摘要:

THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYEDITORIAL ADVISORY BOARD'Chairman: H. J. Cluley ( Wemb/ey)'L. S. Bark (Salford)R. Belcher (Birmingham)L. J. Bellamy, C.B.E. (Waltham Abbey)L. S. Birks (U.S.A.)E. Bishop (Exeter)L. R. P. Butler (South Africa)E. A. M. F. Dahmen (The Netherlands)A. C. Docherty (Billingham)D. Dyrssen (Sweden)'W. T. Elwell (Birmingham)J. Hoste (Belgium)'J. A. Hunter (Edinburgh)H. M. N. H. Irving (Leeds)M. T. Kelley (U.S.A.)W. Kemula (Poland)'G. F. Kirkbright (London)G. W. C. Milner (Harwell)G. H. Morrison (U.S.A.)H. W. Nurnberg (W. Germany)'J. M. Ottaway (Glasgow)' G . E. Penketh (Wilton)'T. B. Pierce (Harwell)E. Pungor (Hungary)D. 1. Rees (London)'R. Sawyer (London)P. H. Scholes (Sheffield)'W.H. C. Shaw (Greenford)S. Siggia (U.S.A.)A. A. Smales, O.B.E. (Harwell)A. Walsh (Australia)T. S. West (Aberdeen)A. L. Wilson (Medmenham)P. Zuman (U.S.A.)A. Towns h e n d (Birmingham)*Members of the Board serving on The Analyst Publications CommitteeREG I0 NAL ADVl SO RY EDIT0 RSDr. J. Aggett, Department of Chemistry, University of Auckland, Private Bag, Auckland, NEWDr. G. Ghersini, Laboratori CISE, Casella Postale 3986, 20100 Milano, ITALY.Professor L. Gierst, Universit6 Libre de Bruxelles, Facult6 des Sciences, Avenue F.-D. Roosevelt 50,Professor R. Herrmann, Abteilung fur Med. Physik., 63 Giessen, Schlangenzahl 29, W. GERMANY.Professor W. A. E. McBryde, Dean of Faculty of Science, University of Waterloo, Waterloo, Ontario,Dr. W.Wayne Meinke, KMS Fusion Inc., 3941 Research Park Drive, P.O. Box 1567, Ann Arbor,Dr. 1. Rubeika, Geological Survey of Czechoslovakia, Kostelni 26, Praha 7, CZECHOSLOVAKIA.Dr. J. Rh%ka, Chemistry Department A, Technical University of Denmark, 2800 Lyngby, DENMARKProfessor K. Saito, Department of Chemistry, Tohoku University, Sendai, JAPAN.Dr. A. Strasheim, National Physical Research Laboratory, P.O. Box 395, Pretoria, SOUTH AFRICA.ZEALAND.Bruxelles; BELGIUM.CANADA.Mich. 481 06, U.S.A.Published by The Chemical SocietyEditorial: The Director of Publications, The Chemical Society, Burlington House,London, W1V OBN. Telephone 01 -734 9864. Telex No. 268001Advertisements: Advertisement Department, The Chemical Society, Burlington House, Piccadilly,London W1V OBN. Telephone 01 -734 9864Subscriptions (non-members): The Chemical Society, Distribution Centre, Blackhorse Road,Letchworth, Herts., SG6 1 HNVolume 103 No 1224@ The Chemical Society 1978March 197

