlorinatedbiphenyls by porous deactivated silica gel co1umn chromatography isdescribed. The complete separation and quantitative recoveries obtainedallow a correct tletcrniination of tliesc compounds. This procedure is licl p-ful in niulti-pesticide residue analysis.Keywords : Polychlorinated biphenyls ; chlorinated pesticides ; porous silica gelcolumn cli voulzatograph yVITTORIO CONTARDI, RENZO CAPELLI, GILDA ZANICCHI andMARC0 DRAG0Istituto di Chimica Generale, Gruppo Ricerca Oceanologica-Genova, Universiti diGenova, Genoa, Italy.Analyst, 1983, 108, 510-514

ISSN:0003-2654    

DOI:10.1039/AN98308FP041

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

April, 1983 SUMMARIES OF PAPERS IN THIS ISSUESolid - Liquid Separation after Liquid - Liquid Extraction:Spectrophotometric Determination of Copper by Extraction of its1- Phenyl-4,4,6- trimethyl- (lH,4H)-pyrimidine-2- thiol intoMolten NaphthaleneviiA selective spectrophotometric method has been developed for the determina-tion of copper after extraction of its l-pheny1-4,4,6-trimethyl-( 1H,4H)-pyrimidine-2-thiol (PTPT) complex into molten naphthalene. The optimumpH range for the extraction is 7.4-11.7. The solid naphthalene containingthe copper - PTPT complex is separated by filtration and dissolved in chloro-form. The absorbance is measured a t 400 nm against a reagent blank. Beer’slaw is obeyed in the concentration range 5.0-60.7 p g of copper in 10 ml ofchloroform solution.The molar absorptivity and sensitivity are 1.23 x lo41 mol-l cm-l and 0.008 11 p.g cm-2, respectively. The interference of variousions has been studied in detail. Conditions have been established for thedetermination of copper in goat liver, human hair and certain alloys.Keywords : Solid - liquid separation ; spectrophotometry ; copper determina-tion ; biological materials analysis ; alloy analysisABDUL WASEY, RAJ KUMAR BANSAL and BAL KRISHAN PURIDepartment of Chemistry, Indian Institute of Technology, New Delhi 110016, India.and MASATADA SATAKEFaculty of Engineering, Fukui University, Fukui 910, Japan.Analyst, 1983, 108, 515-520.Gas Chromatographic - Mass Spectrometric Identification of9,lO-Epoxystearate in Human BloodShort PaperKeywords : Gas chromatography - mass spectrometry ; 9,lO-epoxystearate ; endo-genous component ; bloodGUNNAR A.ULSAKER and GERD TEIENNational Centre for Medicinal Products Control, Sven Oftedals vei 6, Oslo 9, Norway.Analyst, 1983, 108, 521-524.Spectrophotometric Determination of Small Amounts ofNiobium in Steels Using Sulphochlorophenol SShort PaperKeywords ; Niobium determination ; steel analysis ; matrix extraction ; sulpho-chlorophenol S ; spectrophotometryZDENEK CiZEK and VLASTA STUDLAROVACentral Research Institute, Skoda Co., 316 00 Plzeii, Czechoslovakia.Analyst, 1983, 108, 524-528viii SUMMARIES OF PAPERS I N THIS ISSUESpectrophotometric Determination of Cobalt after Coprecipitationof its Morpholine-4-carbodithioate with MicrocrystallineNaphthaleneApril, 1983Short PaperKeywords : Spectrophotometry ; cobalt determination ; morpholine-Ccarbo-dithioate ; GoprecipitationCHAMAN LAL SETHI, ASHOK KUMAR and BAL KRISHAN PURIDepartment of Chemistry, Indian Institute of Technology, New Delhi 110016, India.and MASATADA SATAKEFaculty of Engineering, Fukui University, Fukui, 910, Japan.Annlyst, 1983, 108, 528-530.Spectrophotometric Determination of Some Lanthanides asTetraethylenepentamineheptaacetic Acid ChelatesShort PaperKeywords : Lanthanide determination ; spectroph.otometry ; tetraetlzylene-pentamineheptaacetic acidM.TARAK M. ZAKI, ADEL F. SHOUKRY and MOHAMED B. HAFEZChemistry Department, Faculty of Sciences, United Arab-Emirates University, P.O.Box 15551, Al-Ain, Abu-Dhabi, United Arab Emirates.Analyst, 1983, 108, 531-534.Spectrophotometric Determination of Cerium with MethylthymolBlue in the Presence of Oxalate and Cyanide as Masking AgentsShort PaperKeywords : Cerium determination ; Metlzylthymol Blue ; spectrophotometry ;Pjj'ect of masking agentsAMALIA CABRERA- MARTIN, ROBERTO IZQUIERDO - HORNILLOS,ALBERT0 J.QUEJIDO-CABEZAS and JOSg L. PERAL-FERNANDEZDepartment of Analytical Chemistry, Facultad de Ciencias Quimicas, UniversidadComplutense de Madrid, Ciudad Universitaria, Madrid-3, Spain.Analyst, 1983, 108, 534-537April, 1983 SUMMARIES OF PAPERS IN THIS ISSUEPotentiometric Titration Method for the SimultaneousDetermination of Two Monofunctional Weak Acids of SimilarStrength: Titration to the Initial PotentialShort PaperKeywords : A cid-mixture determination ; potentiometric titration ; titration toinitial potentialHILDA SZALAI and TAMAS L.PAALNational Institute of Pharmacy, POB 450, Budapest 5, H-1372 Hungary.Analyst, 1583, 108, 537-540.Rapid Routine Procedure for the Determination of Anhydrous andHydrated Tripolyphosphate by means of X-ray Powder DiffractionShort PaperKeywords : Tripolyphosphates determination ; X-ray powder difjcactionALESSANDRO MANGIAIstituto di Chimica Generale ed Inorganica, Universiti di Parma, Via M. D’Azeglio 85,43100 Parma, Italy.Analyst, 1983, 108, 540-543.Catalimetric Determination of Iodine in Common SaltShort PaperKeywords : Iodine determi.nation ; common salt analysis ; catalysis ; hexa-chloroantimonate( V ) hydrolysisKAMBHAMPATI SRIRAMAM, BRAHMANDAM S . R.SARMA, AGNI-HOTRAM R. K. VARA PRASAD and KONIDENA KALIDASDepartment of Chemistry, Nagarjuna University, Nagarjunanagar 522 510 (A.P.),India.Analyst, 1983, 108, 543-546.Some Comments on Calibration Proced.uresShort PaperKeywords : Calibration pvocedures ; X-ray fluorescence spectrometryALLAN G. C. MORRISESAB Ltd., Beechings Way, Gillingham, Kent, ME8 6PU.ixAnalyst, 1983, 108, 546-548TUCK IN UNDER FLAP ATHE ANALYST April, 1983 iREADER ENQUIRY SERVICEFor further information about any of the products featured in the advertise- fments in this issue, please write the appropriate A number in one of the 7 iPostage paid if posted in the British Isles but overseas readers must affix 6 ia stamp.boxes below. :;I H I I(Please use BLOCK CAPITALS)NAM E ..................................................... .........................................................................................................OCCUPATION ................................................................................................................................................................. iADDRESS ............................................................................................................................................................................. iSECOND FOLDPostagewill bePaid byLicenseeDo not affix Postage Stamps if posted inGt. Britain, Channel Islands or N. IrelandIBUSINESS REPLY SERVICELicence No. W.D. 106Reader Enquiry ServiceThe AnalystThe Royal Society of Chemistryi 0 : : c : : + IBurlington HousePiccadilly London W1 E 6WFENGLANDTHIRD FOLD2 :4 ;ll:g :$ ! I O I.

ISSN:0003-2654    

DOI:10.1039/AN98308BP045

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

APRIL 1983 The Analyst Vol. 108 No. 1285 Kinetic Static and Stirring Errors of Liquid Junction Reference Electrodes* Donald P. Brezinskit Corning Glass Works Sullivan Research Park Corning N Y 14831 USA Response characteristics of reference electrodes were determined by subjecting them to large changes in ionic strength and transference. Commercial elec-trodes often gave surprisingly poor performance exhibiting slow inaccurate and stirring-dependent responses. Slow response is caused primarily by diffusional entrapment of previously measured solutions within the junction. Large offsets at pH extremes or a t low ionic strength are caused by storage of no-flow electrodes in standard buffers junction charge and improper junction geometry. Stirring potentials are caused by shifts in offset error with local changes in concentration a t the junction surface.Adequate outward flow or diffusion of junction electrolyte serves to suppress these anomalies. Reference problems are often undetectable in the standard pH buffers typically used for calibration. Keywords Reference electrode ; liquid junction ; offset error; kinetics ; stirring potential Variation of the junction potential of the reference electrode is a recognised source of error in ion-selective electrode potentiometry. The junction potential is generally attributed to ionic interdiffusion at a direct (liquid - liquid) interface between the measured sample and the junc-tion electrolyte. Guggenheim1p2 showed that such diffusion potentials can be rendered minimal and reproducible by using a concentrated equitransferent junction electrolyte (typically 4 M potassium chloride solution) at a free diffusion interface with cylindrical sym-metry.The significance given to the residual junction potential under these conditions is different in the American (NBS) and British (BSI) pH scales. The multi-standard NBS sys-tem regards the residual junction potential as an error and the different primary standards are restricted to a range over which junction potential variation is rather negligible (<1 mV).S The single-standard BSI system incorporates the variation in junction potential into the defined pH v a l ~ e . ~ ~ ~ Whatever the theoretical interpretation of the junction potential the significance of practical measurements is based on the presumption that non-negligible junction potentials will be reproducible and equal to those yielded by a standard (definitive) junction.However reference errors (departures from standard junction potentials) seem likely as one proceeds from the optimised solutions and methodologies defining the pH scale into the much broader domain of practical pH measurement. The usual calibration standards being formulated with the intent of minimising junction error are generally moderate with respect to pH buffer value ionic strength and transference; however it is common practice to measure solutions for which these parameters are extreme (e.g. de-ionised water acid plating baths, colloidal soils) Also commercial reference electrodes employ a wide variety of junction materials geometries and electrolytes and as a group bear little resemblance to each other or the definitive junction.Small but significant errors (<3 mV) have been observed with commercial reference electrodes in standard pH buff ers.5 Much larger deviations might be expected in non-standard solutions ; accordingly Illingworth6 recently reported concentration-dependent errors averaging 0.2 pH per decade in a sampling of porous ceramic junction refer-ence electrodes. Although there have been numerous studies pertaining to the magnitude and * Presented a t the 32nd Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Atlantic City NJ March 9th 1981. t Address for correspondence 54798 CR653 Paw Paw MI 49079 USA. 42 426 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol.108 stability of liquid junction potentials @er se,7-lo the kinetics of practical reference electrodes have received scant attention. In particular their contribution to measurement drift and stirring dependence has not been elucidated. Finally reference electrodes with a permanent, gelled electrolyte have become popular primarily because they eliminate electrolyte flow and the consequent need for re-filling. Although this innovation is convenient and removes an unnecessary constraint on junction design namely control of flow its impact on performance remains uncertain. The purpose of the present study was to determine the general response characteristics of practical reference electrodes and how they might be improved. A rigorous test procol was established and applied to a representative selection of commercial reference electrodes.The level of performance obtained was surprisingly poor. Virtually all of the reference electrodes exhibited large transient errors (slow response) under certain conditions. More importantly, some electrodes exhibited large static (after-equilibrium) errors that would invalidate many measurements of practical importance. Finally some electrodes showed substantial shifts in potential when agitated in low ionic strength solutions. This “stirring error” made the correct pH value uncertain and made it difficult to titrate solutions to a desired pH. These various errors are explained and some prototype reference junctions with fast ideal response are presented in support of the theory.Although discussed in the context of pH measurement, these findings also apply to other electroanalytical uses of reference electrodes. Theoretical Definition of Correct Response The diffusion potential being a non-equilibrium phenomenon generally depends on inter-facial configuration. Guggenheiml distinguished three types of definite junction continuous mixture constrained diffusion and free diffusion. These junctions agreed to within 1 mV in the system he studied (0.1-3.5 M potassium chloride solution - 0.1 M hydrochloric acid) but a stable reproducible potential was not obtained unless the interfaces had cylindrical symmetry (all chemical and electrical gradients parallel to a straight line). The diffusion potential for a continuous mixture interface can be computed from the Hender-son equation,ll which can be expressed as where xi is the signed valence of the ith ionic species pi is its relative mobility (velocity/force) and C and C:’ are its molar concentrations in solutions I and 11.Table I gives Henderson TABLE I CALCULATED LIQUID JUNCTION POTENTIALS ( Uinside - Uoutside/mV) FOR INTERFACES BETWEEN DISSIMILAR ELECTROLYTES* Inside 4M l M 1M 10-4 M 4111 1M Outside KC1 KC1 HC1 HC1 equitransferentt equitransferentt 4 M KCl . . 0 -0.7 - 14 -4.6 -0.5 -0.9 1 M KCl . . 0.7 0 - 27 - 4.0 - 0.2 - 0.5 1 M HCl . . 14 27 0 152 14 27 M HC1 . . 4.6 4.0 - 152 0 0.0 0.0 l o 4 M KCl . . 6.1 4.5 - 169 - 27 0.0 0.0 M NaOH 4.9 4.2 - 160 - 33 0.0 0.0 1 M NaCl . - 1.3 - 4.4 - 31 - 43 - 2.2 - 4.9 1 ~ N a 0 H - 8.8 - 19 -33 -132 -9.6 - 20 10-7 M H+ OH- .. 7.8 7.2 -262 -110 0.0 0.0 * From equation (l) using Table I1 mobility values. t With ionic conductivity comparable to KCl p+ = p- = 75. 58 mV = 1 pH unit at 20 “C. No corrections have been made for activity or incomplete dissociation April 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 427 values for interfaces between certain key electrolytes. A study of Table I and equation (1) reveals that a junction potential arises when the net charge diffusion tendency ZzipiCi, changes across the interface. The junction potential between a concentrated and a very dilute electrolyte is determined primarily by the transference characteristics of the concentrated electrolyte. The potential is large if the concentrated electrolyte is very heterotransferent (e.g.hydrochloric acid) but small if it is fairly equitransferent (e.g. potassium chloride solu-tion). In the latter instance the magnitude of the residual junction potential depends logarithmically on the conductance (Cz:piCi) of the dilute electrolyte but not on its trans-ference characteristics. This gives a simple explanation for Picknett's observation12 that the residual junction potentials for different dilute electrolytes are nearly identical at similar levels of specific conductance. Thus the relative merits of different junction electrolytes are largely indicated by their concentration (self-diff usion) potentials [Table 11 (B)]. Although 4 M potassium chloride solution approaches the ideal of a concentrated equitransferent electrolyte and yields small junction potentials against standard buffers appreciable junction potentials may occur between it and concentrated heterotransferent samples (Table I) ; hence such samples cannot be meaningfully measured unless the junction potential is standardised and reproducible.The steady-state response of a capillary junction electrode with 4 M potassium chloride solution is used herein as the defining standard. The ideal practical reference would be one that quickly and reproducibly establishes the same potential relative to the capillary junction electrode in all solutions. A fixed offset is not regarded as error as it can be attributed to a difference in half-cell potential between the tested and stand-ard electrode and would normally be nulled by calibration of the ion meter.i i Avariable offset is indicative of electrode error. TABLE I1 DIFFUSIONAL CHARACTERISTICS OF IONS AND SALTS (A) Limiting ionic conductances13/ (B) Self-diffusion (concentration) potentials/mV* mho cm2 equiv-l (25 "C) H+ . . . . 349.8 K+ . . . . 73.5 NH,+ . . . . 73.5 Na+ . . . . 50.1 Li+ . . . . 38.7 NO,- . . 71.4 c1- . . . . 76.3 SO,2- . . . . 80.0 OH- . . 198.6 KC1 . . 1.1 KNO - 0.9 NH,NO - 0.9 HC1 . . . . -38.0 NaOH . . 35.0 KOH . . 27.2 Li,SO . . 1.4 LiCl . . . . 19.4 NaCl . . . . 12.3 * Table 11 (B) values are shifts in potential in passing from very dilute solution to one which is 10-fold more concentrated. Calculated from equation ( l ) using relative mobilities (p) determined by dividing Table 11 (A) conductances by unsigned valences ( I z I ) .Analysis of Errors Caused by Junction Charge Let (T represent the concentration of fixed ionic charge which for simplicity is assumed to be uniformly distributed within the junction void volume. This charge must be balanced by mobile counter ions of opposite sign. Thus C- = C+ + CT where C- and C+ are the concentra-tions of negative and positive mobile ions. The flux f of a given ionic species is the sum of diffusional electrophoretic and bulk-flow contributions so f = -DdC/dx + CpexE - Cv, where p is the ionic mobility D = pkT is the diffusion coefficient e is the unit charge z is the ionic valence E = -dU/dx is the electric field x is the distance into the junction and v is the flow velocity out of the junction.Local electroneutrality requires the anion and cation fluxes to be essentially equal. Assuming a 1 1 equitransferent electrolyte setting ff = f- and solving for E one obtains E = - F . z ( G + D . ~ ) RT 1 da v , 428 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vole 108 where C = (C+ + C-)/2 is the mean ionic concentration. Substituting equation (2) for E in the flux expressionf = (f+ + f-)/2 one obtains Finally a steady-state profile (aC/at = 0) implies that the flux is everywhere the same (af/ax = 0) but deep within the junction where C = 4 M>U the flux is due solely to outward flow transport s o f = -CJv where C is the concentration of the undiluted junction electrolyte. If da/dx = 0 and v = 0 equation (2) yields E = 0; thus the potential is constant along a uniformly charged unflowing junction in spite of concentration gradients and diffusional transport of an equitransferent electrolyte.Solving equation (3) under these conditions yields a linear concentration gradient. At the surface of the junction however there is an abrupt change in the ratio of free anions and cations so a Donnan-type potential is required to balance charge transport across the interface. Let Co and C denote the mean ionic concentrations just inside and outside the junction surface. Equating the diffusional and electrophoretic fluxes for each species which become large and must cancel throughout the abrupt interface solving for -E and integrat-ing one obtains the boundary potential AU = U, - U = (RT/F)ln(C;/C,) = (RT/F)ln(C,/Ct).Solving the simultaneous equations C$/C = C,/C; and C$ + a = C;, one obtains C$ = 4 C i + (0/2)~ - 4 2 and C; = 2/Ci + (012)~ + 4 2 . Thus the change in potential upon entering the junction is w(RT/F)(o/2Cl) for C > la1 ; -(u/ I aI)(RT/F)ln( 101 /Cl) for C >>. U. This shift in potential has the same polarity as the space charge. Equation (4) is a single-boundary equivalent of the Teorell - Meyer - Sievers membrane potential.14s15 If the concentration C at the junction outer surface is comparable to or smaller than a the shift in potential is large At the inner surface where C = 4 M > O there is no shift in potential. With flow through the junction streaming potentials are predicted. Equation (2) indicates that streaming fields will be inversely proportional to electrolyte concentration.Although such fields are likely to be negligible in the 4 M potassium chloride filled-portion of the junction, they might become appreciable near the end of the junction where C is reduced by diffusional exchange with the solution. Exact analysis of this situation would require solving non-linear equation (3) followed by numerical integration of E to yield the flow potential. However, dimensional analysis indicates that the steady-state flow potential due to diffusional exchange is independent of the flow velocity if junction charge is uniform. Equation (4) with f = -CJv and da2/dx = 0 is invariant under the parameter change v' = pv where j5 is the factor by which flow velocity is changed so the effect of increasing the outward flow velocity is to sharpen the concentration and field profiles and increase the field strength all by the same factor.Thus the potential at corresponding positions is unchanged because the higher field is exactly compensated by shorter distance. If the flow velocity is high enough to confine diffusional exchange to the outer portion of the junction the steady-state end-streaming potential is essentially independent of flow velocity. For example if the flow is increased the end-streaming potential should increase abruptly then decay back to its original level as a shorter steady-state profile is established. A rough approximation for the flow potential in a uniformly charged junction can be derived by assuming that fixed charge does not significantly alter the steady-state exponential con-centration profile obtained in a flowing uncharged junction (Fig.1). In this instance, C = CJ + (C1-CJ)e-Dz'D and integration of equation (2) yields This unbalanced shift in potential would yield a static error. I U(X) = - JEdx = (RT/F)(xflU/2DCj + (o/2C,)h([CJ + (C1 - Cj)e-Dz'D]/Cl>) 0 At large x the second logarithmic term in the brackets approaches a constant value while the first term corresponding to the streaming potential of the 4 M potassium chloride-fille April 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 429 portion continues to increase linearly with distance. The first term can simply be ignored if the streaming potential is negligible when C (k 4 M potassium chloride solution) fills the whole junction.Hence the end-streaming potential should change logarithmically with concentration of the exterior electrolyte with a Nernstian slope that is reduced by the factor 0/2CJ. As a/2CJ will typically be several orders of magni-tude below unity the contribution of the external solution to the streaming potential should be negligible even when C is comparable to (T and gives a significant static error. Then AU = ((T/2CJ) (RT/F)ln(C,/C,). 100% C0i”t I I I I Inside I (4 M KCI) I I I - Y (Flow) I solution I 0% coefi \ Distance from external surface Fig. 1. Exponential concentration profile in junction having outward flow. Total transport = flow + diffusion. dC/dt = vaC/ax + Da2C/ax2 = 0 a t steady state. Hence Cext(x) = Coexte-vz‘D for outside species and Ci*t(x) = Coint (1 -e-vpr’D) for inside species.Experimental Kinetic Test Reference electrodes were transferred between beakers containing about 150 ml of 1 M potassium chloride solution 1 M hydrochloric acid and 10-4 M hydrochloric acid. Potentials were measured versus sealed calomel electrodes left permanently in solution. These “station-ary” references provided stable potentials against which transient responses could be measured. The junctions of the stationary references were covered with perforated silicone-rubber caps plugged with glass-wool so that their potentials would not be affected by stirring. This allowed the tested electrode to be agitated selectively to check for stirring potentials. The transfer procedure was started by first equilibrating the electrode for at least 5 min in the 1 M potassium chloride solution.The electrode was then quickly rinsed in several aliquots of distilled water transferred to the next test solution and vigorously agitated for 5 s. After 4 mins the electrode was again vigorously agitated to indicate the shift due to stirring. The potential was recorded for an additional 1 min without stirring. Finally the electrode was rinsed and transferred into the next solution. Data were taken for all possible transfers between solutions. Particular care was taken in rinsing the electrode prior to its transfer into M hydrochloric acid which was extremely susceptible to carry-over from the 1 M solutions. Static Test Various commercial electrodes were initially evaluated by transferring them together for 15 min in 1 M potassium chloride solution 15 min in 1 M hydrochloric acid and then 30 min in 10-3~ hydrochloric acid.Differences in potential relative to a 1 mm bore glass capillary junction reference were noted at the end of each immersion period when fairly steady values had been attained. The capillary electrode (Fig. 2) was employed by flushing the capillary junction with fresh 4 M potassium chloride solution from the “purge” syringe rinsing the electrode with water and blotting dry immersing it into the solution and slowly pulling the “draw” syringe to move the solution interface into the capillary bore. Static errors were later determined from the dynamic test by noting the differences between the final potentials attained in each solution.However the stationary references slowl 430 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol. 108 drifted relative to each other over several days of immersion. To compensate for this drift, the solution potentials were measured periodically by use of the capillary junction reference, and the observed shifts in potential were subtracted from the experimental data. Finally it was found that a porous ceramic junction used with sufficient outward flux of pure 4 M potassium chloride solution could accurately replace the capillary junction. Unless otherwise specified experiments with ceramic junctions employed a silica-based sintered ceramic of 15% porosity 1 mm diameter and about 5 mm length clad in glass and ground flat at the exterior surface.Cell potentials were measured at 20 "C using a Corning Model 130 pH meter connected to a strip-chart recorder. Corning pH standard solutions were used 0.05 M potassium biphthalate pH 4.00; 0.05 M potassium dihydrogen orthophosphate - sodium hydroxide pH 7.00; and 0.05 M potassium carbonate - potassium tetraborate - potassium hydroxide pH 10.00, " Pu rg e " syringe // Capped stationary reference requiring calibration I Solution I Fig. 2. Capillary junction reference electrode used for deter-mining offsets between stationary references. A calomel internal electrode was used in some experiments. Sulphate-free (uncharged) agarose HSIF grade was obtained from Litex Corp. Celgard 3501 microporous polypropylene film was obtained from Celanese Corporation Greer SC USA, and Goretex fabric from W.L. Gore & Associates Inc. Elkton MD USA. Results and Discussion Because the liquid-junction potential should be determined primarily by the sample's ionic conductance and transference in theory only four types of solutions are needed to test ade-quately the performance of a reference electrode. These are concentrated equitransferent (e.g. 1 M potassium chloride solution) ; concentrated heterotransferent (e.g. 1 M hydrochloric acid) ; dilute equitransferent (e.g. M potassium chloride solution) ; and dilute hetero-transferent (e.g. lo4 M hydrochloric acid). Only one of the dilute solutions need be used as its composition should have little bearing on kinetic effects during transfers to and from the strong solutions. Dilute hydrochloric acid was selected as preferable to dilute potassium chloride solution for detecting steady-state errors with potassium chloride-filled electrodes ; a ca.4 M potassium chloride - 10-4 M potassium chloride interface being a single-species gradient would be insensitive to improper geometry. Fig. 3 shows a typical response to the kinetic test. A very large and persistent transient error occurred upon transfer from 1 M to The initial error was roughly -2 pH units and equilibrium was not reached even after 5 min. The other transfers M hydrochloric acid A@&? 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 431 resulted in smaller but significant transients. In the last transfer from 1 M potassium chloride solution to lo-* M hydrochloric acid the electrode “remembered” its previous exposure to 1 M hydrochloric acid and gave a fairly stable reading that was in error by about -0.6 pH unit.This type of response was characteristic of most commercial reference electrodes particularly those without junction flow (gel electrodes). For the sake of simple comparison the difference between millivolt readings at 20 s and at 5 min was taken as an Table I11 gives kinetic data for a variety of commercial electrodes. 1 M 110-41 1 M KCI 1 M HCI 10-4M HCI 1 M HCI I 1 M KCI I 10-4M HCI v j o 0 al c .- 4- - -50 > E \ - (II .- c. f -100 4-II al U 5 -150 al u1 - 1 -2 1 I 1 1 0 5 10 15 20 25 30 35 40 Time/min Typical response of no-flow reference electrode with 4 M potassium chloride junction electrolyte (5 mm long ceramic junction silver chloride saturated 4 M potassium chloride electrolyte gelled with 2% agarose stored in 4 M potassium chloride solution).Fig. 3. The arrows indicate times of brief agitation. index of “transient error,” and this difference was converted to equivalent pH measurement errors by dividing by +58 mV pH-1. Transient errors going from hydrochloric acid to other solutions were negative in polarity and were moderately large upon transfer to 1 M potassium TABLE I11 REFERENCE ELECTRODE KINETIC RESPONSE (pH, - pH 300 s) ‘l‘he tabulated values represent pH measurement errors at 20 s. combination electrodes. All references except A D2 and G were in Sealed no-flow electrodes are denoted Half-cells were Ag - AgCl except as noted. “gel” ; the others were re-fillable flow-type.A1 A2 A3 B C D1 D2 E F G H Electrode . . . . . . . . From 1 M KCI From 1 M HCI Type Ceramic Asbestos fibre calomel (blown clear) Ceramic gel Ceramic gel (at ground input) Polymer cloth gel Ceramic annulus Sintered PTFE body gel Annular glass channel Ceramic : No agitation 5 s vigorous agitation 20 s vigorous agitation Plastic sleeve (D. J.) : 10% KNOs filled 4 M KCI filled 0.01 M KCI filled 0.01 M KC1 filled, agitated Pt fibre thallium amalgam ? O I M I M -10-4; KCI HCI HC1 0.00 0.11 0.00 0.17 0.00 0.01 0.00 0.00 0.04 -0.06 0.00 0.13 -0.29 0.00 -0.10 0.33 0.00 0.11 0.01 0.01 0.12 -0.24 0.00 0.00 -0.01 0.00 0.22 0.07 0.00 0.24 0.27 0.04 0.00 0.04 0.14 0.00 0.07 0.00 0.00 -1.2 2.4 0.00 -1.3 2.4 0.00 0.00 0.01 To 1 M KCI - 0.11 - 0.16 - 0.01 - 0.07 - 0.10 0.19 - 0.11 -0.24 0.34 -0.46 - 0.14 - 0.08 0.89 0.83 - 0.06 I M HC1 0.03 0.00 0.05 0.02 - 0.03 0.02 0.02 0.02 0.04 - 0.01 -0.06 0.00 - 0.19 - 0.05 0.05 -10-4 M HCI -0.81 - 0.4 -1.3 -1.6 1.6 -0.74 - 2.0 -0.77 - 1.8 - 2.0 - 2.0 -1.9 -1.6 - 1.5 - 4.4 - 3.8 - 1.9 From lo-‘ M HCl r - p T o ~ M I M -10-4~ KCI HCI HC1 -0.01 0.14 0.10 0.00 0.01 0.04 -0.06 - - 0.01 0.03 0.00 - 0.04 - 0.08 0.01 0.00 0.28 -0.03 0.31 0.04 -0.24 0.08 0.15 -0.03 0.45 0.03 0.15 0.00 0.38 0.09 0.60 0.57 0.00 -0.05 0.13 0.02 -0.00 0.10 0.00 -0.09 0.09 -0.40 -0.34 0.21 -0.33 0.00 0.06 0.0 432 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol.108 chloride solution and extremely large upon transfer to M hydrochloric acid. Transfers to 1 M hydrochloric acid generally had positive transient errors. The larger transient error was associated with prior soaking in the weaker electrolyte. Rinsing with water and re-immersion in the same solutions gave fairly small transients. Data for electrode F showed that vigorous agitation had little effect on the transient. Thus the transient is not due to carry-over on the surface of the electrode nor does it depend much on mass transport within the external solution. The processes associated with the transient must take place largely within the electrode itself.Table I11 data for the double-junction electrode (G) indicate little difference in kinetic response between 4 M potassium chloride and 10% potassium nitrate filling solutions. However the use of 0.01 M potassium chloride solution as the junction electrolyte resulted in a very large increase in several of the transients; concentrated filling solutions gave a much faster response. Again stirring had negligible effect on response time. The polarity of transients almost always corresponded to lags rather than overshoots in pH response. The few exceptions included transfers of electrodes B and D1 from 1 M potassium chloride solution to lov4 M hydrochloric acid. These would give overshoot if the 1 M potassium chloride solution had a pH above 4. Static error is the error in potential that persists after the electrode has essentially stopped drifting.Columns 1 4 in Table IV show millivolt data for the static test (simultaneous Electrode A1 A3 . . B c D1 ,. D2 E F TABLE IV REFERENCE ELECTRODE STATIC RESPONSE Electrode potential/mV* 1 M KC1 (15 min) - 44.9 - 28.1 -37.5 - 53.6 - 39.0 -44.5 - 46.0 -49.9 1 M HC1 (15 min) -42.5 - 13.3 - 24.6 - 44.4 - 23.5 - 20.5 -41.2 - 50.2 - M HCl (15 min) -49.2 - 25.0 -7.0 - 75.3 - 14.7 -63.1 -41.9 - 54.9 - M HC1 (30 min)? -45.5 - 22.5 - 13.6 -65.1 - 17.9 - 55.5 -42.2 - 55.2 pH measurement discrepancy r m From 1 M KC1 1 M KCl 1 M HC1 To 1 M HC1 10-3 M HClt M HClt 0.04 -0.01 - 0.05 0.25 0.10 -0.16 0.22 0.41 0.19 0.16 -0.20 - 0.36 0.27 0.36 0.10 0.41 -0.19 -0.60 0.08 0.07 - 0.02 - 0.01 -0.09 - 0.09 * mV relative to calomel capillary reference.t 30-min value; re-agitated vigorously a t 15 min. measurement with all electrodes together in the same beaker). In columns 5-7 the shifts in potential relative to the capillary reference upon changing solutions have been converted to equivalent pH units. Suppose parallel measurements of pH are made using identical pH electrodes but one measurement uses the capillary junction reference and the other uses the indicated reference. After meter calibration to achieve agreement in the solution marked “From,” the electrodes are transferred to the solution marked “To.” The tabulated values are the measurement discrepancies that would be observed in the second solution.These often amount to several tenths of a pH unit. As calibration buffers more closely resemble potassium chloride in transference and as electrode non-idealities are probably minimised in strong potassium chloride electrolytes the discrep-ancies measured relative to potassium chloride (listed in the first two columns) probably reflect the errors that would be experienced in actual use. It was noted that the shifts in potential in column 2 were opposite for two different types of references made by the same manufacturer (D1 and D2) so the results obtained with these electrodes in 0.001 M hydrochloric acid should disagree by over 0.5 pH unit. To confirm a discrepancy pH measurements were performed to compare results obtained with the two references using the same pH electrode.In 1 M potassium chloride containing a trace amount of hydrochloric acid the measurements agreed closely as would be expected if junction anomalies are suppressed by the 1 M potassium chloride solution. In 1 M hydrochloric acid the readings differed by 0.2 pH unit and in M hydrochloric acid the stable readings differed by a full pH unit. The significance of these shifts is the following. These data are shown in Table V April 1983 C I Q 7.00 -0 + 6.99 -6-98 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES TABLE V Reference and pH electrode both exposed Reference 1 in baffle, pH sensor 15-cm reference exposed electrolyte head 1 -cm electrolyte head 1 - ---- - c - I I I I Stir Stir Stir Stir w w Vigorous stirring I I 0 5 10 15 20 433 DEPENDENCE OF MEASURED pH ON REFERENCE ELECTRODE Reference Solution electrode* mV PH 1 M KC1 trace of HC1 D1 210.8 3.49 1 M KC1 trace of HC1 D2 214.2 3.43 1 M HC1 D1 395.4 0.41 1 M HC1 D2 384.2 0.60 M HC1 D2 225.2 3.24 10-4 M HC1 D1 162.8 4.28 * pH electrode pH section of combination electrode D1.Refer-Reference D2 AgCl ence D1 AgCl ceramic-annulus junction. porous polymer body non-refillable. “Stirring potentials” are generated by the reference electrode and normally are a problem only at low ionic strength. In Fig. 4 for example the pH indicated by a glass pH electrode and a ceramic-junction reference electrode in pH 7 standard buffer (0.05 M phosphate) was shifted only slightly (ca.0.013 pH unit) by stirring. This shift originated at the reference electrode as it was completely eliminated by placing the reference electrode within a protective baffle (a plastic syringe body immersed in the solution). Fig. 5(a) shows the much larger stirring effect that occurred when the same electrodes were used to adjust de-ionised water to ca. pH 10 with sodium hydroxide solution. When the solution was stirred as required to distribute titrant the measurement shifted by about -0.2 pH unit. When stirring was stopped the indicated pH slowly drifted back to its unstirred value. This stirring effect made it difficult to adjust the solution to the desired final pH reading and moreover it was unclear which pH value was correct stirred or unstirred.Again shifting the reference to the unstirred baffle restored the quiescent pH value whereas placing the pH electrode within the baffle had no effect. The slowness of recovery when stirring was stopped was due primarily to continued movement of the solution. Recovery from longitudinal agitation of the electrode which did not disturb the bulk solution was much faster [Fig. 5 ( b ) ] and suggested that the intrinsic decay time for the stirring potential was in the order of a few seconds BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol. 108 5 U ; 9.8 .- 0 - 9.7 9.6 434 10.0 I TI Q 9.9 c, . 