ISSN:0003-2654    

DOI:10.1039/AN97803FX009

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

ANALAO 103 (1 224) 1 93-304 (1 978)ISSN 0003-2654March 1978THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYCONTENTS193 REVIEW. Derivative Formation in the Quantitative Gas-chromatographicAnalysis o f Pharmaceuticals. Part Il-J. D. Nicholson223 Determination of Metiamide (SKF 92058) by Differential-pulse Polarography-P. E. Anderson and W. Franklin Smyth227 Simultaneous Kinetic Determination o f Mixtures o f Ions by Digital Analysis o fAm per0 metric or Pot en t i o me t r i c Data-An n e - Marie Al brec h t - Gary, Jean - Pa u ICollin, Pierre Jost, Philippe Lagrange and Jean-Paul Schwing233 Simple Semi-automated Procedure for Determining the Lipophilic Nature o fOrganic Compounds-G. Voss238 Determination o f Phosphorus in the Presence o f Iron(lll) and lron(l1)-S.McLeod and (the late) A.R. P. Clarke246 Relationship Between Activity and Concentration Measurements of PlasmaPotassium-D. M. Band, J. Kratochvil, P. A. Poole Wilson and Tom Treasure252 Quantitative Determination of Texturised Soya Protein by a StereologicalTechnique-F. 0. Flint and M. V. Meech259 Single-cell Proteins from Methanol : Elucidation o f the Structures o f theUnsaturated Fatty Acids-L. Cavalli, A. Landone, G. Cancellieri and A. Zotti268 Determination o f Morphine in Raw Opium, Opium Tincture and Other OpiatePreparations by Gas - Liquid Chromatography-Joint Committee on Pharma-ceutical Analysis284 Determination o f Hexachlorophane in Cosmetic and Toilet Preparations-Joint Committee on Pharmaceutical AnalysisSHORT PAPER296 Use o f Non-segmented High-speed Continuous Flow Analysis f o r the Deter-mination of Calcium in Animal Feeds-W.D. Basson and J. F. Van Staden300 Book ReviewsSummaries of Papers in this fssue-Pages iv, v, viii, xPrinted by Heffers Printers Ltd Cambridge EnglandEntered as Second Class at New York, USA, Post OfficSelected Annual Reviewsof the Analytical SciencesVolume 4CONTENTSL'Advances in Voltammetric Techniques,' byB. Fleet and R. D. Jee'High-frequency Electrodeless PlasmaSpectrometry,' by B. L. SharpPp. vi + 73 f 9.50ISBN 0 85990 204 8CS Members' price f3.00Orders should be sent direct, with remittance, orthrough your usual bookseller to-THE CHEMICAL SOCIETYDistribution Centre,Blackhorse Road, Letchworth,Herts. SG6 1 HN, EnglandCS Members must write direct to the above addressenclosing the appropriate remittance.JOURNALS * BOOKSMONOGRAPHSOrders for all publicationsformerly published by the Societyfor Analytical Chemistry shouldbe sent direct or through abookseller toTHE CHEMICAL SOCIETY,Distribution Centre,Blackhorse Road, LetchworthHerts., SG6 1 HNNOTICE TO SUBSCRIBERS(other than Members of the Society)Subscriptions for The Analyst, Analytical Abstracts and Proceedings shouldbe sent to:The Chemical Society, Distribution Centre,Blackhorse Road, Letchworth, Herts., SG6 1 HNRates for 1978The Analyst, Analytical Abstracts and Proceedings (including indexes):(a) The Analyst, Analytical Abstracts and Proceedings .. . . .. . . f99.00(b) The Analyst, Analytical Abstracts printed on one side of the paper, andProceedings . . . . . . a . . . . . . . .. . . f105.00The Analyst and Analytical Abstracts without Proceedings (including indexes) :The Analyst, and Analytical Abstracts printed on one side of the paper( c ) The Analyst, and Analytical Abstracts . . .. .. . . .. . . f87.00(d) . . f 93.00(Subscriptions are NOT accepted for The Analyst and/or for Proceedings alone)Analytical Abstracts only (two volumes per year, including indexes):(e) Analytical Abstracts . . . . . I . . . . . . . . . . f67.00(f) Analytical Abstracts printed on one side of the paper . . . . . . . . f73.0