9.8 C -9.7 Agitate Agitate - -- I 1 I I --a)pH and reference Reference in electrodes both baffle; pH Both exposed I exposed I exposed -,Hard Ti me/m in Stir A 1 15 I 0 I I I 0 20 40 60 Time/s Fig.5. Large stirring potentials observed at low ionic strength. Same electrodes as in Fig. 4. (a) Solution stirred by magnetic bar; (b) longitudinal agitation of reference electrode. Explanation of the Various Errors The observed transient errors can be explained largely by diffusional entrapment of the previously measured solution within the physical junction. During electrode immersion, 4 M potassium chloride within the outer region of the junction is diffusionally exchanged for external solution species. When the reference is rinsed and transferred to the next solution, this exchanged layer intervenes between the 4 M potassium chloride junction electrolyte and the new solution. This layer disturbs the junction potential because diffusion potentials are not a transitive property.That is given solutions A B and C the diffusion potential at the interface A/C is not the sum of the potentials at the interfaces A/B and B/C. This is seen clearly from Table VI which compares the diffusion potentials for “sandwiched” and “direct” interfaces. The calculated differences in diffusion potential are in general agreement with the magnitudes of the observed transients. to 1 M hydrochloric acid where the observed transient is much smaller than predicted by the model. This might be expected as the residual concentration of hydrochloric acid and potassium chloride at the junction surface at time of transfer is undoubtedly much higher than 10-4 M, and would substantially reduce the diffusion potential against 1 M hydrochloric acid.The sole exception is the transfer from TABLE VI CALCULATED AND OBSERVED TRANSIENTS FOR “SANDWICH” INTERFACES Type of transfer A I 7 Type of interface 1 M KCI -+ 1 M HCl/mV 1 M HCI -D 10-4 M HCl/mV M HCI -+ 1 M HCl/mV 1 M HCI -P 1 M KCl/mV (A) Sandwich Inner ~ M K C I / ~ M K C I -1 ~ M K C I / ~ M H C I 14 ~MKCI/~O-~MHCI 5 ~ M K C I / ~ M H C I 14 Outer . . . . 1 M KCI/1 M HCl 27 1 M HCl/lO-4 M HCl - 152 M HC1/1 M HC1152 1 M HC1/1 M KCI - 27 Sum 28 - 138 157 -13 (B) Direct . . . . 4 M KCI/1 M HCI 14 4 M KCl/10-4 MHCI 5 4 M KCI/1 M HCI 14 4 M KC1/1 M HCI I (C) Difference (transient expected fromA -B) . . (Fig. 2) . . . . (D) Observed transient 14 -- 20 143 143 -130 30 -14 18 The potential at interface A/B is generally negligible compared with that at B/C if A is 4 M potassium chloride solution so in effect the layer of previously measured solution temporarily replaces 4 M potassium chloride solution as the junction electrolyte.Error kinetics are deter-mined by the time required to disperse this layer and obtain a direct interface. Under no-flow conditions the kinetics are determined by the duration of prior immersion with the thicknes April 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 435 of the junction being the limiting factor at the steady state. A faster response can be obtained by using a thinner junction. The diffusional relaxation time for a junction of thickness X is approximately T = X2/?r2D where D is the diffusion coefficient.Taking D = 2 x cm2 s-l for a typical electrolyte the relaxation time for a 0.5-cm junction would be roughly 30 min. This could be reduced to 20 s by reducing the effective junction thickness to 0.06 cm. This presumes that the junction permeability is low enough to keep the potassium chloride - sample interdiffusion profile from extending substantially beyond the physical junction. Thin-membrane junctions were made from Goretex fabric (microporous PTFE on a nylon backing, PTFE side out) and Celgard 3501 (microporous polypropylene 25 pm x 0.25 mm diameter, 1.8 kQ). As expected these no-flow junctions yielded very fast responses; their transient errors were respectively only -0.05 and -0.02 pH unit at 20 s in the difficult transfer from 1 to However these membranes proved impractical for general use owing to problems with inconsistent wetting, excessive porosity organic fouling etc.Outward flow establishes a steady-state profile in which the concentration of external ionic species decreases exponentially from the junction surface while the junction electrolyte (potassium chloride solution) shows a comparable increase (Fig. 1). The relaxation distance for such profiles is D/v where v is the flow velocity and D is the diffusion coefficient. When the junc-tion is transferred into a new solution the concentration of old solution species a t the junction surface should decay exponentially with a relaxation time of D/v2. To achieve a decay constant of 20 s the required velocity is v = (D/i)0-5m 10-3 cm s-l a very modest value.With a typical junction ceramic (1 mm diameter 15% porosity) this corresponds to a flow-rate of only 4 p1 h-1. However the junctions of conventional silver - silver chloride electrodes rapidly become clogged with precipitated silver chloride so the response becomes slow owing to lack of adequate flow.lG High flux with low total flow can be obtained by restricting the flow to a small aperture. Fig. 6 shows an “annular-pore” junction formed by inserting a silica fibre through a hole in 125-pm poly(viny1idine fluoride) film. Even small flow-rates (ca. 2 pl h-l) yielded very fast responses. The annular shape of the pore helped to prevent clogging by particulates. How-ever practical problems with the junction included a requirement for a positive flow and for a separate restrictor to control the flow-rate.The geometry of the interface between the two liquids is not well defined at the pore and depends on flow-rate; unstable results were obtained if a positive flow was not maintained. M hydrochloric acid and below 0.01 pH unit in all other transfers. Outward flow of junction electrolyte is another way of attaining a fast response. f - Quartz fibre PVF Film 125’ pm _F_ Outside solution Fig. 6. Annular-pore reference junction. With an outward flow-rate of 1.6 p1 h-l the transient error at 20 s was below 0.01 pH in all solution transfers. Junction resistance in 4 M potassium chloride solution was 2.6 kn. For the sake of discussion the observed “static” errors can be separated into two classes: very slow transient errors with discrepancies at 15 or 30 min which would eventually dis-appear with prolonged soaking and true static errors which would persist no matter how long the soaking.Some of the observed static errors probably belong to the first class (particu-larly those in 1 M hydrochloric acid all of which have the same polarity as the transient errors). The diffusional relaxation time for a 1-cm junction is roughly 1.4 h. If such an electrode is routinely stored in pH 7 buffer standard (0.05 M phosphate) between measurements as is usual practice then pH 7 buffer not 4 M potassium chloride solution effectively serves as the Some of the tested references (C D1) had junctions longer than 1 cm 436 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol. 108 junction electrolyte.Response problems would not necessarily be evident during calibration in various NBS-type standard buffers as these are fairly similar in transference and ionic strength. Upon insertion into the solution to be measured a fairly stable diffusion potential corresponding to the phosphate - sample interface would rapidly be established at the outer surface of the junction ; however with very dilute or non-equitransferent samples this potential may differ considerably from the correct value obtained against 4 M potassium chloride solution. To determine the possible magnitude of such errors a junction using pH 7 buffer was evalu-ated. An electrode body with ceramic junction was filled with gelled pH 7 buffer (0.05 M phosphate 2% agarose) and inner contact was made via a miniature 4 M potassium chloride solution bridge.The observed static error was 0 mV in 1 M potassium chloride solution +48 mV in 1 M hydrochloric acid and -14 mV in The observed value in 1 M hydrochloric acid corresponded well to the calculated value (54 mV) but the static error in M hydrochloric acid was smaller than expected (-44 mV). In distilled water an initial static error of about -65 mV was observed, Results for this electrode are shown in Fig. 7. M hydrochloric acid. J pH7 I M 1 M 1 0 - 4 ~ 1 M I M 1 0 - 4 ~ buffer KCI HCI HCI HCI KCI HCI C .- U C .-L' I Solution composition I Distillea water pH7 p H 4 pH7 pH 10 p H 7 > E buffer buffer buffer buffer buffer 1 (p==106 ~2 cm) m 0, .-0) 0 -1.3 mV Q) W U U g t U g o --65 mV -50 -- 0 L E z Ip I From pH 7 -100 I x f f e r -- t 9 H -501 d c 0 1 d - :L -100 2 0 5 From pH 7 buffer 1 M HCI Y-l 10 15 20 25 Fig.7. Anomalous response of no-flow junction saturated with pH 7 standard buffer. Ceramic double-junction reference filled with 0.05 M phosphate buffer gelled with 2% agarose. The arrows indicate times of brief agitation A@ril 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 437 which held steady for a while before drifting to more negative values. This initial error is probably caused by the unbalanced transference of phosphate buffer. Although the electrode response was very slow in certain transfers it was surprisingly fast in the transfers from to 1 M hydrochloric acid and then to 1 M potassium chloride solution; however the potentials obtained were erroneous.Finally the electrode was tested for performance in pH 4 7 and 10 buffer standards. Static errors relative to pH 7 buffer were below 8 mV and there was no observable kinetic or stirring error. During two-point calibration the electrode would appear to perform perfectly giving no indication that there would be problems in other solutions. Even with three-point calibration slope and offset adjustment would reduce the discrepancies at pH 4 7 and 10 to only *0.02 pH unit, which is likely to be ignored. Errors of the above type may be expected in routine use of no-flow gel electrodes and clogged flow-type electrodes with long junctions. The second class of static errors are those persisting at the steady state.Such errors are evident in Table IV; the long-term offsets for electrodes B and D1 in M hydrochloric acid (relative to 1 M potassium chloride solution) were substantially positive even though the short-term transients were negative in polarity. A likely explanation is the presence of fixed positive charge within the junction pores as demonstrated below. Large shifts in potential were observed with cessation of flow in the “annular-pore” junction of Fig. 6. This indicates non-standard interfacial geometry as another possible source of static error. Junction geometry may be more critical in solutions of low ionic strength where inter-diffusion can drastically change the concentration of the sample while having negligible effect on the junction electrolyte.Thus when 4 M potassium chloride solution flows out of an iso-lated pore into a dilute electrolyte a broad continuous-mixture type of interface results but when the dilute electrolyte flows inward into the potassium chloride solution interdiffusion causes the concentration interface to collapse to the immediate vicinity of the pore resulting in an interface with constrained and likely spherical diffusion. According to Guggenheim’s criteria,l the former interface should be very stable and the latter unstable. This may explain the empirical observation7 that outward flow is required for stability of leak-type junctions in dilute electrolyte. Similarly in this study the 4 M potassium chloride open-capillary junction was immediately stable to better than 1 mV after insertion into 1 M potassium chloride solution and 1 M hydrochloric acid but tended to drift slowly by a few millivolts after insertion into In contrast flat-surfaced ceramic junctions with adequate positive flow stabilised quickly in M hydrochloric acid probably because of greater control over the geometry and mixing pattern at the liquid interface.Other possible sources of static error include the introduction of extrinsic charge or ion specificity by clogging of junctions with solution precipitates (e.g. Ag,S and Hg,S in sulphide-containing solutions) and also the redox sensitivity of electronically conductive junction materials (e.g. platinum) when clogged ; however these possibilities were not investigated in detail. The origin of stirring errors is less evident.A seemingly attractive hypothesis is that the stirring potential is an electrokinetic effect analogous to a streaming potential and arises from physical displacement of a layer of charged solution. However such an explanation is untenable for several reasons. Firstly it is hard to envisage how fluid motion parallel to the junction surface could induce a substantial electrokinetic potential perpendicular to the surface. Also when the solution outside and within the junction is the same (e.g. loA4 M hydrochloric acid throughout) the stirring potential disappears (see below). This would not be expected if the stirring potential was due to a surface electrokinetic effect. Further it seems particularly unlikely that the potential is due to sweeping of counter ions away from a charged exterior surface as the counter ionic cloud is centred roughly 0.03 pm from the surface at M ionic strength (and decreases as I - l I 2 ) while the thickness of the “stagnant” Nernst diffusion layer exceeds 10 pm for even vigorous stirring17 Finally the diffusional relaxation time correspond-ing to a 0.03 pm cloud thickness is below 1 ps and voltages induced by current pulses applied through reference electrodes gave observed decay times below 200 ps suggesting that electro-kinetic effects should dissipate several orders of magnitude faster than the relaxation time observed for the stirring effects about 3-6 s in Fig.5 ( b ) . In contrast a relaxation time of 3-6 s corresponds to diffusion into a layer of roughly 120-170 pm which falls in the thickness range of Nernst diffusion layers in moderately stirred to unstirred solution (50-500 pm).Therefore the stirring potential is more likely associated with disturbance of free-diffusion Thus a fast response is no guarantee of accuracy. M hydrochloric acid 438 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol. 108 profiles extending beyond the junction into the adjacent convection-free layer of solution. When the external solution is stirred constrained diffusion is imposed at the external junction surface whereas free diffusion may propagate beyond the external junction surface in unstirred solution (Fig. 8). Thus stirring could affect the junction potential by altering the interfacial geometry or by reducing the ionic strength at the junction surface which would increase the boundary potential due to junction charge [equation (4)].Both mechanisms are supported by experiment. Large changes in potential (3040 mV) were observed upon agitation of very slow-flowing pore electrodes where interfacial geometry was the only likely factor. Con-versely stirring potentials were also observed with 4 M potassium chloride solution-filled ceramic junctions agitated in M potassium chloride solution where the single-species interface precludes geometric effects. Uniform CI Inside Physical Outside solution I junction I solution Fig. 8. Effect of stirring on junction concentration profile. The boundary layer thicknesses and con-centrations have been greatly exaggerated for clarity. Stirring changes the outer profile from free to con-strained diffusion and decreases the concentration at the junction surface.With conventional reference electrodes the boundary-potential explanation of stirring error seems plausible because diffusion should yield significant concentrations of potassium chloride at the exterior surface of even fairly thick junctions. For example the surface potassium chloride concentration of a no-flow 1 cm long 10% porosity junction should be about 0.004 M in rapidly stirred solution (0.01-cm Nernst diffusion layer) and about 0.02 M in unstirred solu-tion (0.05-cm Nernst layer). Ironically the stirring error should disappear with very thick junctions leaving only a large static error With junctions of moderate thickness the static error should reach a limiting value in very dilute solutions and not be nearly as great as when the external solution is drawn into the junction by negative flow.Finally both static and stirring errors should be greatly suppressed with a positive flow of electrolyte through thick junctions. For example a 0.001 cm s-l positive flow would yield potassium chloride transport equivalent to a 0.2-cm (=D/lO%v) junction. The junction-charge theory of static and stirring error is confirmed in Fig. 9 which shows the changes in potential of a ceramic-junction reference electrode as solution stirring and junc-tion electrolyte flow are varied. The observed behaviour is as expected for a fixed negative charge in the junction. Confirmatory features include the following a large offset error but no stirring potential after dilute external electrolyte has been drawn deep into the junction (a i k 1); a minimum offset with stirring potential when potassium chloride solution is delivered to the junction surface by outward diffusion or flow of junction electrolyte (b-f m) ; an inverse relationship between stirring potential and outward flow-rate (b c n) ; the absence of a streaming potential with continuous outward flow which keeps the junction filled with 4 M potassium chloride solution (n); the presence of a streaming potential after prolonged inward flow which fills part of the junction with dilute electrolyte.(h j) ; and the absence of an “end-streaming” potential due to diffusional exchange (p).Also as predicted the polarity o April 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 439 the stirring error is the same as that of the static offset; and the polarity of the streaming potential which depends on direction of flow matches that of the static offset when flow is outward.According to the junction-charge theory clogging of junction pores by extraneous agents could aggravate static and stirring errors. Partially clogged pores have high surface to volume ratios which should increase the ionic-strength threshold for errors due to surface charge. Also the clogging agent might contribute surface charge of its own. For example Fig. 10 shows the reference response of a flow-type ceramic silver - silver chloride combination elec-trode which though having a clogged junction was considered functional by its users. Tests in pH 4 7 and 10 buffers indicated only a slow response and moderate relative offsets.How-ever when the electrode was transferred from pH 7 buffer to M hydrochloric acid the electrode showed very large positive static and stirring errors indicating the presence of posi-tive charge in the junction. On transfer to 1 M hydrochloric acid the potential decreased overall as the charge effects were suppressed whereas normally the initial shift is positive owing to transient diffusion potentials. M hydrochloric acid the potential increased to an even higher offset error (exceeding + 100 mV). Normally a negative diffusional transient would be observed with this transfer. The increased offset error in M hydrochloric acid after exposure to 1 M hydrochloric acid is probably due to H+ adsorption to materials within the junction.This H+ adsorption phenomenon is also evident to a lesser extent in earlier data with unclogged ceramic junctions. In Figs. 3 and 7, for example the stirring potentials are initially positive after transfer from 1 to M hydro-chloric acid but are negative after prolonged exposure to dilute solution. This change When the electrode was returned again to 0 head after drawingext. soh. deep into junction 0 > E -10 Fig. 9. 40cm zcm ocm -10 ocm -1ocm- head cm KCI Junction electrolyte (4 M KCI) pressure head I 10cmhead l c m ocm ocm Streaming potential develops as dilute M) electrol9e fills junction (h) Stirring potential is absent as solution is same within and outsidc junction (i) M KCI at junction outer surface causes large static error (9) 0 10 20 Ti me/mi n 30 40 Offset diminishes Stirring potential, stir error returns residual offset error as KCI concentration increases at surface are inversely related to positive flow-rate Flow-proportional streaming pote-"-' (plus fixed offsc., when M KCI fills stirrin! 7 potent cm Stir 10 cm Continuous stirring Brief intervals of +llOcm head -1 briefly stopped Qualitative confirmation of space-charge theory of static and stirring errors.Ceramic junction in M potassium chloride solution ueYsu.s similar capped no-flow junction 440 BREZINSKI KINETIC STATIC AND STIRRING Analyst Vol. 108 probably reflects Hf desorption from pore walls with net surface charge changing from positive to negative.Thus some of the kinetic behaviour of reference electrodes e.g. anomalously slow recovery from stirring is caused by slow adsorption or desorption of charge within the junction. Boundary potentials due to junction charge seem a likely explanation for the errors observed by Illingworth.6 A highly charged junction characterised by partial clogging or very hetero-geneous pore size would act as a Nernstian concentration sensor shunted by a liquid-junction leak yielding the observed sub-Nernstian logarithmic dependence on salt concentration. Accordingly he found that errors were diminished considerably when potassium chloride electrolyte was allowed to seep out and evaporate on the junction surface prior to electrode use. Tim e/m i n Solution composition 10-4 M 1 M 10-4 M 1 M \ HCI HCI HCI KCI PH J iuffer 0 5 10 15 20 25 Time/mi n Fig.10. Anomalous response of reference electrode with protein-clogged junction. The electrode was a silver - silver chloride combination with a ceramic junction used for 6 months, including storage in pH 7 buffer and measurement of protein solutions. The junction resistance was about 10 kR. The arrows indicate times of brief agitation. Conclusion Overall the performance of commercial reference electrodes was found to be surprisingly poor considering the very widespread application and presumed accuracy of pH measurement and the relative lack of attention to reference electrode performance characteristics and problems. The seemingly “passive” role of reference electrodes has probably resulted in many of their performance aspects being overlooked.Also reference problems tend to be suppressed in standard buffers where accuracy is usually checked. Therefore users may see no reason to disbelieve erroneous readings obtained in non-standard environments that have to be taken at face value. The observed errors apart from belying the significance typically displayed by pH meters, are large enough to be of practical consequence-they often correspond to many-fold differences in H+ activity. pH monotoring and control are key aspects in chemical processing and environ-mental protection and these fields frequently require measurement at extreme pH or low ionic strength. Problems may arise even under moderate conditions; some enzyme buffers prepared using the Fig.10 electrode were off by 0.5 pH unit April 1983 ERRORS OF LIQUID JUNCTION REFERENCE ELECTRODES 441 In view of these large errors the use of special equitransferent solutions for low ionic strength measurements seems of dubious benefit. An exactly equitransferent solution in place of 4 M potassium chloride solution would eliminate a roughly 6 mV shift in liquid junction potential between calibration buffers and pure water (lo-’ M). However much larger static errors were observed with some conventional junctions at ionic strengths as high as M. Also equitransferent filling solutions are often less concentrated than 4 M and a reduction in electrolyte concentration should result in a roughly proportionate increase in static and stirring error due to junction charge [equation (4)].As discussed above and by Picknett,l2 the residual junction potential for 4 M potassium chloride solution in very dilute solutions should be a predictable logarithmic function of solution conductance. Conductance measurement and compensation particularly if accomplished directly by the ion meter are a possible alternative that could yield greater accuracy than replacing 4 M potassium chloride solution with a more equitransferent but less concentrated electrolyte. Colloidal suspension effects were studied by Jenny et aZ.,18 who observed differences as large as 5 pH units between the pH measured in a sediment of electrically charged particles such as clays and ion-exchange resin beads and the pH measured in the supernatant fluid.This pH discrepancy was attributed to a large junction potential between the reference electrode and the sediment. However their conclusion appears dubious in the light of the present findings. Provided that the concentration of the potassium chloride junction electrolyte is high at the junction surface potassium chloride should diffuse a distance into the sediment suppressing the boundary potential at the sediment - junction interface in accord with equation (4). Further the gradient of junction potassium chloride diffusing from the surface to the interior of a uniformly packed sediment should generate no additional potential [equation (2) with do/dx ZI = 01. Thus the over-all junction potential should be relatively low. On the other hand equation (4) predicts a much larger shift in potential a t the boundary between the charged sediment and the dilute supernatant.Owing to this potential difference hydrogen ions should be partitioned unequally across the boundary yielding equal electrochemical potentials U + (RT/F)ln aH+ (sensed by the pH electrode) but unequal chemical potentials and true pHs. This view is supported in a separate study.lg The origin of actual junction potentials is considerably more complex than has generally been assumed and it is clearly inadequate to view the reference junction in practical terms as a mere “leak.” However much present-day technology and practice seem to be based on an inadequate understanding of the requirements for satisfactory junction performance. To prevent static and stirring errors the porous junction material should exhibit a low charge to volume ratio at low ionic strength even after exposure to pH extremes.Boundary potentials due to residual junction charge should be further reduced by an adequate concentration of potassium chloride delivered to the junction surface by diffusion or flow. Delivery of ample potassium chloride is also especially critical to accurate pH measurement in charged sediments and suspensions. Fast response requires either adequate outward flow or very thin junctions. Most commercial electrodes particularly those designed for process applications do not meet these criteria. Also the extent to which junction performance is rapidly degraded by silver chloride clogging has not been appreciated (vix. “flow-type” silver - silver chloride electrodes).In view of the low toxicity and excellent thermal stability of the silver - silver chloride electrode the clogging problem is worth preventing and electrode designs with pure 4 M potassium chloride electrolyte are discussed elsewhere.l6 Electrode performance claims are usually based on tests in standard buffers which are largely ineffective in differentiating be-tween reference electrodes. Tests similar to those described could provide the basis for developing improved electrodes and more meaningful performance specifications. Improvements in user technique and awareness are also suggested. Thick junctions without flow should not be stored in standard buffers or be used for critical or continuous measurement. It is a common practice to store electrodes in very dilute solutions to “condition” them for low ionic strength measurements.This approach is completely wrong as virtually all of the junction anomalies are increased by reduction of ionic strength within the junction. Instead, reference electrodes should be stored in 4 M potassium chloride solution which suppresses anomalies and minimises junction clogging. The prevalent recommendation for continuous stirring during ion-selective electrode measurements may need qualification because stirring generally (but not always)16 increases the reference error at low ionic strength. In blood-gas and other electroanalytical instrumentation the reference junction is sometime 442 BREZINSKI located in a recess or downstream and is connected to the measured sample by a segment of running buffer.This configuration could constitute a sandwiched (double) interface which can cause static errors and should be avoided. Likewise an intervening particle-free layer between a measured colloid and the concentrated junction electrolyte must be avoided. The above recommendations are necessarily general because the requirements for optimum performance will vary with the electrode type and application. Specific procedural recom-mendations must be determined by appropriate user - manufacturer testing. A practical fast-responding no-flow junction based on the thin-membrane approach seems feasible but is subject to stringent material requirements. In addition to strength chemical durability and immunity to fouling the membrane material must have low porosity while maintaining a low charge to volume ratio.Further fast response in such a junction seems incompatible with accuracy in highly charged colloids because a low potassium chloride con-centration at the exterior junction surface would give rise to a colloidal boundary potential, whereas response will be slow if the interdiffusion profile extends substantially beyond the physical junction. Fortunately nearly ideal general-purpose response seems attainable with conventional flow-type reference electrodes. It is hoped that this study will contribute to improved accuracy in electroanalytical methodologies. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. References Guggenheim E. A. J Am. Chem. SOC. 1930 52 1315. Guggenheim E. A. J . Phys. Chem. 1930 34 1758. Bates R. G. “Determination of pH,” Second Edition Wiley Toronto 1973 p. 85. Covington A. K. Anal. Chim. Acta 1981 127 1. Bates R. G. CRC Crit. Rev. Anal. Chem. 1981 10 247. Illingworth J. A. Biochem. J . 1981 195 259. Covington A. K. in Durst R. A. Editor “Ion Selective Electrodes,” NBS Special Publication 314, Planck M. Ann. Phys. 1890 39 161. Bass L. Trans. Faraday SOC. 1964 60 1656. Hafemann D. R. J . Phys. Chem. 1965 69 4226. MacInnes D. A. “Principles of Electrochemistry,” Reinhold New York 1939 p. 232. Picknett R. G. Trans. Faraday SOC. 1968 64 1059. Dean J. A. Editor “Lange’s Handbook of Chemistry,” Eleventh Edition McGraw-Hill New York, Teorell T. Proc. SOC. Exp. Biol. Med. 1935 33 282. Meyer K. H. and Sievers G. F. Helv. Chim. Acta 1936 19 649. Brezinski D. P. Anal. Chim. Acta 1982 134 247. Bockris J. O’M. and Reddy A. K. N. “Modern Electrochemistry,” Plenum Press New York 1970, Volume I pp. 199 and 729; Volume 11 p. 1058. Jenny H. Nielson T. R. Coleman N. T. and Williams D. E. Science 1950 112 164. Brezinski D. P. Talanta 1983 in the press. National Bureau of Standards Washington DC 1969 Chapter 4 (general review). 1973 p. 6-30. Received July 5th 1982 Accepted September 24th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800425

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Analyst April 1983 Vol. 108 PP. 443-451 Analytical Use of the Kinetics of Complex Formation: Simultaneous Determination of Iron and Cobalt by Differential Kinetic Methods 443 L. Ballesteros and D. P6rez-Bendito” Department of Analytical Chemistry Faculty of Sciences University of Cdrdoba Cdrdoba Spain The kinetic determination of iron and cobalt mixtures without a prior separa-tion is described. The methods are based on the differential reaction rate between pyridoxal thiosemicarbazone and these metallic ions. Various ratios at the M level of the two ions can be determined photometrically by using either the “logarithmic extrapolation” or the “single-point” methods. These two methods are compared. Keywords Iron and cobalt simultaneous determination ; pyridoxal thio-semicarbazone ; kinetics of complex formation ; differential kinetic methods This paper reports part of a general study of the use of thiosemicarbazones in kinetic analysis.So far we have reported results on catalytic - kinetic analysisl-4 and catalytic titrations.596 We have now extended these studies with the use of thiosemicarbazones in differential kinetic analysis. We describe here the kinetics of the complex formation between pyridoxal thiosemi-carbazone (PT) S II CH=N-NH-C- NHp and iron and cobalt in order to establish procedures for the simultaneous determination of these two ions by means of differential kinetic methods. The methods are based on the fact that the formation of the cobalt - PT complex is rapid whereas the formation of the iron - PT complex is very slow.Several differential rate methods for the analysis of mixtures of metals based on redox rea~tions,~-ll catalysed reactions12-15 and ligand-exchange r e a c t i o n ~ l ~ - ~ ~ has been reported. However only one application of the reactions of complex formation in differential kinetic analysis has been published involving the analysis of mixtures of nickel and cobalt.2s The technique described in this paper for the analysis of iron and cobalt mixtures is very simple and the use of a stopped-flow technique common in the procedures cited above is not necessary. Experimental Apparatus A Perkin-Elmer 575 ultraviolet - visible spectrophotometer with 1 .O-cm glass cells and equipped with an electronic thermostat was used for the kinetic measurements.A Radio-meter PHM 62 pH meter with a combined glass - calomel electrode and a Hewlett-Packard HP-85 computer were also used. Reagents All reagents were of analytical-reagent grade. Pyridoxal thiosemicarbazone was synthesised by the condensation of pyridoxal chloro-hydrate with thiosemicarbazide.27 A 0.1 yo m/V solution of the reagent in “’-dimethyl-formamide was used. * To whom correspondence should be addressed 444 BALLESTEROS AND PI~REZ-BENDITO USE OF KINETICS OF Analyst Vol. 108 Standard solutions of iron(II1) (43.3 pg ml-l) and cobalt(I1) (47.6 pg ml-1) were pre-pared by dilution of iron( 111) nitrate and cobalt nitrate solutions respectively and standardised gravimetrically. Sodium acetate - acetic acid buffer solution [C (total concentration) = 1.70 M and pH = 3.851 was used.Procedure for the Simultaneous Determination of Iron and Cobalt Logarithmic extrapolation method To a 10-ml calibrated flask containing up to 30pg of iron and cobalt in a mixture add 3 ml of 1 M potassium nitrate solution 2 ml of sodium acetate - acetic acid buffer solution (pH 3.85) and an appropriate volume of distilled water so that when 0.5 ml of 0.1% reagent solution is finally added (with great care and last to prevent prior mixing of the solutions), the total volume becomes 1O.Oml. The mixture is warmed in a water-bath at 25 & 0.1 "C for 5 min then shaken (time t = 0) and one portion is transferred into a 1.0-cm cell thermo-stated at 25 5 0.1 "C. The change in absorbance at 425 nm with time is recorded between minute 2 and minute 3.Then the mixture is allowed to stand until development is complete (30 min) and the absorbance is measured (A,). The treatment of the data obtained is indicated later for the mixture analysis. Single-p oint met hod The samples are prepared as described above but the absorbance is measured exactly 1 min after shaking the samples. The mixture is allowed to stand until the reaction is complete (30min) to obtain A,. From a previously plotted calibration graph the molar fraction is obtained together with the amounts of iron and cobalt as indicated later. Results and Discussion Spectrophotometric Characteristics of Iron - PT and Cobalt - PT Complexes In a weakly acidic medium (pH 3.95) PT forms yellow - brown chelates with iron(II1) and cobalt(II) the solutions of which are stable and have maximum absorption at 425 and 440 nm respectively.The absorbance of both complexes at 425 nm is not affected appreci-ably by temperature ionic strength dielectric constant or acetate ion concentration. The molar absorptivities are 1.62 x lo4 and 1.15 x 104 1 mol-l cm-l respectively. The metal to ligand ratio for the iron - PT and cobalt - PT complexes were determined by Job's method at two different wavelengths for each complex (425 and 440 nm for the iron complex and 440 and 470 nm for the cobalt complex). A ratio of 1 3 was found in both instances. The reagent probably behaves as a bidentate ligand forming octahedral complexes. The con-dition stability constants from Job's method data were calculated (logK = 16.4 * 0.2 for the iron complex and logK = 16.9 & 0.5 for the cobalt complex).In order to establish the oxidation state of the metal ions in each complex the effect of oxidising and reducing agents was studied. Whereas the iron - PT complex is formed in the presecce of peroxydisulphate or periodate it is not produced in the presence of ascorbic acid or hydroxylamine and it is concluded that the trivalent iron(II1) ion forms the iron - PT complex. In the presence of the same oxidising and reducing agents the cobalt - PT complex is formed and the oxidation state cannot be determined. On the other hand complex formation does not take place when PT and cobalt(I1) are mixed in an inert atmosphere. However when a current of air is bubbled through the solution rapid formation of the cobalt -PT complex is observed.Therefore the dissolved oxygen oxidises the cobalt(I1) to cobalt(III) which forms the cobalt - PT complex. Kinetic Study of Complex Formation recording of the absorbance - time curves at 425 nm (Fig. 1). after 25-30 min for the iron - PT complex and after 2-3 min for the cobalt - PT complex. The kinetics of the formation of these complexes were followed spectrophotometrically by The reactions are complet April 1983 COMPLEX FORMATION DETERMINATION OF FE AND Co 445 Effect of Reaction Variables Variation of the ionic strength has little influence on the rate constant and the ionic strength was fixed at 0.3 M with potassium nitrate. On the other hand an increase in the NN'-dimethylformarnide concentration decreases the rate constant and a 5% concentration of the organic solvent was chosen for the formation of both complexes.0.8 I 0.6 8 m $ 0.4 u) 2 0.2 0 10 20 30 Time/m in Fig. 1. Absorbance versus time graphs a t 426 nm for (A) iron - PT complex and (B) cobalt - PT complex. CM = 2.60 pg ml-l Fe(II1) or 2.85 pg ml-l Co(I1); CB = 2.08 x 10-4 M; pH = 3.96; T = 26 "C; and p = 0.3 M (KNO,). Efect of temperatwe The effect of temperature on the rate constant for both complexes was studied in the range 2045 "C. The rate constant increased as the temperature increased and in both instances a temperature of 25 "C was selected. From the rate constant values and the Arrhenius equation the activation energies for the two complex formation reactions were calculated and were found to be 17.32 and 2.55 kcal mol-l for the iron - PT and cobalt - PT reactions respectively.As the activation energy for iron - PT complex formation is high temperature control is critical. With the electronic thermostat used a high degree of temperature control is possible as is shown by the small relative error (less than 1%) in the determination of the rate constant for the iron - PT complex. Efect of reagent concentration The effect of the reagent concentration on the rate constants was studied in the range 2 x lo4 - 5 x lo4 M for the iron - PT complex and 1.0 x M for the cobalt - PT complex. These concentration ranges are small because the lower limit is con-ditioned for total completion of complex formation reaction. On the other hand larger amounts of the reagent cannot be used because the complex formation reactions are very rapid especially with the cobalt - PT complex.A 2.08 x 10-4 M concentration of the reagent (0.5 ml of 0.1% solution) was selected. In order to establish the partial orders of reaction with respect to the reagent concentra-tion the logarithm of the rate constant was plotted against the logarithm of the reagent concentration. The reaction rate of complex formation is directly proportional to the square root of the PT concentration for the iron - PT complex and to the square of the PT concentration for the cobalt - PT complex. Therefore the partial orders are 0.5 and 2, respectively. The differences found in the partial orders may be due to the nature of the two reactions being different as is explained in detail later.- 2.5 446 BALLESTEROS AND P~REZ-BENDITO USE OF KINETICS OF Analyst VoZ. 108 Efect of acetate ion The influence of the acetate ion was studied in the range 10-3 - lo- M for both complexes. The rate constant of iron - PT complex formation depends greatly on the sodium acetate concentration. A logarithmic plot showed that the reaction rate is inversely proportional to the acetate ion concentration. In contrast the rate constant for cobalt - PT complex formation is independent of the acetate ion concentration and therefore this reaction is of zero order with respect to acetate ion concentration. The rate constant ratio kc,/kF, increases as the acetate concentration increases and the determination of cobalt and iron mixtures is favoured by high acetate concentrations.An acetate concentration of 0.34 M (2 ml of buffer solution) was chosen because larger amounts of acetate would excessively and disadvantageously delay the iron - PT complex formation. Efect of $H The variation of the rate constants with pH was studied by addition of various amounts of sodium hydroxide to a fixed amount of acetic acid giving a pH range of 3.4-5.3 for the iron - PT complex and 3.04.5 for the cobalt - PT complex. At pH 3.95 the rate constant of iron - PT complex formation attains a minimum value. Therefore we selected a buffer solution of pH 3.85 which yields a final pH of 3.95 probably owing to the effect of the organic solvent added. In this study it is necessary to take into account that as the acetic acid is a weak acid the acetate ion concentration increases as the pH increases.This is not of importance in cobalt - PT complex formation but it is critical in iron - PT complex formation. In the latter instance the rate constant depends on both H+ and the acetate ion when the pH is changed. To prevent this difficulty an attempt was made to obtain partial orders of reaction with respect to H+ and AcO- based on the theoretical calculation of a corrected rate constant obtained by means of a fixed acetate concentration as reference, expressed by From this kcor value a partial order of -1 with respect to the acetate concentration is found. The same partial order with respect to the H+ concentration was also obtained. The same results for orders with respect to H+ and AcO- were found by using a computer program in which pair orders for hydrogen and acetate ions from -2 -2 to 2 2 (half by half units) were employed (i.e.arbitrarily assign a value to the partial order of reaction with respect to H+ and another different value to the partial order of reaction with respect to AcO-) and assuming as optimum orders those which give a smaller relative standard deviation in the values obtained. Rate Equations From this kinetic study the following equations are suggested for the formation of the iron - PT and cobalt - PT complexes in sodium acetate - acetic acid medium at pH 3.95 : The various kinetic dependences indicated above are summarised in Table I. = kFe (PT)t (Fe3+) (AcO-)-l (H+)-l d(Fe - PT) dt = kco (PT) (Co2+) d(C0 - PT) dt where kpe and kco are the conditional rate constants.From these equations we can say that the formation of the iron - PT complex is a reaction of ligand substitution between the reagent and the acetate ion. This is supported by the fact that the acetate ion forms colour-less complexes with iron(II1) and mainly because the rate of iron - PT complex formation decreases as the acetate concentration increases. As the conditional stability constants of the acetate complexes are knownm (log k = 3.4 log k = 6.1 and log k = 8.7) we calcu April 1983 q -1.0 - a A -1.5 COMPLEX FORMATION DETERMINATION OF FE AND Co --447 TABLE I SUMMARY OF KINETIC DATA PARTIAL ORDER OF REACTION Iron - PT complex Cobalt - PT complex A A \ r 1 Concentration Partial Concentration Partial Species range order range order Metal .. 0.3-3.0 pg ml-l 1 0.4-3.0 pg ml-1 1 PT . . . . 2 x 10-4-5 x ~ O - * M 0.6 1 X lO-L2.5 X w 4 M 2 AcO- . . . . 10-2-10-3 M -1 10-*-10-3 M 0 H+ . . 5 x 10-6-4 x 1 0 - 4 ~ -1 3 x 10-5-10-SM 0 lated the molar fractions of the different iron species in the solution (ao = 1.9 x 2.2 x 10-3; cc2 = 0.05; a3 = 0.95). could be written as a = Therefore the over-all reaction of ligand substitution Fe(CH,COO) + 3PT -+ Fe(PT) + SCH,COO-On the other hand we can assume that the reaction between cobalt(I1) and PT involves complex formation as the rate constant depends only on the metal and reagent concentra-tions. Determination of Conditional Rate Constants Under experimental conditions of constant pH and with a large excess of sodium acetate and reagent with respect to the metallic ion concentration the reactions are pseudo-first order with respect to iron(II1) and cobalt(I1) ions.Therefore under these conditions the integrated equation in absorbance terms can be written as Log ( A - A ) = log A - - kM t 2.303 where k is the conditional rate constant and Ao A t and A are the absorbances after a time t = 0 t equal to any time and t equal to a very long time respectively. A graph of log(A - A ) against time (Fig. 2) is a straight line whose slope allows the determination of the conditional rate constant of the complex formation reaction. The average values thus obtained for 11 determinations are k, . PT = 0.248 0.002 min-l and kc PT = 2.6 -j= 0.1 min-1. :'. B I I I I 0 60 120 180 Ti m e/s Fig.2 Graph of log(A,-At) vevssus time for the conditional rate constant calculation for (A) iron -PT complex and (B) cobalt - PT complex. Conditions as in Fig. 1 448 Differential Kinetic Determination of Iron and Cobalt We consider the following pseudo-first-order reactions : BALLESTEROS AND P~REZ-BENDITO USE OF KINETICS OF Analyst VoZ. 108 ka kB A + R - P B + R - P where k and k are the conditional rate constant and k > k,. When these reactions proceed simultaneously and their rate constants are independent of each other (the sum of the concentration of A and B reacting to form a comrnon product P is given) at any time, t we can write where [Ale and [Bl0 are the initial concentrations of A and B; [ A ] [B] and [PIt are the concentrations of A B and P respectively at time t and [PIm is the final concentration of P.Equation (1) is the basis of differential kinetic methods of which we used two for the simultaneous determination of iron and cobalt. Logarithmic Extrapolation Method When the faster reacting component of the mixture A has reacted to completion the term [Ale exp(- kAt) in equation (1) becomes negligible and by taking the logarithm of both sides of the equation we obtain and expressed as a function of the absorbance this equation becomes where eFe is the molar absorptivity of the iron - PT complex and I is the cell path length (1.0 cm). A graph of lo@ - A ) against time (Fig. 3) is a curve ending in a straight line which, when extrapolated to t = 0 allows us to obtain the initial concentration of iron [Fe3+Io.The initial concentration of cobalt [Co2+l0 in the mixture can be obtained by difference from the final absorbance Am. -0.7 -0.8 - 4 cn .-I -0.9 -1.0 I I 0 60 120 180 Ti me/s Fig. 3. Logarithmic extrapolation method ; log(A 00 - A t) as a function of time according to equation (3) for the simultaneous determination of iron and cobalt. CBI = 0.69 pg ml-' Fe(II1) and 0.96 pg ml-' Co(I1); CR = 2.08 x M ; pH = 3.95; T = 25 O C ; and p = 0.3 M (KNO,) April 1983 Single-point Method COMPLEX FORMATION DETERMINATION OF FE AND Co Dividing equation (1) by the total concentration C = A + Bo we obtain 449 where Ct is the total concentration at time t. TABLE I1 EFFECT OF DIVERSE IONS ON THE SIMULTANEOUS DETERMINATION OF 1.04 pg ml-l OF IRON AND 1.43 pg ml-1 OF COBALT Amount toleratedlpg ml-l Ion Be(I1) .. Sr(I1) . . Ba(I1) . . Pb(I1) . . Mn(I1) . . Bi(II1) . . As(II1) . . Cr(II1) . . Cd(I1) . . Zn(I1) . . Cyanide . Phosphate Chloride . Bromide Thiocyanate :4#-;) - - Added as . . Be(NO,), . . Ca(NO,) . . Sr(N03) Ba(NO,) Pb(NO,) Mn(NO,) . . Bi(NO,) Na,AsO . . CrC1 Cd(NO,) Zn(NO,) KCN N%PO . . NaCl KBr NaSCN * . - * Mg(N03)2 r -Logarithmic extrapolation method 26 600 260 600 1000 1000 1000 26 260 2 25 10 6 6 600 600 260 1 Single-point method 26 1000 260 600 1000 1000 1000 10 260 2 26 6 6 6 260 600 100 For a previously fixed time a graph of Ct/Co against the molar fraction of one of the components gives a calibration graph from which the molar fraction of that component in the sample can be calculated.As the final absorbance Am and the molar fractions of both metallic ions are known we can write [A1 0 x A = [A10 + [BIO From these two expressions the initial concentrations [A]. and PI0 can easily be calculated. The optimum time of measurement (topt) is established according to the Lee and Kolthoff expression29 : and is 1 min in this instance. Effect of Diverse Ions A study was made to determine the tolerance limits of various ions that may be present in the simultaneous determination of iron and cobalt using both the logarithmic extra-polation and single-point methods.The cations tested were added as nitrates or chlorides and the anions were added as sodium or potassium salts. Relative errors of less than &5y0 were considered negligible. Similar concentration levels of copper(II) nickel(II) vanadium-(V) silver(1) gold(II1) and palladium(I1) interfere positively owing to complex formatio 450 BALLESTEROS AND P~~REZ-BENDITO USE OF KINETICS OF Artalyst VoZ. 108 TABLE I11 ANALYSIS OF SOME SYNTHETIC MIXTURES BY THE LOGARITHMIC EXTRAPOLATION METHOD Added/pg ml-1 Found/pg ml-l Relative error yo - 7+ - Cobalt Iron Cobalt Iron Cobalt Iron 0.69 0.95 0.68 0.97 - 2.30 2.11 0.69 1.90 0.71 1.87 2.30 -1.68 1.38 0.96 1.36 1.00 -2.31 6.26 1.38 1.90 1.42 1.85 2.89 -2.63 with PT and similar levels of molybdenum(VI) tungsten(VI) antimony(III) tin(II), tartrate oxalate and EDTA interfere negatively by inhibition of the complex formation reactions.The permissible amounts of other ions are shown in Table 11. Determination of Iron and Cobalt in Synthetic Mixtures The logarithmic extrapolation and single-point methods have been applied to the deter-mination of various synthetic mixtures containing different relative proportions of both iron(II1) and cobalt(I1) ions. Tables I11 and IV summarise the results obtained. It can be concluded that when the mixture contains similar concentration levels of both cations the errors are small. On the other hand when the ratio of these concentrations is very large a greater error in the determination of the metallic ion added in the smaller amount is observed.However it should be noted that this disadvantage is common to every additive method for the simultaneous analysis of mixtures. TABLE IV ANALYSIS OF SOME SYNTHETIC MIXTURES BY THE SINGLE-POINT METHOD Added/pg ml-l Found/pg ml-l - Iron Cobalt Iron Cobalt 0.35 0.35 0.35 0.69 0.69 0.69 0.69 1.38 1.38 2.08 0.48 0.96 1.90 0.48 0.95 1.90 2.86 0.95 1.90 0.95 0.36 0.33 0.33 0.67 0.69 0.67 0.71 1.38 1.35 2.10 0.47 0.97 1.93 0.50 0.96 1.94 2.83 0.96 1.96 0.93 Relative error yo Iron 1.99 -4.47 -4.74 -2.52 - 1.00 -3.44 2.71 - 0.43 - 2.68 0.87 Cobalt 1.86 1.27 6.44 1.08 1.85 - 0.97 0.91 2.89 -2.80 -2.13 The single-point method is more advantageous than the logarithmic extrapolation method because in general it gives smaller errors and the relative range of determination of the two ions in mixtures is wider.The accuracy and precision of the two methods applied to a mixture containing 0.87 pg ml-1 of iron(II1) and 0.95pgml-1 of cobalt(I1) are given in Table V. It can be seen that the methods prcposed here are more precise than most of those reported in the literature. TABLE V ACCURACY AND PRECISION OF THE DIFFERENTIAL KINETIC METHODS USED IN THE DETERMINATION OF IRON AND COBALT MIXTURES Relative standard Method Metal Relative error yo deviation yo (n = 11) Logarithmic extrapolation . . . . Fe (0.87 pg ml-l) -0.91 0.64 Co (0.96 pg ml-l) 0.76 0.65 Single-point . . . . Fe (0.87 pg ml-l) -0.94 0.69 Co (0.95 pgml-l) 0.68 1.0 A$ril 1983 COMPLEX FORMATION DETERMINATION OF FE AND Co References 451 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. Perez-Bendito D. ValcArcel M. Ternero M. and Pino F. Anal. Chim. Acta 1977 94 405. Ternero M. Pino F. Perez-Bendito D. and ValcArcel M. Michrochem. J. 1980 25 102. Ferrer J. L. and PBrez-Bendito D. Anal. Chim. Acta 1981 132 157. Moreno A. Silva M. PCrez-Bendito D. and ValcArcel M. Talanta in the press. Ternero M. Pino F. PCrez-Bendito D. and ValcArcel M. Anal. Chim. Acta 1979 109 401. Ternero M. Perez-Bendito D. and ValcArcel M. Microchem. J. 1981 26 61. Ingle J. D. Jr. and Crouch S. R. Anal. Chem. 1971 43 7 . Ohashi K.Kawaguchi H. and Yamamoto K. Anal. Chim. Acta 1979 111 301. Yonekubo T. Nagaosa Y. and Nakahigashi Y. Nippon Kagaku Kaishi 1974 2 269. Nagaosa Y. Yonekubo T. Satake N. and Seto R. Bunseki Kagaku 1972 21 215. Nagaosa Y. and Yonekubo T. Bull. Chem. SOC. Jpn. 1973 46 1677. Worthington J. B. and Pardue H. L. Anal. Chem. 1970 42 1157. Deming S. N. Anal. Chem. 1971 43 1726. Alekseeva I. I. Ruzinov L. P. Chernysheva L. M. and Kachaturian E. G. Izv. Vyssh. Ucheb. Wolff C. M. and Schwing J. P. Bull. SOG. Chim. FY. 1976 1 679. Tanaka M. Funahashi S. and Shirai D. Anal. Chim. Acta 1967 39 437. Mentasti E. Anal. Chim. Acta 1979 111 177. Ridder G. M. and Margerum D. W. in Waninen E. Editor “Essays on Analytical Chemistry (in Ito S. Haraguchi K. and Nakagawa K. Bunseki Kagaku 1978 27 334. Ito S. Haraguchi K. Nakagawa K. and Yamada K. Bunseki Kagaku 1977 26 554. Pausch J. B. and Margerum D. W. Anal. Chem. 1969 41 226. Albrecht-Gary A. M. Collin J.-P. Jost P. Lagrange P. and Schwing J.-P. Analyst 1978 103, Funahashi S. Yamada S. and Tanaka M. Anal. Chim. Acta 1971 56 371. Kopanica M. and Stara V. Collect. Czech. Chem. Commun. 1976 41 3275. Yatsimirskii K. B. Khachatryan A. G. and Budarin L. I. Dokl. Akad. Nauk SSSR 1973 211, Kitagawa T. and Fugikawa K. Nippon Kagaku Kaishi 1977 7 998. Perez-Bendito D. and Valckrcel M. Afinidad 1980 18 336. Sommer L. and Pliska K. Collect. Czech. Chem. Commun. 1961 26 2754. Lee T. S. and Kolthoff I. M. Ann. N . Y . Acad. Sci. 1951 53 1093. Zaved. Khim. Khim. Tekhnol. 1974 16 1445. Memory of Professor Anders Ringbom),” Pergamon Press Oxford 1977 p. 529. 227. 1139. Received August 20th 1982 Accepted October 18th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800443