ISSN:0003-2654    

DOI:10.1039/AN97803BX011

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

i V SUMMARIES OF PAPERS IN THIS ISSUE MaYch, 1978Summaries of Papers in this IssueDerivative Formation in the Quantitative Gas -chromatographicAnalysis of Pharmaceuticals. Part I1A ReviewSummary of ContentsSilylationHydrolysisInteraction of amino and carbonyl groupsUse of NN-dimethylformamide tlialkyl acetalsOxime formationCarbamate formationPyrolysisReductionOxidationTsothiocyanate and isocyanate formationCyclisation of biguanides ant1 guanidinesPhosphorus-containing derivativesAmide formationPhenylhydrazone formationKeywords : Review ; l3,havmaceutical analysis ; devivatieie fovmatioiz ; gasch vonzatogi.ap h yJ. D. NICHOLSONMedicines Testing Laboratory, Pharmaceutical Society of Great Britain, 36 YorkPlace, Edinburgh, EH1 3HU.AJzalyst, 1978, 103, 193-222.Determination of Metiamide (SKF 92058) by Differential-pulsePolarographyMetiamide (SKF 92058) undergoes anodic oxidation a t the dropping-mercuryelectrode with the formation of mercury salt (s) .Differential-pulse polaro-graphic analysis has been carried out on several formulations in 0.1 M sodiumhydroxide solution as supporting electrolyte using the peak correspondingto this oxidation process. The effects of degradation products and anti-oxidants on this determination have been investigated in addition to com-paring the polarographic method with one involving high-performance liquidchromatography.Keywords : Metiamide detevvnination; difevential-pulse polavography ; anti-oxidants ; degradation productsP.E. ANDERSONAnalytical Development Laboratory, Smith, Kline and French Laboratories Ltd.,Welwyn Garden City, Hertfordshire, AL7 1EY.and W. FRANKLIN SMYTHDepartment of Chemistry, Chelsea College, University of London, Manreaa Road,London, SW3 6LX.Analyst, 1978, 103, 223-226,Mnrch, 1978 SUMMARIES OF PAPERS I N THIS ISSUESimultaneous Kinetic Determination of Mixtures of Ions by DigitalAnalysis of hmperometric or Potentiometric DataElectroanalytical techniques allow the kinetic determination of concentrationsof ions in their mixtures when the reactions used proceed a t sufficiently slowvelocity. Amperometry and potentiometry with digital data acquisition andtreatment have been used for determinations of calcium ancl magnesium ionsbased on the exchange reactions between lead(I1) and alkaline earth metalcomplexes of ( t r a m - 1,2-cyclohexylenedinitriloj tetraacetic acid.Over a con-centration range of 10-4-5 x lo-’ M precision and accuracy are about 5 - 1 0 ~ ~ oin amperometry and 5 yG in potentiometrp.Keywovds : Calcium detevmination ; magnesiuin deteiwzination ; electvoamlysis ;kinetics ; automationANNE-MARIE ALBRECHT-GARY, JEAN-PAUL COLLIN, PIERRE JOST,PHILIPPE LAGRANGE and JEAN-PAUL SCHWING&ole Nationale Sup@rieure de Chimie, 1 Rue I3laisc T’ascal, 67000 Strasbourg, France.il azalyst, 1978, 103, 227- 232.Simple Semi- automated Procedure for Determining theLipophilic Nature of Organic CompoundsThis paper describes a semi-automatecl procedure for determination of therelative lipophilicities of organic compounds xvithin a homologous series.Theprinciple of the method is based on the ability of flexible polymer tubing toabsorb highly lipophilic materials more readilv than less lipophilic materials.The experimental assembly consists of a Technicon proportioning pumpthat generates two continuously flowing streams of an aqueous solution of thetest compound and an ultraviolet spectrophotometer equipped with flow celland recorder. The first stream passes tlirough the flow cell without beingpumped through polymer tubing, thus representing the reference sample Thesecond stream, however, is “tube-partitioned.” It arrives in tlie flow cellafter the first stream and is recorded as a second peak s!iortly after the refer-ence sample.The ratio of the peak heights was found to be a measure ofthe lipophilic nature