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

452 Analyst April 1983 VoL. 108 pp. 452-456 Improvements to the Oxygen Flask Combustion Procedure for Assay of Halogenated Organic Compounds Duncan Thorburn Burns and Binod K. Maitin Department of Analytical Chemistry The Queen’s University of Belfast Belfast BT9 5AG An improved oxygen flask procedure is described for the accurate precise and rapid determination of chlorine bromine and iodine in halogenated organic compounds. Quantitative recoveries were achieved by using neutral hydrazine hydrate in the absorbing solution which may be titrated potentio-metrically directly with silver nitrate solution. Hydrazine hydrate does not interfere under acidic conditions. Accurate and precise results were obtained when the equivalence points were determined by Gran’s method.Keywords Halogen determination ; oxygen $ask combustion ; hydrazine hydrate ; potentiometric titration ; Gran’s method Several procedures for the complete decomposition of organic compounds have been described of which the oxygen flask combustion method is the most widely used. This method was devised by Hempell who applied it to the determination of sulphur in coal and organic materials. As Macdonald has noted293 few further applications appeared until SchOniger4s5 adapted the method to the micro-scale. Subsequently the method has been applied widely and a very large number of modifications and applications have been described and d i s ~ u s s e d . ~ ~ 9 ~ - ~ ~ In oxygen flask combustion problems may arise from the incomplete combustion of samples. Although filter-paper is the most common material used as sample carrier for solid samples, other materials such as cigarette paper polythene and poly(methy1 methacrylate) sheets have also been u ~ e d .~ ~ + Many workers add some auxiliary combustible material to assist complete decomposition. Glucose sucrose potassium or sodium nitrate paraffin wax naphthalene, phthalic acid benzoic acid and dodecanol have each been used as combustion aids.3*7,12J3 To overcome problems due to sublimation or volatilisation of certain highly halogenated aromatic compounds Awad et aL.12 recommended repeated folding of the paper carrier in all directions, so that several layers surround the sample. Complete absorption of the combustion products is an important factor in these determina-tions.In order to convert the liberated halogens into an ionic form it is normally necessary to use a reducing absorption medium. A large variety of reducing agents have been used for example neutral and alkaline hydrogen peroxide alkaline sulphur dioxide saturated aqueous hydrogen sulphite alkaline hydrogen sulphite alkaline sodium arsenite acidic sodium nitrite, alkaline sodium tetrahydroborate( 111) dilute ammonia solution ammoniacal peroxide and alkaline hydrazine ~ulphate.~$l~ Of these alkaline hydrogen peroxide was commonly used for the determination of chlorine and bromine but not for iodine.13 Alkaline hydrazine sulphate is now more favoured than alkaline peroxide and can be used for determination of chlorine, bromine and iodine.15-19 Detailed collaborative studies organised by Lalancette and co-workers showed that satisfactory results for the three halogens could be obtained using alkaline hydrazine sulphate.20,21 Mazzeo-Farina and MazzeoZ2 have suggested the use of alkaline hydrazine hydrate instead of alkaline hydrazine sulphate.In this study quantitative recoveries of the halide ions have been achieved using neutral hydrazine hydrate. Several organic halogen compounds have been analysed by potentiometric titration of the solution after oxygen flask combustion with silver nitrate titrant the equivalence points were deter-mined by Gran’s method.23 Experimental Apparatus The following apparatus was used an Orion 901 Microprocessor Ionalyser to read 0.1 mV; a calibrated ALGA Micrometer Syringe burette of 0.5 ml capacity; a Pye silver billet indicator electrode; an EIL mercury - mercury(1) sulphate reference electrode; a magnetic stirrer; and a Cahn 21 automatic electrobalance THORBURN BURNS AND MAITIN 453 Reagents and Solutions Sodium chloride 90.99*yo.Potassium bromide 99.9*yo. Potassium iodide 99.75*y0. Silver nitrate solutiopz 0.2 M. Cellulose powder. Paperfor m e as samele carrier. Nitric acid 4 M. Aluminium jhotassiztm sulfihate solution 0.05 M. Ammonium vanadate solution 1% m/V. Bromophenol blue indicator solution in ethanol 0.05% mlV. BDH Chemicals Ltd. Hydraxine hydrate 99-100yo. BDH Chemicals Ltd. Doubly distilled water was used throughout the work for making up the solutions dilutions and rinsing of glassware etc. All items of calibrated glassware used were certified “A” grade.Procedure Accurately weigh by difference an amount of the sample to contain 0.04 & 0.002 mmol of halide ion in the sample mass. Place it on a filter-paper cut into an L shape (3 x 3 cm with a 1 x 3 cm tail at one edge). Add 30-40 mg of dry cellulose powder and make a parcel (1 x 1 cm) by folding the paper containing the sample in such a way so that the sample is covered with several layers of paper in all directions. Place an additional strip of paper (1 x 6 cm) in between the folds fold it once over the parcel and then wrap the parcel completely with the tail of the original paper. The unfolded portion of the additional strip is used as a fuse for ignition of the parcel. Clamp the parcel into a platinum gauze (1 x 1.5 cm) attached to a thin tungsten rod sealed into a ground-glass stopper (B24/29).Ignite the fuse and combust the sample in a 500-ml Erlenmeyer flask (neck B24/29) that has been flushed previously with oxygen and contains 10 ml of water and 5 drops of hydrazine hydrate. After the combustion is complete shake the flask gently on a mechanical shaker for 15 min and then allow it to stand for another 15 min. Carefully transfer the contents into a 100-ml beaker. Wash the stopper platinum gauze and walls of the flask into the beaker using small portions of water taking care to ensure that the total volume does not exceed 60 & 5 ml. Whilst stirring add 3 drops of bromophenol blue indicator followed by addition of 4 M nitric acid dropwise until the yellow colour of the indicator is just restored then add a few drops in excess.Add 4 drops of ammonium vanadate solution heat the solution gently on a hot-plate for a few minutes until it almost reaches boiling-point remove and stir for 30 s. Add 2 ml of aluminium potassium sulphate solution and dilute to 75 ml. Place the beaker in a brown-glass container and titrate potentiometrically with 0.2 M silver nitrate solution as follows. Add 0.05-ml aliquots at the beginning of the titration and 0.01-ml aliquots in the vicinity of the equivalence point in 8-10 steps. Note the potential readings, allowing a 1-min waiting time after each addition of the titrant. Determine the equivalence points by Gran’s method using data after the equivalence point. Standardise the silver nitrate solution by titrating a set of solutions containing standard inorganic halides equivalent to the halide content to be determined in the sample (i-e.about 0.04 mmol). Titrate another set of solutions containing known amounts of inorganic halides as above in the solutions obtained after combustion of a similar amount of paper cellulose powder and all other reagents used in the combustion of organic samples. From the titre values obtained in these titrations and those obtained in the standardisations above calculate the blank values. Separate standardisation and blank determinations are required for each halide. Calculate the percentage of halogen(s) present in the samples after deducting the blank values. AnalaR BDH Chemicals Ltd. AnalaR BDH Chemicals Ltd. AnalaR BDH Chemicals Ltd.AnalaR BDH Chemicals Ltd. Whatman CCL41 dried at 105 “C. AnalaR Hopkin and Williams. Cut from Whatman No. 1 filter-paper. AnalaR Hopkin and Williams. May and Baker. Dissolved using a few millilitres of 2.5 M sulphuric acid. Place a magnetic-stirring bar in the beaker. Allow to cool to room temperature. The reagent blanks are determined as follows. Results and Discussion In order to achieve rapid and complete combustion of the halogenated organic compounds, sucrose dodecanol glucose starch and cellulose powder were examined as combustion aids. * Percentage purity from gravimetric assay 454 THORBURN BURNS AND MAITIN IMPROVEMENTS TO OXYGEN Anahst VOJ. 108 Cellulose powder was found to be the best for complete and uniform burning of the samples. Sucrose and glucose "sparked" whilst burning and loss of sample was possible.Starch burned very slowly and often gave black residues. Dodecanol combusted quickly and the paper was burnt before the complete combustion of the sample which then caused soot formation and only partial combustion of the sample. Because of high vapour pressure many halogenated materials tend to distil out of the flame leading to low results owing to partial combustion. Partial combustion and soot formation also resulted when samples were not carefully and tightly wrapped. To avoid such problems the sample wrapping procedure as suggested by Awad et a1.12 was adopted. The size of the platinum gauze was also found to be critical; it should be small enough so that it remains entirely inside the flame during the burning process otherwise soot formation and distillation are unavoidable.After studies of the combustion of the compounds attention was given to the absorption stage of the method. Neutral hydrogen peroxide was found to give satisfactory results for compounds containing chlorine and bromine. Iodine was however absorbed and reduced in alkaline peroxide but as noted earlier by Lalancette et al.,13 the absorbing solutions occasionally remained yellow indicating the presence of free iodine. In addition hydrogen peroxide must be completely eliminated before titration of iodide. The decomposition of hydrogen peroxide is difficult free iodine was found to appear on acidification even after boiling the solution after combustion for several minutes. Similar problems arose when alkaline sodium tetrahydro-borate(II1) was used as reducing agent.Alkaline hydrazine sulphate has been considered a better reducing agent and used by many workers to give satisfactory results for chloride, bromide and iodide. In this study hydrazine hydrate has been shown to be more convenient than hydrazine sulphate alkaline conditions were not required as had been suggested by Mazzeo-Farina and Mazzeo. Further it was found that removal of any excess of hydrazine was not necessary provided the solution is kept acidic during the titration. Earlier workers have removed excess of hydrazine by using hydrogen peroxide and then eliminated the excess of hydrogen peroxide by boiling. Although details were not given it would appear that TABLE I ASSAY OF ORGANIC CHLORINE COMPOUNDS Compound l-Chloro-2,4-dinitrobenzene* C1C6H,(N0,), S-Benzylthiuronium chloride C6H,CH,SC( NH)NH,Clt p-Chlorobenzoic acid .. ClC,H,COOH Hexachlorobenzenes . . c6c16 Sample mass/mg . . 8.621 8.607 8.610 8.617 8.593 . . 8.616 8.647 8.649 8.645 8.647 8.639 . . 6.927 6.907 6.888 6.902 6.907 6.903 . . 2.089 2.092 2.093 2.083 Calculated yo Found yo Mean % 17.50 17.54 17.55 17.59 17.52 17.56 17.54 17.49 17.54 17.50 17.53 17.50 17.54 17.42 17.45 22.65 22.61 22.60 22.56 22.55 22.64 22.63 22.59 74.70 74.39 74.32 (74.28)y 74.25 74.15 74.50 Standard Apparent deviation, +0.05 f0.03 error yo % +O.Ol f0.05 -0.05 f0.04 -0.38 A0.15 (+ 0.04) TI * Microanalytical-reagent grade ; relative molecular mass 202.56 ; and melting-point 50-51 "C.BDH t Microanalytical-reagent grade; relative molecular mass 202.72 ; and melting-point 176-177 "C. Microanalytical-reagent grade; relative molecular mass 156.57 ; and melting-point 240-241 "C. 5 Microanalytical-reagent grade purity expected Q 99.5% ; relative molecular mass 284.8 1 ; and 7 The values in parentheses are based on the purity found by GLC. Chemicals Ltd. BDH Chemicals Ltd. BDH Chemicals Ltd. melting-point 228-230 "C. BDH Chemicals Ltd. Purity found by GLC 99.44% April 1983 FLASK COMBUSTION FOR HALOGENATED ORGANICS 455 Childs et al.15 have also titrated directly without the removal of the excess of hydrazine. Recently Chengl* also titrated the final solution directly.Many workers7J4~24-27 have indicated that bromate or iodate may be formed after the com-bustion. In order to overcome this problem a few drops of ammonium vanadate were added to the acidified solution after combustion and the solution was heated for a few minutes. This procedure catalyses the conversion of any bromate or iodate if formed after combustion into bromide or iodide in the presence of hydrazine hydrate. For the analysis of compounds containing chlorine alone such treatment is not required but as it did not show any adverse effects it was also followed for compounds containing chlorine to maintain a single procedure for all sample types. During studies of the potentiometric titration of halides using silver nitrate solution a con-centrated titrant was found to give sharp potential breaks therefore avoiding dilution during the titration.28 The use of a concentrated titrant also improved the stability of the potential responses.A calibrated microburette was used for accurate and precise delivery of the small volumes of the titrant. It was also found advisable to use similar amounts of halides in samples and standards. The titrations were monitored using a silver billet indicator electrode and a mercury - mercury(1) sulphate reference electrode. In a study of electrode response i t was noticed that it was necessary to allow 1 min after each addition of the titrant in order to obtain reproducible potential readings. Aluminium potassium sulphate was added to avoid the problems that may occur due to ad~orption.~~ Equivalence points were evaluated by Gran's method using 8-10 points after the equivalence point.These results were found to be better than those calculated by other graphical or numerical methods. The paper used as sample carrier was found to give a blank value. The blank values were different for each halide and varied with the different methods of equivalence point evaluations examined. It is essential to calculate the blank separately for each halide titration. It was found that the blanks were more reproducible when measured by difference as in the procedure herein than when measured directly. Several standard compounds and other organic compounds containing single halogens were assayed by using the final procedure described above and results obtained are given in Tables 1-111.It is seen that these are satisfactory except for hexachlorobenzene which was exam-ined by gas - liquid chromatography but when the results are corrected for over-all purity they are then in good agreement. Compound Bromobenzoic acid* BrC,H,COOH N-Bromosuccinimidet C,H,O,NBr Bromoacetanilide: CH,CONHC,H,Br TABLE I1 ASSAY OF ORGANIC BROMINE COMPOUNDS Standard Sample Calculated Apparent deviation, masslmg yo Found yo Mean yo error % % . . 8.552 39.75 39.89 39.82 +0.07 f0.09 8.522 39.73 8.567 39.71 8.541 39.87 8.532 39.78 8.525 39.93 7.538 45.03 7.591 45.14 7.596 44.87 7.568 44.83 . . 7.601 44.89 45.16 45.01 +0.12 f0.15 . . 9.068 37.33 37.23 37.35 +0.02 fO.08 9.135 37.37 9.111 37.38 9.121 37.32 9.115 37.43 * Microanalytical-reagent grade; relative molecular mass 201.02 ; and melting-point 256-257 "C.t Relative molecular mass 177.99 melting-point 180 O C ; Aldrich Chemical Co. 99% material recrystal-$ Relative molecular mass 214.07 ; melting-point 169 "C; BDH Chemicals Ltd. 99-101% material BDH Chemicals Ltd. lised five times from benzene. recrystallised five times from ethanol 456 THORBURN BURNS AND MAITIN TABLE I11 ASSAY OF ORGANIC IODINE COMPOUNDS Compound p-Iodonitrobenzene* C6H4N021 N-Iodosuccinimidet CH,CONICOCH, o-Iodobenzoic acid$ IC,H,COOH Sample masslmg . . . . 10.983 10.982 10.951 10.982 10.987 . . 9.930 9.930 9.930 9.912 9.914 9.924 . . 10.953 10.934 10.924 10.931 10.942 Standard Calculated Apparent deviation, % Found % Mean yo error % % 50.96 50.92 50.86 -0.10 f0.20 50.88 51.11 50.83 50.55 56.19 56.19 56.52 56.64 56.10 51.40 51.15 51.37 51.18 56.40 56.42 56.34 -0.06 f0.22 51.16 51.48 51.32 +0.16 f0.14 * Relative molecular mass 249.01 ; melting-point 172-173 “C ; Aldrich material recrystallised five times t Relative molecular mass 224.99 g ; melting-point 203-204 “C; Hopkin and Williams Ltd.laboratory-$ Microanalytical-reagent grade; relative molecular mass 248.03 ; and melting-point 162-163 “C. from ethanol. reagent grade material recrystallised five times from dioxan - carbon tetrachloride. BDH Chemicals Ltd. 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. References Hempel W. 2. Angew Chem. 1892 13 393. Macdonald A. M. G. Analyst 1961 86 3. Macdonald A. M. G. in Reilley C. N. Editor “Advances in Analytical Chemistry and Instru-Schoniger W. Mikrochim. Acta 1955 123. Schoniger W. Mikrochim. Acta 1955 869. Patai S. “The Chemistry of Carbon - Halogen Bond Part 1,” John Wiley New York 1973. MAzor L. “Analytical Chemistry of Organic Halogen Compounds,” Pergamon Press Oxford 1975. Ma T. S. and Rittner R. C. “Modern Organic Elemental Analysis,” Marcel1 Dekker New York 1979. Ingram G. “Methods of Organic Elemental Microanalysis,” Chapman and Hall London 1962. Levy R. Pure APPl. Chem. 1972 29 629. Griepink B. and van Sandwijk A. Mikrochim. Acta 1969 1014. Awad W. I.Gawargious Y. A. Hassan S. S. M. and Milad N. E. Anal. Chim. Acta 1966 36, Lalancette R. A. Lukaszewski D. M. and Steyermark A. Microchem. J. 1972 17 665. Celon E. Mikrochim. Acta 1969 592. Childs C E. Meyers E. E. Cheng J. Laframboise E. and Balodis R. B. Microchem. J. 1963 7, Nara A. Kobayashi N. and Honba K. Microchem. J. 1975 20 200. Rittner R. C. and Ma T. S. Mikrochim. Acta 1976 243. Cheng F. W. Microchem. J. 1980 25 86. Campiglio A. and Traverso G. Mikrochim. Acta 1980 I 485. Lalancette R. A. Steyermark A. Lukaszewski D. M. and Kostrazewski P. L. J . Assoc. Ofl. Lalancette R. A. and Steyermark A. J . Assoc. Of. Anal. Chem. 1974 57 26. Mazzeo-Farina A. and Mazzeo P. Microchem. J . 1978 23 137. Gran G. Analyst 1952 77 661. Dixon J. P. “Modern Methods in Organic Microanalysis,” Van Nostrand New York 1968. Olson E. C. in Kolthoff J. M. and Elving P. J. Editors “Treatise on Analytical Chemistry Part MAzor L. PApay K. M. and Klatsmanyi P. Talanta 1963 10 557. Belcher R. and Fildes J. E. Anal. Chim. Ada 1961 25 34. Thorburn Burns D. and Maitin B. K. Anal. Proc. 1982 19 324. Lingane J. J . “Electroanalytical Chemistry,” Second Edition Interscience New York 1958. mentation,” Volume 4 John Wiley New York 1965 p. 75. 339. 266. Anal. Chem. 1973 56 888. 11,” Volume 14 John Wiley New York 1971 p. 1. Received August 6th 1982 Accepted November l l t h 198

ISSN:0003-2654    

DOI:10.1039/AN9830800452

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Analyst April 1983 Vol. 108 pp. 457-463 457 Evaluation of Equivalence Points in the Potentiometric Titration of Mixtures of Halides Duncan Thorburn Burns Binod K. Maitin and Gyula Svehla Department of Analytical Chemistry Queen’s University of Belfast Belfast BT9 5AG The accurate and precise location of the equivalence points in the potentio-metric titration of halide ions can be achieved numerically by using Gran’s method. The results have been shown to be better by this method than those obtained on the same titration data by the differential methods of Kolthoff of Fortuin and of Hahn. Keywords ; Equivalence points ; potentiowaetric titration ; halide ions ; Gran’s method Simple graphical methods for the location of equivalence points are tedious and prone to sub-jective errors.Methods involving mathematical calculations of the equivalence points with-out plotting the titration curves are relatively rapid and convenient to make using a pro-grammable calculator. In addition they yield results with improved accuracy and precision. One of the most useful numerical methods for end-point location is that due to Gran,l which uses functions that linearise sigmoidal potentiometric titration curves. For each type of titration two functions were derived one for the region before and the other for the region after the equivalence point. For example for argentimetric titrations using a silver indicator electrode the appropriate form of the Gran functions F and FA for the regions before and after the equivalence points respectively are and where Vo is the initial volume of the solution being titrated VT is the volume of titrant added and E is the measured potential (volts).Each function is linearly dependent on the volume of the titrant and both become zero at the equivalence point. Various modifications have been Anfalt and Jagnerg have given estimates of the precision and accuracy expected from a variety of numerical and graphical procedures to evaluate equivalence points from potentio-metric titration data. Gran’s method and the multi-parameter method are expected to give the least systematic errors and highest precisions. Because the expected precisions attainable were similar the comparatively simpler Gran’s method was chosen and has been compared with alternative methods involving simple calculations such as those of Kolthoff and co-workers,1°-12 Fortuin13 and Hahn.14 These latter three methods make use of the three or four largest potential steps obtained in the equivalence-point region resulting from the addition of equal volumes of titrant.Each method is based on the assumption that the stoicheiometric equivalence point and the inflection point of the titration curve are coincident. which all involve computer-based calculations. Experimental Mixtures containing 0.04 mmol of each halide ion (for binary) and 0.03 mmol of each halide ion (for ternary mixtures) were titrated with 0.2 M silver nitrate solution using a silver billet indicator and a mercury - mercury(1) sulphate reference electrode as described e1~ewhere.l~ An AGLA micrometer syringe burette and an Orion 901 Microprocessor Ionalyser were used for measurement of potentials.A Metrohm Titroprocessor 636 Autotitrator 635 was also used in the fixed volume addition and fixed waiting time modes; the volume - potential data for the entire titration were printed out and used for equivalence-point calculations.16 Results and Discussion The first series of results concerned titrations of single halides. The silver nitrate titrant was Standardisations and standardised using pure inorganic halides (re-assayed gravimetrically) 458 THORBURN BURNS et al. EVALUATION OF EQUJVALENCE Analyst VoZ. 108 blank determinations from the paper carrier and reagents were carried out separately for each halide as described elsewhere.16 Each method of calculation of the equivalence point was applied systematically to each set of titrations for the standardisation the blank evaluation and for the assay.Several organic compounds containing chlorine bromine or iodine were analysed by poten-tiometric titration with silver nitrate solution after oxygen flask combustion. The results calculated by each method are summarised in Tables 1-111. Of the three differential methods (Kolthoff’s Fortuin’s and Hahn’s methods) only Fortuin’s method gave good results for all three halides. Kolthoff’s method was satisfactory for chloride and bromide but not for iodide. Hahn’s method gave similar results to that of Kolthoff except for the assay of hexachlorobenzene where precision was affected by a single low result. The increase in apparent errors must arise from the method of data treatment because the compounds assayed were pure or reference materials.This view is in accord with that of Ebel TABLE I COMPARISON OF RESULTS OBTAINED BY DIFFERENT METHODS OF EQUIVALENCE-POINT EVALUATION FOR CHLORIDE Chloride found yo Chloride ReDlica- I L \ Compound l-Chloro-2,4-dinitro-benzene-Mean . . Standard deviation . . chloride*-Apparent error . . S-Benzylthiuronium Mean . . Standard deviation . Apparent error . . p-Chlorobenzoic acid*-Mean . . Apparent error . . Standard deviation . . (pufity expected 4 99.5%, purity by GLC 99.44%) Hexachlorobenzene*-Mean . . * . Apparent error . Standard deviation . . calculated tibn % No. Kolthoff 17.50 1 17.’53 2 17.57 3 17.65 4 17.53 5 17.57 .. 17.57 . . +0.07 . . f0.05 17.49 1 17.63 2 17.48 3 17.42 4 17.48 5 17.39 6 17.38 . . 17.46 ,. . . -0.03 . . f0.09 22.65 1 22.59 2 22.63 3 22.59 4 22.76 5 22.70 6 22.64 . . 22.64 . . -0.01 . . f0.08 74.70 1 74.47 (74.28)t 2 74.37 3 74.57 4 74.45 . . 74.47 * . . . -0.23 * . . . f0.08 (+0.19) t Fortuin 17.55 17.61 17.66 17.56 17.51 17.58 + 0.08 f 0.06 17.64 17.52 17.48 17.50 17.44 17.43 17.50 +0.01 fO.08 22.65 22.64 22.60 22.82 22.74 22.55 22.67 + 0.02 fO.10 74.52 74.41 74.72 74.40 74.51 - 0.19 (+0.23)t f0.15 Hahn 17.52 17.59 17.54 17.48 17.53 17.53 + 0.03 f 0.04 17.44 17.42 17.46 17.51 17.46 17.41 17.45 f 0.04 22.62 22.69 22.69 22.77 22.69 22.52 22.66 +0.01 f0.08 - 0.04 74.82 74.71 74.67 72.99 74.30 - 0.40 (+0.02)t f0.87 Gran before 17.62 17.70 17.68 17.62 17.61 17.65 + 0.15 f0.04 17.47 17.48 17.61 17.55 17.59 17.49 17.53 + 0.04 f 0.06 22.60 22.70 22.70 22.72 22.69 22.63 22.67 + 0.02 f0.05 74.68 74.77 74.86 74.76 74.77 +0.69 (+ 0.69) t f 0.07 Gran after 17.54 17.69 17.62 17.66 17.54 17.55 + 0.06 f 0.03 17.54 17.53 17.50 17.54 17.42 17.45 17.50 + 0.01 k0.05 22.61 22.56 22.55 22.64 22.63 22.59 22.60 - 0.05 f 0.04 74.39 74.25 74.15 74.50 74.32 -0.38 (+ 0.09) t f 0.15 * Microanalytical-reagent grade BDH Chemicals Ltd.Poole Dorset. t The values in parentheses are calculated on the basis of results obtained by gas - liquid chromatographic analysis of the sample; the errors calculated on the basis of these results are shown in parentheses April 1983 POINTS I N POTENTIOMETRIC TITRATION OF HALIDES TABLE I1 COMPARISON OF RESULTS OBTAINED BY DIFFERENT METHODS OF EQUIVALENCE-POINT EVALUATION FOR BROMIDE 459 Compound Bromobenzoic acid*-Mean . . Apparent error . . Standard deviation . . N-Bromosuccinimide t-Mean . . ,. Apparent error . . Standard deviation . . Bromoacetanilide$-Mean . . Apparent error . . Standard deviation . . . . . . Bromide found % Bromide P A 1\ calculated Replication % No. 39.75 1 2 3 4 5 6 7 .. . . 44.89 1 2 3 4 5 6 . . 9 . 37.33 1 2 3 4 5 Kolthoff 39.79 39.80 39.84 39.80 39.72 39.66 39.65 . . 39.74 . . -0.01 . . f0.08 44.88 44.71 44.57 44.77 44.83 44.76 . . 44.75 . . -0.14 . . fO.11 37.31 37.30 37.26 37.28 37.28 . . 37.29 . . -0.04 . . f0.02 Fortuin 39.91 39.97 40.13 39.84 39.88 39.70 39.64 39.87 +0.12 f0.16 44.97 44.52 44.35 44.64 44.70 44.53 44.62 - 0.27 f0.21 37.36 37.27 37.33 37.24 37.18 37.28 f0.07 -0.05 Hahn 39.87 39.96 40.07 39.80 39.83 39.66 39.60 39.83 + 0.08 f0.16 44.93 44.75 44.70 45.00 44.76 44.81 44.83 - 0.06 f0.12 37.32 37.24 37.27 37.31 37.34 37.30 f0.04 - 0.03 Gran before 39.93 40.20 39.98 39.80 39.94 39.47 39.65 39.84 + 0.09 f0.25 44.89 44.42 44.41 44.70 44.74 44.60 44.63 -0.26 kO.19 37.37 37.31 37.33 37.23 37.13 37.27 - 0.06 hO.10 Gran after 40.03 40.09 40.14 39.86 39.80 39.67 39.67 39.89 +0.14 f0.19 44.90 44.67 44.38 44.6 1 44.80 44.73 44.68 -0.21 k0.18 37.35 37.31 37.32 37.26 37.14 37.27 - 0.06 f 0.08 * Microanalytical-reagent grade BDH Chemicals Ltd.t Relative molecular mass 177.99; melting-point 180 “C. Aldrich (99%) material recrystallised five $ Relative molecular mass 214.07 ; melting-point 169 “C. BDH Chemicals Ltd. (99-101%) material times from benzene. recrystallised five times from ethanol.and Seuringl’ who refer to these differential methods as approximate and suggest that the errors arise because only a few results near the equivalence points are used. Potential readings near the equivalence points are subject to “potential transfer errors,” which arise from slow attainment of solution equilibria and of the electrode potential and from time constants of the measuring circuits. Here it was noted that “potential transfer errors” were more frequent in the titration of iodide probably owing to the very abrupt potential change near the equivalence point. In an attempt to improve the precision and accuracy of the results Gran’s method was then applied to data before and data after the equivalence point using a programmable calculator (Texas TI59) with a printout.A program was devised to calculate the Gran’s functions and the line of linear regression with volume for a variable number of data points using a least-squares fit. After entering the initial volume of titrand (V,,) for a calculation based on data prior to the equivalence point the next stage was to enter data potential (E in volts and volume of titrant VT in millilitres) for 8-10 points considered to be prior to the equivalence point. The data were treated system-atically in order to calculate the equivalence point from the points diagnosed as being before the equivalence point. The equivalence point (ie. the volume where Gran’s function = 0) and the correlation coefficient were calculated from the first three Such errors occurred only occasionally with chloride and bromide.The procedure was as follows 460 THORBURN BURNS et U l . EVALUATION O F EQUIVALENCE AfiUhySt VOJ. 108 data points and printed out. The next data point was then included and the equivalence point and Correlation coefficient re-calculated from the new regression line using all of the four points and so on until all the data points were incorporated into the series of least-square fits. The equivalence point was normally determined using the condition that the equivalence volume must be greater than the volume co-ordinate of the last acceptable data point. In addition to this criterion the values of the correlation coefficients for each calculation printout were inspected. Normally the correlation remains close to unity for the points prior to the equivalence point and decreases when points beyond the equivalence points are included.Occasionally the correlation coefficient decreased for the last data point accepted as prior to the equivalence point owing to “potential transfer errors” as described earlier in such instances this point was rejected and the equivalence point calculated from the previous point. TABLE I11 COMPARISON OF RESULTS OBTAINED BY DIFFERENT METHODS OF EQUIVALENCE-POINT EVALUATION FOR IODIDE Compound P-Iodonitrobenzene*-Mean . . Apparent error . . Standard deviation . . N-Iodosuccinimidet -Mean . . Apparent error . . Standard deviation . . o-Iodobenzoic acid:-Mean . . Apparent error . . Standard deviation . . Iodide calculated, YO 50.96 56.40 .. 51.16 f . Replication No. 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 . . Iodide found yo f h \ Kolthoff 50.54 50.88 50.74 50.66 50.62 . . 50.69 . . -0.27 . . f0.13 56.42 56.88 56.70 56.93 57.03 56.79 . . 56.79 . . +0.39 . . f0.21 51.80 51.87 51.85 51.84 51.67 . . 51.81 . . +0.65 . . fO.08 Gran Gran Fortuin Hahn before after 50.92 50.39 50.77 50.92 50.95 51.14 51.39 60.88 51.07 50.49 50.80 61.11 50.95 50.32 51.23 50.83 50.93 50.46 51.34 50.66 50.96 50.56 51.11 50.86 f0.06 f0.33 f0.30 k0.20 56.37 56.51 56.81 56.42 56.45 56.69 56.81 56.19 56.42 66.58 57.12 56.19 56.58 56.79 56.94 56.52 56.62 56.93 57.06 56.64 56.46 56.57 57.03 56.10 0.00 -0.4 +0.15 -0.10 56.48 56.68 56.96 56.34 hO.10 f0.16 f0.13 f0.22 51.41 51.84 52.02 51.48 51.45 51.88 51.81 51.40 51.38 51.55 51.95 51.15 51.42 51.82 51.87 51.37 51.25 51.63 51.91 51.18 +0.08 +0.28 +0.56 -0.06 51.38 51.74 51.91 51.32 +0.22 +0.58 +0.75 tO.16 kO.08 h0.15 fO.08 f0.14 * Relative molecular mass 249.01 ; melting-point 172-173 “C ; Aldrich material recrystallised five times t Relative molecular mass 224.99; melting-point 203-204 “C.Hopkin and Williams laboratory-from ethanol. reagent grade material recrystallised five times from dioxan - tetrachloromethane. Microanalytical-reagent grade BDH Chemicals Ltd. For calculations based upon data after the equivalence point any pre-equivalence-point data are rejected on the basis that the volume co-ordinate of the point entered must not be less than the equivalence point calculated; before it was occasionally necessary to reject a point on the grounds of a reduced correlation coefficient.Fig. 1 shows the computer printout of such an evaluation process. The left-hand column contains the data as they were fed in; Vo is the initial volume of the sample in the cell 75.00 cm3; V the volume of the titrant added in the first instance 0.15 cm3; E is the corresponding electrode potential -0.32 V and S is th April 1983 POINTS I N POTENTIOMETRIC TITRATION OF HALIDES 461 counter for the data pairs. In the right-hand column the first figure is the intercept the second the slope of the Gran function versm titrant volume plot calculated by linear-regression analysis followed by the value of the equivalence point (V,) and the coefficient of correlation [(R) in this instance always negative as we have a negative slope] and finally the counter N is again printed out.From the values it is easy to select V = 0.2064 cm3 as the final result. FLP-?& 75. 00 a. 15 -0.321 I . 0. ! 6 -0. 116 2 0. !7 -0. ?!U 3. U. 13 -0. 301 4. 0. 19 -0. 269 5. 0. 20 -0. 266 6 . 0.21 -0. a37 7. 0. 22 -0.230 a. REF 1. Printout for Fig. point evaluation. VE = 0.206 4 ml. 'dE R N 'd E R N VE R N 'VE R ti YE R N VE R N end-The results for these calculations are shown in Tables 1-111. For chloride and bromide both data before and after the equivalence point yielded satisfactory results.Although differences in precision and accuracy obtained by using both sets of data were small the results using data after the equivalence point appeared to be better. For hexachlorobenzene both precision and accuracy were relatively poor. Examination of a sample of the hexachlorobenzene by gas chromatography using an integrator revealed that the compound was slightly impure (99.44%) and the result obtained by data after the equivalence point confirmed this view. The results for iodide using data after the equivalence points were distinctly better than those obtained using data prior to the equivalence point. Further results by Gran's method were better than those obtained by the differential methods except for that of Fortuin which gave comparable accuracy and precision.The second series of experiments concerned the analysis of organic compounds containing bromide and chloride. Standardisations were carried out using mixtures of individually weighed pure organic compounds containing single halides. The blank determinations were carried out using the procedure as described before.15 Each method of calculation was again applied systematically to each set of titrations for standardisation blank determination and for the assays. The data obtained using Gran's method for data prior to the equivalence point were clearly superior to the differential methods including that of Fortuin which was previously found to be satisfactory for the single halide titrations. The use of the before equivalence-point data is also supported by the correlation-coefficient data.It was found that for bromide the correlation coefficients are close to unity only for data points prior to equivalence points while for chloride (second equivalence point) there was no significant difference between correlation results using before and after the equivalence-point data. As the chloride equivalence volume is determined by the difference of the second and first equivalence points the same calculation criterion was used for both the equivalence points and values closer to the true values were obtained. These results show the importance of the method of data examination and that for single halide determinations satisfactory results can be obtained using either Fortuin or Gran's method but for mixtures of chloride and bromide Gran's method is superior (e.g.see Table IV). The results are summarised in Table IV 462 THORBURN BURNS et al. EVALUATION OF EQUIVALENCE Artalyst VoZ. 108 TABLE IV COMPARISON OF RESULTS OBTAINED BY DIFFERENT METHODS OF EQUIVALENCE-POINT EVALUATION Bromide found yo Chloride found % L I A \ I i Replica- Kolt- Gran Gran Kolt- Gran Gran Compound tion No. hoff Fortuin Hahn before after hoff Fortuin Hahn before after 1- [l-( 4-Bromophenylmethy1)-4-~i~eridinvll-~-chloro-2-~ tri-fl;okmeth$lj-l If-benzimi‘dazole (CaoHmBslFaNtJ*- 1 17.14 17.06 17.34 17.01 17.13 7.42 7.47 7.37 7.48 7.47 2 17.12 17.02 17.38 17.07 17.09 7.41 7.45 7.36 7.45 7.36 - . . - - - . - - 4 16.90 16.90 K s i i637 16.84 7.49 7.49 7.58 7.48 7.59 5 17.11 17.02 17.29 17.07 17.18 7.45 7.50 7.41 7.47 7.47 6 16.74 16.79 16.86 16.92 16.81 7.54 7.50 7.38 7.47 7.47 7 16.90 16.90 16.83 16.94 17.10 7.48 7.48 7.41 7.48 7.35 8 17.12 17.03 17.19 17.05 17.28 7.55 7.58 7.46 7.59 7.47 9 16.96 16.94 16.74 16.90 16.86 7.47 7.49 7.62 7.49 7.68 10 17.02 16.99 17.38 16.94 17.10 7.48 7.49 7.34 7.48 7.48 Mean .. . . . . . . . . 17.01 16.97 17.12 16.99 17.05 7.47 7.49 7.43 7.49 7.47 Standarddeviatiod‘ . . . . . f0.13 h0.08 f0.27 &-0.06 50.16 f0.05 f0.03 f0.10 f0.04 *0.08 Apparent error +0.11 +0.07 +0.22 $0.09 +0.15 -0.03 -0.01 -0.07 -0.01 -0.03 N-( 4-Bromophenyl)-N‘-( 2-chloro-1 20.64 20.68 20.50 20.69 20.66 9.24 2 20.37 20.40 20.59 20.57 20.48 9.32 3 20.40 20.46 20.59 20.61 20.50 9.26 4 20.49 20.56 20.58 20.64 20.92 9.20 5 20.35 20.40 20.54 20.58 20.47 9.29 6 20.36 20.39 20.65 20.55 20.44 9.27 9 20.38 20.43 20.57 20.65 20.47 9.38 8 20.39 20.44 20.69 20.58 20.43 9.26 9 20.40 20.46 20.53 20.59 20.38 9.23 Mean .. . . 20.42 20.47 20.57 20.61 20.53 9.27 Standard deviatioi’ . . . . . . *0.09 fO.09 f0.04 f0.04 f0.17 k0.05 Apparent error -0.24 -0.19 -0.09 -0.05 -0.13 +0.10 9.24 9.30 9.21 9.16 9.26 9.26 9.37 9.22 9.21 9.25 k0.06 - 0.08 9.34 9.24 9.27 9.10 9.22 9.18 9.08 9.20 9.15 9.04 9.14 8.99 9.29 9.24 9.15 9.05 9.21 9.16 9.26 9.17 9.16 9.08 9.24 9.19 9.22 9.17 9.29 9.16 9.20 9.17 +0.01 +0.03 0.00 fO.11 f0.04 kO.09 a-Bromo-fi-chloroacetophenone (ClC,H,COCH,Br) 3- 1 34.65 34.43 34.72 34.50 34.64 15.00 15.13 15.07 15.03 15.16 2 34.03 34.18 34.30 34.21 34.20 15.37 15.28 15.03 15.38 15.13 3 34.44 34.34 35.23 34.31 34.61 15.14 15.21 14.87 15.22 15.13 4 34.42 34.30 35.35 34.33 33.77 15.04 15.11 14.78 15.13 15.42 5 33.83 33.94 34.08 34.25 34.08 15.34 15.27 15.03 15.19 15.15 6 34.74 33.50 34.82 34.38 34.80 15.11 15.22 15.08 15.24 15.15 7 33.92 34.02 34.21 34.17 34.03 15.45 15.40 15.46 15.39 15.24 8 33.66 33.75 33.96 33.91 33.86 15.30 15.25 15.02 15.31 15.10 9 33.82 33.93 34.05 34.07 34.12 15.28 15.23 15.32 15.27 15.05 Mean .. . 34.17 34.04 34.52 34.24 34.23 15.23 15.23 15.07 15.24 15.17 Standard deviation’ . . f0.40 f0.30 50.52 h0.17 f0.36 f0.16 f0.09 k0.21 k0.12 kO.11 Apparent error -0.05 -0.18 +0.30 +0.02 +0.01 +0.05 $0.05 -0.11 +0.06 -0.01 * BCR CRM 073. Calculated bromide l6.90% chloride 7.50%.t BCR CRM 071. Calculated bromide 20.66% chloride 9.17%. 3 Calculated bromide 34.22y0 chloride 15.18%. Aldrich 98% material recrystallised five times from ethanol. The reasons for the over-all superiority of the Gran-functions procedure to evaluate the equivalence point over that of Fortuin are two-fold and arise from a consideration of the propagation of errors and the reliability of the data used in each example. Gran’s method is based directly on potential measurements at discrete values of volumes added ; Fortuin’s method relies on the ratio of difference between three successive potentials measured in the vicinity of the equivalence point after addition of equal volume increments of the titrant which compounds any random potential errors four-fold. A further advantage of Gran’s method is that each point is independent of each other.Gran’s method is applied to data away from the equivalence-point region where “potential transfer errors” can arise whilst Fortuin’s method requires data points that are necessarily from the equivalence-point region. Both methods rely on achieving a Nernstian electrode response which with mixed halides only applies to data well before the equivalence point due to co-precipitation problems. Hence Gran’s procedure remains valid for mixtures as well as for single halides. The over-all procedure described here was also applied to evaluate equivalence points in the titrations of iodide - chloride mixtures18 and of binary and ternary mixtures of halides using a microprocessor controlled aut otit rat or.1 April 1983 POINTS IN POTENTIOMETRIC TITRATION OF HALIDES References 463 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Gran G. Analyst 1952 77 661. Liteanu C. and Cormos D. Talanta 1960 7 25. Liteanu C. and Cormos D. Talanta 1960 7 32. Ingman F. and Still E. Talanta 1966 13 1431. McCallum C. and Midgley D. Anal. Chim. Actct 1973 65 155. Midgley D. and McCallum C. Talanta 1974 21 723. Magallanes J . F. and Caridi A. F. Anal. Chim. Acta 1981 133 203. Mascini M. “Ion Selective Electrode Reviews,” Volume 2 Pergamon Press Oxford 1980 p. 17. Anfalt T. and Jagner D. Anal. Chim. Acta 1971 57 165. Kolthoff I. M. and Laitinen H. A. “pH and Electrotitrations,” John Wiley New York 1944, Kolthoff I. M. and Sandell E. B. “Textbook of Quantitative Inorganic Analysis,” Macmillan, Kolthoff I. M. and Furman N. H. “Potentiometric Titrations,” Second Edition John Wiley New Fortuin J . M. H. Anal. Chim. Acta 1961 24 175. Hahn F. L. FreseniusZ. Anal. Chem. 1958 163 169. Thorburn Burns D. and Maitin B. I<. Analyst 1983 108 452. Thorburn Burns D. and Maitin B. K. J . Indian Chem. Soc. in the press. Ebel S. and Seuring A. Angew. Chem. Int. Ed. Engl. 1977 16 157. Thorburn Burns D. and Maitin B. K. unpublished data. p. 110. New York 1952 p. 488. York 1949 pp. 95-96. Received August 6th 1982 Accepted November 17th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800457