of the compounds investigated. Substituted phenyl-ureas and sulphonamides were screened by the new procedure and the resultscompared with those obtained by more conventional approaches using linearregression analysis.Keywords : Lipophilicity ; automated analysis ; phenyluvea hevbicides ; szd-p honamidesG. VOSSAgrochemicals Division, CIBA-GEIGY Ltd., Basle, Switzerland.A n d j J s t , 1978, 103, 233-237.Determination of Phosphorus in the Presence ofIron(II1) and Iron(I1)Both manual and automated techniques have been developecl for thedetermination of phosphorus in the presence of iron(II1). Interference fromiron(II1) has been substantially eliminated by its reduction with hydroxyl-ammonium chloride prior t o the formation of tlie molybdophosphate bluecomplex using ascorbic acid.The small but reproducible error clue to theformation of iron(I1) is dependent on the iron to phosphorus ratio present inthe digest solution but is significant only when this ratio is more than 500.Both methods have been applied successfully to the determination ofphosphorus in digests and extracts of soil materials where the ratio is far belowthe value of 500, thus making a correction unnecessary.Keywords : Phosphorus detevmination ; soil analysis ; spectvophotowetry ; ivoninterferenceS. McLEOD and (the late) A. R. P. CLARKECommonwealth Scientific ancl Industrial Research Organisation, Division o f Soils,Private Bag No.2, Glen Osmond, South Australia 5064.A n d ~ ~ s t , 1978, 103, 238-245THE ANALYST March, 1978 viMATHEMATICAL MODELS IN WATER POLLUTION CONTROLedited by A. James, Head o f Division of Public Health Engineering, University of Newcastle-upon-Tyne.This book attempts to present a comprehensive account of the use of mathematical models in the controlof water pollution. After an introductory section outlining the mathematical techniques, the rest of thebook is devoted to examples of the use of models in various branches, especially in the design andoperation of waste treatment plants and the management of lakes, rivers, estuaries and the sea.0471 99471 5 approx. 480 pages In Press approx. $27.00/€15.00POTENTIOMETRIC WATER ANALYSISby D.Midgley and K. Torrance, both o f Central Electricity Research Laboratories, Leatherhead, Surrey.With increasing demands on water for industrial, agricultural and domestic consumption, resources arebeing used and re-used more intensively than ever before. In the past, non-specific parameters such asconductivity, pH and biological oxygen demand have largely sufficed as measures of water quality, but theneed in many circumstances for the determination of specific chemical substances i s becoming morepressing. A number of species can now be determined by means of potentiometry with ion-selective elec-trodes, this technique having the advantage of being easily adaptable to different circumstances.0471 99532 0 approx. 384 pages In Press approx. $28.00/€15.00AIR POLLUTION CONTROL Part 3edited by W. Strauss, University of Melbourne, Victoria, Australia.This discusses source and ambient sampling and monitoring of gaseous pollutants. It also covers monitor-ing, sampling, and physical and chemical analysis, including particle size and concentrating on traceelements, of other pollutants. It discusses the integration of physicochemical methods into a regionalcontrol system, including bases for selection of particular networks. Hydrogen sulfide in pollutant concen-trations from both anthropogenic and natural geothermal sources i s examined. (Environmental Scienceand Technology Series)0471 83323 1 approx. 560 pages In Press approx. $31.70/f18.70COLORIMETRIC DETERMINATION OF NONMETALS 2nd Ed.edited by D.F. Boltz and J.A. Howell, Western Michigan University.This describes the principles and practices involved in colorimetric analysis of non-metals. (ChemicalAnalysis Series Vol. 8)0471 087505 approx. 560 pages In Press approx. 841.85/f23.2