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

FOGG, CHAMS1 AND ABDALLA 465 - Delay coil (a) : Waste 4 m x Glassy Sample 0.58 mm carbon Pump injection detector Cadmium sponge column : Waste 10cm x detector 1 mm cadmium wire (c) -t- : Waste 0.58 mm carbon Pump 'ample 11.5 cm x detector injection 1-14 mm tube Fig. 1. Flow injection systctns uscd: (a) batchwise reduction niethod ; (b) continuous reduction method ; ancl (c) direct reduction mcthotl. Reagents Standard sodium nitrate solution, 1 x 10-1 M. Dissolve 2.13 g of sodium nitrate in water and dilute to 250 ml in a calibrated flask. Staizdard sodizrm ititritc solzitioit, 1 x 10-1 M. Dissolve 1.73 g of sodium nitrite in water and dilute to 250 ml in a calibrated flask. Acidic bromide eluent. Dissolve 100 g of potassium bromide in 350 ml of water, add 138 ml of concentrated hydrochloric acid, cool the solution, dilute to 500 ml and mix.Hydrochloric acid solzition, 0.3 M. Dilute 13.8 ml of concentrated hydrochloric acid to 500 ml with water. Spongy cadmiztm. Place one or two zinc rods (analytical-reagent grade, approximately 10 g total) in 250 ml of 20% m/V cadmium sulphate solution and allow to stand for 3 h at room temperature, moving the rods around periodically. Remove the rods, decant off the super- natant liquid from the deposited cadmium and wash the cadmium twice with distilled water. Cover the granules with water and macerate them for 10-15 s to produce a uniform fine size. Store under water. Prepare more dilute standard solutions by dilution. Prepare more dilute standard solutions by dilution. Preparation of the Reductor Column An Econo-Column (Bio-Rad Laboratories) was used in this work.Make a small mark on the side of the column about 0.5 cm above the spongy cadmium. Allow distilled water to flow through the column and store with the cadmium sponge under water. Before use, activate the cadmium sponge column by passing 0.3 M hydrochloric acid solution through it, and then wash with water again. Bring the water level down to the 0.5-cm mark. A suitable column for packing is about 10 cm long with an inner diameter of 7 mm. Fill the reductor column to about 7 cm with the spongy cadmium. Procedure Place a 10-ml calibrated flask below the column to collect emerging solution. Transfer slowly by pipette 1 ml of standard nitrate solution (1 x l O - S l x 10-1 M) to the top of the cadmium column, placing the tip of the pipette near the top of the cadmium bed in a manner such that the bed is not disturbed.Allow the solution to run through the column and into the466 FOGG et al. : FLOW INJECTION VOLTAMMETRY Analyst, Vol. 108 flask until the level of solution reaches the mark above the bed, then add about 2 ml of water at a time until the solution in the calibrated flask reaches the 10-ml mark. Mix, inject 25 p1 of the solution into acidic bromide eluent and note the reduction signal at the glassy carbon electrode. A 4-m delay coil and a flow-rate of 5 ml min-1 were used in this work. Determination of Nitrate using Continuous Reduction to Nitrite The reagents and column used in this study were the same as those used above for the batch- wise reduction of nitrate to nitrite.The only difference in this method was that the reductor column was connected to the sample injection valve such that solution passed through the sample loop of the flow injection valve after passing through the reductor. Sample solution was passed continuously through this system and the nitrite concentration in the sample solution was monitored at several stages by injecting the contents of the loop into the acidic bromide eluent in the flow injection voltammetric system. The system is shown in Fig. l(b). Determination of Nitrate by Direct Injection with a Flow Injection System Incorporating a Cadmium Wire Reductor Analytical-reagent grade cadmium wire of 1 mm diameter was used. In the procedure recommended, a 10-cm length of cadmium wire is inserted into an 11.5-cm length of PTFE tube (i.d.1.14 mm). This tube is then connected between the sample injection valve and the delay coil. The cadmium wire has to be reactivated as required and the frequency at which this has to be done depends on the amount of reducible material that has passed over it. The wire is taken out of the PTFE tube and washed vigorously with dilute hydrochloric acid before being replaced in the system. Aliquots (25 pl) of nitrate solution are injected directly into the system. A delay coil length of 4 m and a flow-rate of 5 ml min-l were used in this work. The flow injection system used is shown in Fig. l(c). Other reagents were prepared as described above. TABLE I BATCHWISE REDUCTION OF NITRATE : EFFECT OF LENGTH OF CADMIUM SPONGE COLUMN Nitrate and nitrite concentrations before dilution in the column = 1 x M.Length of reductor column/cm . . .. . . 2.8 7.0 12.8 Signal for nitrate/pA . . 1 . .. . . 0.91 1.31 1.13 Signal for nitrite/pA . . .. . . . . 1.87 1.68 1.43 Signal for nitrite without reductor/pA . . . . 1.87 1.87 1.87 Results In the preliminary study of the batchwise reduction of nitrate solutions the effect of the length of the spongy cadmium column was studied first (see Table I ) ; control tests were also carried out using nitrite solutions to see the effect of the reductor on nitrite. Optimum reduc- tion of nitrate at the 1 x Clearly nitrite itself is being extensively reduced on longer columns. The effect of flow-rate through the reductor column was not critical at this level of nitrate, the current signals obtained being 1.17 and 1.15 pA at 2 and 6 ml min-l, respectively. The extents of reduction at various levels of nitrate and nitrite from 1 x to 1 x 10-1 M are compared in Table 11.The current signal M level is effected with a 7-cm column. TABLE I1 BATCHWISE REDUCTION OF NITRATE : SIGNAL SIZE AT VARIOUS PRE-DILUTION CONCENTRATION LEVELS Pre-dilution concentration of nitrite or nitrate/M f 1 o 1 x 10-5 1 x 10-4 1 x 10-3 1 x 10-2 1 x 10-1 Signal for nitrate/pA . . .. . . 0,026 0.040 0.200 1.63 10.8 28.0 Signal for nitrite/pA . . . I . . 0.026 0.041 0.212 1.80 16.0 93.3 Signal for nitrite without reductor/pA. . 0.026 0.041 0.217 2.13 16.2 93.3April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 467 values given for nitrite that had not been passed through the column (but had been similarly diluted) show the near rectilinearity of the nitrite signal up to about the 1 x loA3 M nitrite level.A reduction yield of SO-90~o can be achieved for nitrate at the 1 x loA3 M level; the yield falls rapidly at higher concentrations. 1 x M and 0.05 x 10-3-1 x 10-3 M were rectilinear. Signals obtained at the latter level are shown in Fig. 2; coefficients of variation (3-10 determinations) are typically <2.5% for sequential injections in this procedure and in the other procedures. Calibration graphs in the ranges 0.05 x C i Time d Fig. 2. Batchwise reduction of nitrate: calibration signals obtained a t the pre- dilution 0-1 x 10-3 M level of nitrate. Pre-dilution nitrate concentration : A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 10-3 M.The effect of the volume of sample solution that has passed through the reductor (and the sample injection loop) in the continuous reduction method at the 1 x 10-5 M level is shown in Table 111; the values given are for single runs. A reasonably steady value is attained after about 15 ml of sample solution has passed; there is a further sliglit increase up to about 50-100 ml and a slight decrease is obtained after about 130 ml. When nitrite at tlie same level is passed tlirough the system a similar signal to that of nitrate is observed. The signals obtained when 25 ml of solution have passed at levels of nitrate from 1 x M are shown in Fig. 3. An approach to rectilinearity is apparent up to about 1 x In a preliminary study tlie use of an eluent 3.2 >I in hydrochloric acid and 20% m/V in potassium bromide, tvhich is optimum for the determination of nitrite by direct injection,l was confirmed as being optimum also when nitrate was injected directly and reduced on-line.The effect of the length of cadmium wire used is shown in Table I V ; the optimum length is 10 cm. to 1 x M. TABLE 111 CONTINUOTJS REDUCTION METHOD : SIGNAL OBTAINED AT VARIOUS SAMPLE VOLUMES PASSED THROUGH COLUMN Nitrate and nitrite concentrations = 1 x M. Volume passed/ml . . . . .. 5 15 25 35 50 100 225 Signal for nitrate/pA . . . . 0.156 0.184 0.186 0.190 0.202 0.212 0.187 Signal for nitrite/pA . . . . 0.158 0.164 0.172 0.169 0.182 0.194 0.222468 FOGG et al.: FLOW INJECTION VOLTAMMETRY Annlyst, Vol.108 Time d Fig. 3. Continuous reduction method : signals obtained a t various nitrate con- centration levels: A, 0; B, 1 x C, 1 x 10-5; D, 1 x 10-4; E, 1 x and F, 1 x lo-' M. The decrease in signal with longer lengths is probably due to increased dispersion or possibly further reduction of nitrite. The signals obtained for various levels of nitrate and nitrite injected into the recommended system are given in Table V. These are compared with the nitrite signal attained in the absence of the cadmium wire. The yield of nitrate is seen to vary from 18% at the 1 x M level. The size of the signal is increased by increasing the residence time by using a slower flow-rate; the signal varied from 0.71 pA at 1.5 ml min-1 to 0.35 pA at 8 ml min-1. Signals obtained in the range 0.1 x lOw4-l x M level to 4% at the 1 x M are shown in Fig.4. TABLE IV DIRECT INJECTION METHOD : EFFECT OF LENGTH OF CADMIUM WIRE ON SIGNAL Nitrate concentration = 1 x M. Length of wirelcm . . .. 0 2 5 10 20 40 Signal/pA . . .. . . 0.015 0.087 0.238 0.381 0.365 0.341 Discussion Nitrite sample solutions can be injected into an acidic bromide eluent in a flow injection system and determined voltammetrically at a glassy carbon electrode by the reduction signal of the nitrosyl bromide pr0duced.l In this work the extension of this method to the determina- tion of nitrate has been studied. Commonly, nitrate is reduced to nitrite and determined by visible spectrophotometry using the diazotisation properties of the nitrite. The first procedure given here simply illustrates the use of the flow injection voltammetric method previously TABLE V DIRECT INJECTION METHOD : SIGNAL SIZE AT VARIOUS COT 1' CENTRATION LEVELS Length of cadmium wire = 10 cm.Equivalent concentration of nitrate or nitrite/M f A 7 0 1 x 10-6 1 x 10-5 1 x 10-4 1 x 10-3 i x 10-2 Signal for nitrate/pA . . .. . . 0.0198 0.0198 0.0234 0.115 0.474 2.038 Signal for nitrite/pA . . .. . . 0.0198 0.0300 0.122 0.734 4.70 25.8 Reduction yield of nitrite from Signal for nitrite without wire/pA . . 0.0198 0.0300 0.126 0.833 5.55 34.7 nitrate, % . . .. .. .. - - 18 15 10 4April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 469 I , I Time Fig. 4. Direct injection method: calibration signals obtained in the range 0.1-1 x 10-4 M nitrate: A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 1 0 - 4 ~ .published in the determination of nitrate after it has been reduced first to nitrite by means of a cadmium sponge column. Clearly cadmium ions introduced into the sample during reduction do not interfere and most other reagents commonly used to reduce nitrate to nitrite would not interfere with the determination. Flow injection analysis has so far been little used for the intermittent analysis of sample streams or of large solution samples in which the determined concentration may be varying. If part of the sample stream is allowed to flow through the sample loop of the flow injection valve, the solution can be sampled and the determinand determined as required. Alterna- tively, and possibly more satisfactorily, sample can be pumped from the stream, or from a large solution sample, as required through the sample loop at specified times to allow sampling and determination.Suitable laboratory applications might include solution kinetic studies and tablet dissolution or drug availability studies. This study has suggested an application in which nitrate sample may be pumped at pre-determined times through a reductor column and sample loop, and may then be sampled and the resulting nitrite determined in the flow injection system. The preliminary work carried out so far indicates that the only difficulty would be in maintaining the level of activity of the reductor column. The introduction of an occasional reactivation cycle may solve this problem. The ideal procedure, however, would involve direct injection of the nitrate sample solution into the flow injection system.Results obtained using a cadmium wire to reduce nitrate on-line have been reported here. Yields of nitrite with this system have been low (<20%) and depend on the condition of the surface of the cadmium wire. Nevertheless, the procedure as developed so far may have applications in areas where nitrite levels need to be monitored with only low precision. By means of valves before and after the cadmium wire it may prove to be possible to regenerate the cadmium wire in situ. The results presented here indicated three possible approaches to the determination of nitrate using a flow injection voltammetric method for nitrite. Further work is planned to develop these systems for automatic use, but the present results will allow analytical chemists experienced in particular application areas to assess whether these approaches hold any advantages for them.Full experimental details have been given so that other workers can reproduce our conditions and results exactly before making any modification to suit their particular requirements. A standardisation cycle would also be needed. A.Y.C. thanks the Lebanese University for leave of absence and the Lebanese Government M.A.A. thanks the University of Khartoum for leave of absence and The authors thank Mr. I. W. Burns and Mr. G. M. for financial support. The British Council for financial support. Telling of Unilever Research for helpful discussions. References 1. 2. Fogg, A. G., Bsebsu, N.K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1462. Received November lst, 1982 Accepted December 21st, 1982April, 1983 OF NITRATE AFTER REDUCTION TO NITRITE 469 I , I Time Fig. 4. Direct injection method: calibration signals obtained in the range 0.1-1 x 10-4 M nitrate: A, 0; B, 0.2; C, 0.4; D, 0.6; E, 0.8; and F, 1.0 x 1 0 - 4 ~ . published in the determination of nitrate after it has been reduced first to nitrite by means of a cadmium sponge column. Clearly cadmium ions introduced into the sample during reduction do not interfere and most other reagents commonly used to reduce nitrate to nitrite would not interfere with the determination. Flow injection analysis has so far been little used for the intermittent analysis of sample streams or of large solution samples in which the determined concentration may be varying.If part of the sample stream is allowed to flow through the sample loop of the flow injection valve, the solution can be sampled and the determinand determined as required. Alterna- tively, and possibly more satisfactorily, sample can be pumped from the stream, or from a large solution sample, as required through the sample loop at specified times to allow sampling and determination. Suitable laboratory applications might include solution kinetic studies and tablet dissolution or drug availability studies. This study has suggested an application in which nitrate sample may be pumped at pre-determined times through a reductor column and sample loop, and may then be sampled and the resulting nitrite determined in the flow injection system.The preliminary work carried out so far indicates that the only difficulty would be in maintaining the level of activity of the reductor column. The introduction of an occasional reactivation cycle may solve this problem. The ideal procedure, however, would involve direct injection of the nitrate sample solution into the flow injection system. Results obtained using a cadmium wire to reduce nitrate on-line have been reported here. Yields of nitrite with this system have been low (<20%) and depend on the condition of the surface of the cadmium wire. Nevertheless, the procedure as developed so far may have applications in areas where nitrite levels need to be monitored with only low precision. By means of valves before and after the cadmium wire it may prove to be possible to regenerate the cadmium wire in situ. The results presented here indicated three possible approaches to the determination of nitrate using a flow injection voltammetric method for nitrite. Further work is planned to develop these systems for automatic use, but the present results will allow analytical chemists experienced in particular application areas to assess whether these approaches hold any advantages for them. Full experimental details have been given so that other workers can reproduce our conditions and results exactly before making any modification to suit their particular requirements. A standardisation cycle would also be needed. A.Y.C. thanks the Lebanese University for leave of absence and the Lebanese Government M.A.A. thanks the University of Khartoum for leave of absence and The authors thank Mr. I. W. Burns and Mr. G. M. for financial support. The British Council for financial support. Telling of Unilever Research for helpful discussions. References 1. 2. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1462. Received November lst, 1982 Accepted December 21st, 1982

ISSN:0003-2654    

DOI:10.1039/AN9830800464

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

470 Analyst, April, 1983, Vol. 108, +PI 470-475 Fully Automatic Flow Injection System for the Determination of Uranium at Trace Levels in Ore Leachates Thomas P. Lynch, Arthur F. Taylor and John N. Wilson Chemical Analysis Branch, BP Research Centre, Chertsey Road, Sunbury-on-Thames, Middlesex, T W 16 7LN An automatic flow injection system is described for the determination of uranium in ore leachates. Following injection from an autosampler, the leachate is extracted with a solution of tributyl phosphate in heptane, which removes uranium, and the organic phase is separated. The extract is reacted with an ethanolic solution of 2- (5-bromo-2-pyridylazo) -5-diethylaminophenol (BrPADAP) and benzyldimethyltetradecylammonium chloride (zephiramine) and the resulting ternary complex with U(V1) is measured spectrophoto- metrically a t 579 nm.The lower limit of determination is 0.1 p.p.m. of uranium and up to 50 samples per hour can be analysed. In terms of speed and sensitivity this improves significantly on published procedures using segmented flow systems. The technique is ideal for process control and can be applied to the analysis of ores following mineralisation. Keywords ; Flow injection ; ore leachates ; spectrophotonzetry ; uranium determination The exploitation of low-grade ores involves processes in which it is required to control uranium a t trace levels in residues, leachates and effluents, thus necessitating the use of sensitive analytical procedures that ideally are also rapid and selective. A number of relevant methods have been published, predominantly colorimetric, the usual approach being to use a non- specific chromogen together with a process for the suppression or removal of interference~.~-~~ Segmented flow systems have been used to automate the steps of colour development and measurement.13J4 Uranium is first separated from the leachate and concentrated by a manually performed solvent extraction, which is relatively time consuming.We were faced with a situation in which it would be necessary to analyse large numbers of acid leachates containing from less than 1 and up to 100 p.p.m. of U,O, in order to control ore processing, initially on a pilot plant and later on a production scale. We therefore proposed to in- corporate the solvent extraction process into a fully automatic system and at the same time to increase sensitivity so that leachates containing 0.5pgml-1 of U30, or less could be determined with a precision and speed suitable for process control.Experimental Choice of Reagent Dibenzoylmethane has been used in several colorimetric procedures for the determination of uranium following a solvent extraction process.lp2 Although we successfully adapted this procedure the sensitivity of the reagent (molar absorptivity 1.8 x lo4 1 mol-l cm-l) was barely adequate for our needs. The uranium benzoate - Malachite Green complex has been reported to have a molar absorptivity of 8.3 x 104 1 mol-l cm-l.l0 We were able to produce intensely coloured complexes using this reagent, but the colour was unstable, and a double solvent extraction was necessary to avoid interferences.Other w ~ r k e r s ~ - ~ s ~ ~ have shown the pyridyl- azo dyestuff 2-(5-bromo-2-pyridylazo)-5-diethylaminophenol (BrPADAP) to be a highly sensitive and effective reagent for the determination of uranium. In using this reagent we were able to make two modifications to published procedures that lowered the minimum level of determination. We combined an ethanol solution of the reagent with benzyldimethyl- tetradecylammonium chloride (zephiramine ; Fluka) , which together form a ternary complex with uranium. We observed an increase in molar absorptivity from 5.8 x lo4 to 6.8 x lo4 1 mol-l cm-l and a small shift in the absorption peak from 575 to 579 nm, at which wavelength interference from the background colour of the reagent is somewhat less significant (Fig.1). Following extraction of uranium with tributyl phosphate in heptane we formed theLYNCH, TAYLOR AND WILSON 471 al K m e $ 0.4 n Q: 0.2 ' 0 500 540 580 620 Wavelengthlnm Fig. 1. Absorption spectra of 1, re- agent blank Venus ethanol; 2, uranium - BrPADAP - zephiramine complex zw- sus reagent blank, Amax. = 579 nm; and 3, uranium - BrPADAP complex veYsus reagent blank, Amax. = 575 nm. coloured complex under non-aqueous conditions, in the presence of pyridine. This avoids the need for a dilution with relatively large volumes of ethanol and consequent reduction in sensitivity, in order to maintain a single phase system when aqueous buffers are used. As a result of these modifications an over-all increase in sensitivity of between five and ten times was achieved.The segmented flow systems referred to13914 both use BrPADAP, but in view of our intention to include a solvent extraction step the diffusion dependent flow injection analysis (FIA) procedure developed by RfiZiEka and Hanse1-1~~9~7 appeared to us to be more appropriate. This simple and flexible technique has recently been reviewed18919 and Karlberg and co-workers have described FIA systems incorporating solvent extraction steps for the analysis of certain pharmaceuticals.20921 The extraction of uranium and formation of the ternary complex appeared ideal processes to incorporate in a continuously flowing system. Reagents Dissolve 700 g of A1(N0,),.9H20, 9 g of NaF and 2.6 g of Na2S0, in about 900 ml of de-ionised water.Add 18 ml of glacial acetic acid and dilute to 1 1 with de-ionised water. Dissolve 0.1 g of BrPADAP and 10 g of zephiramine in about 500 ml of absolute ethanol and add 50 ml of pyridine. Mix 100 ml of tributyl phosphate with 900 ml of heptane. Aluminium nitrate salting solution. BrPADAP. Organic extractant, 10% V/V tributylphosphate in heptane. Filter all reagents through a coarse filter-paper such as a Whatman No. 541. Dilute to 1 1 with absolute ethanol. Apparatus A Technicon Autosampler is interfaced with the lOO-pl capacity sample injection loop via a Rheodyne system as follows: the Auto- sampler probe operates a microswitch, which sequentially activates a pair of Rheodyne 7163 solenoid valves. These operate a 5001 pneumatic activator, which charges or discharges a 5020 rotary sample valve.The system flow is controlled by two Watson Marlow peristaltic pumps, one fixed speed (JIHRK 18) and one variable speed (501s). Colour is measured by a Unicam SP6 spectropliotometer with a Hellma 1-cm tubular flow cell, Type OS178.12, the A flow diagram of the system is shown in Fig. 2.472 Organic extractant Alum in ium nitrate reagent Water LYNCH et al. : AUTOMATIC FLOW INJECTION 2.0 L r 1.2 L r 2.6-m coil - 1.2 L 0.9-m coil 1100-pl sample r Aqueous Analyst, Vol. 108 Flow-rate/ BrPADAP ml min-' reagent 1.4 I 4.5-m coil Fig. 2. Flow diagram of the system. signal from which is connected to a single pen chart recorder. In the reagent pump manifold Tygon tubing is used for aqueous solutions and Solvaflex tubing for organic solutions.Trans- mission and mixing tubes are of Altex PTFE tubing of 0.8 mm i.d. The phase separator (Fig. 3) is a modification of a design suggested by Betteridge,22 the ports being threaded to take Altex fittings (Anachem Ltd., Luton). Correct operation of phase separators (or de-bubblers in air segmented streams) is always important for good base-line stability, but there is one factor in our system that makes it critical. When the alkaline BrPADAP reagent is added to the organic extractant any entrained aqueous phase becomes partially miscible and this causes precipitation of aluminium hydroxide (from the aluminium nitrate) in the tubing and spectrophotometer cell. The precipitate is also stained red by absorbed BrPADAP. To allow better control of the take-off to waste ratio a separate, variable speed pump is used for the waste line, rather than relying on the conventional approach of selection of the appropriate pump tube size, with a single pump.0 0 1 0 0 A 1 nil 0 0 B Bottom C 3 Fig. 3. Diagram of the phase separator showing A, Perspex block; B, Neoprene gasket; and C, PTFE block. The Neoprene gasket is sandwiched between the PTFE block and the Perspex block and the assembly is bolted together with eight screws. The mixed aqueous-organic phase enters a t port 1 and the aqueous phase plus a trace of the organic phase are pumped t o waste from port 2. The organic phase is removed a t port 3, after which it is mixed with reagent.April, 1983 FOR DETERMINATION O F u I N ORE LEACHATES Results and Discussion 473 Calibration graphs were prepared by processing solutions of uranyl nitrate either in water or in an aqueous solution containing 25 g 1-1 of sulphuric acid.They were linear up to 50 p.p.m. of U,O,, but such was the sensitivity of the method that we normally operated in the range 0-20 p.p.m. of U,O,, samples with higher concentrations being diluted before injection. A typical calibration graph is shown in Fig. 4. a) (D 0.1 2 s n Q ' 6 min ',5 a) I Timehin (u ; 0.1 5 5: a I I I 0 4 8 12 16 Amount of U308, p.p.m. Fig. 4. Absorption spectra for (a) typical calibration peaks where the values on the peaks represent p.p.m. of U,O,; and (b) corresponding calibration graph. Chart speed, 1 min cm-l. Table I shows results for the determination of uranium in a series of leach liquors containing nominally 25 g 1-1 of sulphate.The first set of results was obtained using aqueous calibration standards. The second and third sets show results for the same samples by standard additions and against calibration standards containing appropriate levels of sulphate. Standard TABLE I DETERMINATION OF URANIUM IN ACID LEACHATES Each result is the mean of triplicate determina- tions, with no difference between any two indi- vidual results within a set exceeding 0.1 p.p.m. U,O, content, p.p.m. - 1 1.6 1.7 1.8 2 6.8 7.4 7.6 3 8.4 9.2 9.6 4 13.3 15.3 15.3 5 9.2 10.5 10.6 6 9.1 10.3 10.4 7§ Sample No. a* b t c: 8.6 10.1 - * Against sulphate-free standards. t By standard additions using microlitre additions of strong uranium solutions. $ Against standards containing 25 g 1-l of sulphuric acid, the nominal concentration of the samples. 5 A 10 p.p.m.standard in 25 g 1-' of sulphuric acid.474 LYNCH et al. : AUTOMATIC FLOW INJECTION Analyst, Vol. 108 additions were made in microlitre amounts using strong solutions of uranium so that sample volumes were effectively unchanged. Whilst the second and third sets of results are in very good agreement, the relatively low results of the first set illustrate the suppression of uranium extraction by sulphate ion reported by Walker and Vita.23 This is confirmed by sample 7, which is a 10 p.p.m. uranium standard with 25 g 1-1 of sulphate present. In Table I1 the suppression is quantified by comparing sets of 5 and 10 p.p.m. standards containing from TABLE I1 SUPPRESSION OF THE EXTRACTION OF URANIUM IN THE PRESENCE OF SULPHATE ION Concentrations indicated compared with sulphate-free standards SOd2-/g 1-1 5 p.p.m.U,O, 0 5.0 6 4.8 10 4.4 16 4.4 20 4.3 26 4.4 30 4.3 10 p.p.m. U,O, 10.0 9.7 8.9 8.9 8.8 8.8 8.8 0 to 30 g 1-1 of sulphate with a sulphate-free calibration. The suppression of 12-14% is constant within experimental error for from 10 to 30 g 1-1 of sulphate, although it is increased if the sample is injected directly into the aluminium nitrate reagent rather than into a separate aqueous stream (Fig. 2). Table I11 shows short-term repeatabilities at nominal levels of 1, 10 and 20 p.p.m. of U,O, using model leachates containing 25 g 1-1 of sulphate. Neutralisation of free acid does not modify the suppression. TABLE I11 SHORT-TERM REPEATABILITY U,O, (nominal Repeatability (95%), * level), p.p.m.n %-I p.p.m. U,O, 1 11 0.016 0.05 10 11 0.069 0.22 20 11 0.176 0.66 * For repeatability values, duplicate results by the same operator should be considered suspect if they differ by more than the amount stated in more than one in twenty determinations. It is particularly important in the processing of low-grade uranium ores to be able to monitor accurately trace levels of uranium in leachates, effluents and residues. Precision, speed and specificity are desirable characteristics of the analytical control method. One of the outstand- ing advantages of flow analysis is that system constants and timing sequences are maintained within far closer limits than is possible with manual processes; and, as in the present procedure, this results in excellent repeatability (Table 111).Good recoveries are substantially confirmed by the agreement with standard additions measurements (Table I). The sensitivity of the system is adequate for the determination of 0.5 p.p.m. of U,O, and the determination limit may be lowered if necessary by increasing the optical path length and by optimising the dis- persion, which is dependent on sample volume, flow-rates and lengths of the various elements of the flow line. The simplex approach of Nelder and Mead24 is a means of achieving this optimisation and Wade25 has recently described the application of a modified version to FIA systems. A novel feature of our system is the use of a second variable speed pump to control the off-take of the aqueous phase from the separator, and although this may seem extravagant it has contributed significantly to the reliability and long-term stability of the system.April, 1983 FOR DETERMINATION OF u IN ORE LEACHATES 475 Selectivity always presents a problem with naturally occurring materials of complex and variable composition. We found a complexing solution based on that proposed by Francoisl was ideal for the materials we were analysing.Of the well known interferences, Ce(IV), which is extracted, was reduced to Ce(III), which is not extracted, by sulphite; titanium and zir- conium were complexed by fluoride; and thorium was complexed by acetic acid. Aluminium nitrate prevents interference from common cations such as Fe(II1) and maintains a distri- bution coefficient, which, subject to the effect of sulphate described, ensures virtually complete extraction of uranyl ion into the organic phase.There is no reason why the composition of this solution should not be modified to process leachates from ores with different composition and interfering species. The anion may be determined easily in unknown solutions by ion chromatography, but in practice the sulphate concentration in a given series of leachates will usually be known and of relatively constant level. As has been shown, calibration standards may be formulated accordingly. Finally, considering speed of analysis, incorporation of the extraction process into the automatic system is a substantial advantage, as it enables leachates from the production pro- cess to be analysed without any pre-treatment. We were able to analyse up to 50 such samples in 1 hour.The sulphate effect does not present a great problem. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. References Francois, C. A., Anal. Chem., 1958, 30, 50. Hennisch, G. W., Mikrochim. Acta, 1970, 258. Perez-Bustamente, J. A., and Palomares Delgardo, F., Analyst, 1971, 96, 407. Johnson, D. A., and Florence, T. M., Anal. Chim. Acta, 1971, 53, 73. Pakalns, P., Anal. Chim. Acta, 1974, 69, 211. Johnson, D. A., and Florence, T. M., Talanta, 1975, 22, 253. “Method of Uranium Determination in Pregnant Solution,” Report No. 1841 of the South African National Institute for Metallurgy, Johannesburg, 1976. Fujinaga, T., Kuwamoto, T., and Ozaki, T., Nippon Kagaku Kaishi, 1976, 12, 1852; Chem. Abstr., 1977, 86, 164764. Pollock, E. M., Anal. Chim. Acta, 1977,88, 399. Dubey, S. C., and Nadkarni, M. N., Talanta, 1977, 24, 266. Lyle, S. J., and Tamizi, M., Anal. Chim. Acta, 1979, 108, 267. Strelow, F. W. E., and Van Der Walt, T. N., Talanta, 1979, 26, 537. Prall, J . R., “The Semi-Automatic Determination of Traces of Uranium,” US Government Report Lyle, S. J., and Tamizi, M., Anal. Chim. Acta, 1980, 121, 341. Hung, S.-C., Qu, C.-L., and Wu, S.-S., Talanta, 1982, 29, 629. RbZiCka, J., and Hansen, E. H., Anal. Chim. Acta, 1975, 78, 145. RbtiCka, J., and Hansen, E. H., Anal. Chim. Acta, 1979, 99, 37. RbiiCka, J., and Hansen, E. H., Anal. Chim., Acta, 1980, 114, 19. Betteridge, D., Anal. Chem., 1978, 50, 832A. Karlberg, B., and Thelander, S., Anal. Chim. Acta, 1978, 98, 1. Karlberg, B., Johansson, P.-A., and Thelander, S., Anal. Chim. Ada, 1979, 104, 21. Betteridge, D., personal communication. Walker, C. R., and Vita, 0. A,, Anal. Chim. Acta, 1973, 67, 119. Nelder, J. A., and Mead, R., Comput. J., 1965, 7, 308. Wade, A. P., Anal. Proc., 1983, 20, 108. NLCO 1091, 1972. Received November 8th, 1982 Accepted December 2nd, 1982

ISSN:0003-2654    

DOI:10.1039/AN9830800470

出版商:RSC    

年代:1983

数据来源: RSC