ISSN:0003-2654    

DOI:10.1039/AN97803FP017

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

...V l l l SUMMARIES OF PAPERS IN THIS ISSUE MaYch, 1978Relationship Between Activity and Concentration Measurements ofPlasma PotassiumAn ion-selective electrode incorporating valinomycin in a poly(viny1 chloride)membrane was used for examining potassium in blood and plasma. Usingan electrode cell thermostated a t 37 O C and a saturated potassium chlorideliquid junction, the activity of plasma potassium appeared to be very similarto that for similar concentrations of potassium in physiological saline.An apparent difference in potassium activities between the plasma of wholeblood and separated plasma was found to be comparable to the discrepancybetween the pH of whole blood and plasma, and was attributed to an effectof red blood cells on the liquid junction of the measuring cell.The magnitude of the liquid junction potential between a bridge solutionand blood is a major uncertainty both in the determination of activities ofions by electrode measurements (e.g., the biological pH scale) and in themeasurement of potential differences across biological membranes.Theadvent of the highly selective plastic membrane electrodes enables thisproblem to be reexamined.Keywords ; Plasma protein determination ; potassium activity ; ion-selectiveD. M. BAND, J. KRATOCHVIL, P. A. POOLE WILSON and TOMTREASURESherrington School of Physiology and the Intensive Care Unit, St. Thomas’sHospital, London, SE1 7EH.Analyst, 1978, 103, 246-251.electrodes ; ualinowyciiz ; liquid junction potentialQuantitative Determination of Texturised Soya Protein by aStereological TechniqueA method for the determination of texturised soya protein in comminutedmeat products is proposed.The microscopical technique described isstereological and uses a 42-point Weibel graticule on 10-pm sections of meat -soya mixtures stained with toluidine blue contained in an aqueous mountant.Results are presented in terms of volume in volume fractions. The accuracyof the method depends largely on the number of points counted and anequation is suggested for determining the number to be counted for a givenaccuracy. The method can be used for other constituents of comminutedmeats, e.g., rind and rusk, and is applicable to any thin sections of materialin which the constituents are sufficiently contrasted.Keywords : Soya protein determination ; steveology ; meat and soya ynixtuvesF.0. FLINT and M. V. MEECHProcter Department of Food and Leather Science, University of Leeds, Leeds,LS2 9 JT.AHalyst, 1978, 103, 252-258THE ANALYST ix March, 1978THE QUEEN'S UNIVERSITYOF BELFASTMSc COURSE inANALYTICAL CHEMISTRYApplications are invited for admission to thisestablished 12 month full-time MSc coursewhich provides a comprehensive training inthe theory and practice of modern chemicaland instrumental methods of analysis. Appli-cants should normally possess an honoursdegree (or equivalent) in chemistry or cognatesubjects. Part-time courses are available.The Science Research Council has recognisedthe course for tenure of its Advanced CourseStudentships.A description booklet and application formscan be obtained from Professor D.ThorburnBurns, Dept. of Chcmistry, Queen's Universityof Belfast, Belfast, BT7 1 ",Northern Ireland.ADVERTISERSPLEASE NOTEAll advertisements forTHE ANALYSTshould from now on be addressed t oour ownAdvertisement Department,The Chemical Society,Burlington House,Piccadilly, London W1V OBNTel: 01-734 9864Please send all space orders, copy,enquiries etc. to this address.WILMAD GLASS CO.I N CO R PO RATEDOffer a complete service to Spectroscopists.Their range includes:NMR sample tubesEPR quartz sample tubesPrecision bore tubingVials and Septum bottles for sample storageCoaxial systemsVortex plugs etc.For details contact their Exclusive U.K.agents:FLUOROCHEM LIMITEDDinting Vale Trading Estate,Dinting Lane, Glossop,Derbyshire, SKI3 9NU.Telephone: Glossop (04574) 4917 or 2855Telex: 669960NewEURO-STANDARDnow availableE. S ,877-1Furnace Dust (LD Converter)Certified for the following elements :Fe, Si, Ca, Mg, Al. Ti, Mn, P, S, Na,K, F, V, Cr, Ni, C, Zn, Pb. Cu and AsFull details obtainable from :Bureau of Analysed SamplesLtd.Newham Hall, Newby,Mlddlesbrough, Cleveland TS8 9EATelephone: Middlesbrough 31721 X SUMMARIES OF PAPERS I N THIS ISSUE March, 2978Single-cell Proteins from Methanol : Elucidation of the Structuresof the Unsaturated Fatty AcidsThe unsaturated fatty acids contained in single-cell proteins prepared fromLi 70 yeasts grown on methanol were studied in order to ascertain the positionof unsaturation and the stereochemistry of the double bonds.The acids wereextracted and esterified to the corresponding fatty acid methyl esters(FAMEs) as described in previous reports. The location of the double bondswas obtained by gas-chromatographic - mass-spectrometric analysis of thetrimethylsilyloxy derivatives of the FAMEs. The stereochemistry wasdeduced by nuclear magnetic resonance spectroscopy using an incrementaladdition study with Eu(fod), on FAMEs and on the corresponding epoxyderivatives. Three model FAME samples, methyl oleate (9c-C18:l), methylelaidate (9t-Cls:1) and methyl linoleate (9c, l 2 ~ - C , ~ : ~ ) , were compared. Mostof the unsaturated fatty acids (about 7476) .present in the single-cell proteinshave structures identical with those known in animal lipids.The three mostabundant unsaturated fatty acids display the following structure : 9c-Cl,,,(22.5y0), 9c-C1,,, (19.4%) and 9c,12c-ClS:, (29.2%). All double bondspossess a cis-configuration; the position of the double bond is at C9-C10 for themonounsaturated acids and C9-C10 and C12-Cl3 for diunsaturated acids.Keywords : Single-cell proteins ; methanol; unsaturated fatty acid structures ;gas chromatography - mass spectrometry ; N M R spectroscopyL. CAVALLI, A. LANDONE, G. CANCELLIERI and A. ZOTTIEUTECO SPA, Centro Ricerche, Via Reali 4, 20037 Paderno Dugnano (Milan), Italy.Analyst, 1978, 103, 259-267.Determination of Morphine in Raw Opium, Optium Tincture andOther Opiate Preparations by Gas - Liquid ChromatographyReport prepared by the Joint Committee of the Pharmaceutical Society andthe Analytical Division of The Chemical Society on Pharmaceutical Analysis.Keywovds : Morphine determination ; vaw opium ; opium tinctuve ; opiatepreparations ; gas - liquid chromatographyANALYTICAL METHODS COMMITTEEThe Chemical Society, Burlington House, London, W1V OBN.Analyst, 1978, 103, 268-283.Determination of Hexachlorophane in Cosmetic andToilet PreparationsReport prepared by the Joint Committee of the Pharmaceutical Society andthe Analytical Division of The Chemical Society on Pharmaceutical Analysis.Keywords : Hexachlorophane determination ; cosmetic and toilet preparations ;thin-layer chromatography ; diflevential spectrophotometvy ; Sephadex LH20chromatographyANALYTICAL METHODS COMMITTEEThe Chemical Society, Burlington House, London, W1V OBN.Analyst, 1978, 103, 284-295.Use of Non-segmented High-speed Continuous Flow Analysisfor the Determination of Calcium in Animal FeedsShort PaperKeywords : Calcium determination ; animal feed analysis ; non-segmentedcontinuous $ow analysisW.D. BASSON and J. F. VAN STADENDepartment of Inorganic and Analytical Chemistry, University of Pretoria, Pretoria,Republic of South Africa.Analyst, 1978, 103, 296-299

ISSN:0003-2654    

DOI:10.1039/AN97803BP021

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

MARCH 1978 The Analyst Vol. 103 No. 1224 Derivative Formation in the Quantitative Gas-chromatographic Analysis of Pharmaceuticals. Part I I" A Review J. D. Nicholson Medicines Tes2ing Laboratory, Pharmaceutical Society of Great Britain, 36 York Place, Edinburgh, EH1 3H U Summary of Contents Silylation Hydrolysis Interaction of amino and carbonyl groups Use of NN-dimethylformamide dialkyl acetals Oxime formation Carbamate formation Pyrolysis Reduction Oxidation Isothiocyanate and isocyanate formation Cyclisation of biguanides and guanidines Phosphorus-containing derivatives Amide formation Phenylhydrazone formation Keywords : Review ; fdzarrnacezttical analysis ; derivative formation ; gas chromatography Silylation Silylation is now a universal technique for forming derivatives of compounds that possess functional groups with reactive hydrogen atoms, such as alcohols, phenols, amines and acids.The hydrogen atom is replaced with a silicon atom to which are attached suitable groups, of which the most frequently encountered is methyl, which gives rise to trimethylsilyl (TMS) derivatives. The resultant product is usually thermally stable, volatile and of low polarity.262 Various reagents are available for silylation reactions and will be considered individually below. The reaction conditions vary greatly and depend on the reactivity of the group being derivatised and the degree of steric hindrance to which it is subjected. The more reactive compounds are derivatised by merely dissolving them in the appropriate reagent. If solu- bility problems are encountered, a mutual solvent such as pyridine or dimethylfonnamide can be used.Such solvents also posseSs a degree of catalytic activity, but catalysis is normally achieved by the addition of trimethylchlorosilane (TMCS). The latter is a common additive, to the extent of l-lOyo, in some commercially available silylating mixtures. For difficult silylations it may be necessary to use more potent reagents and/or to heat the reaction mixture in order to achieve complete reaction in a reasonable period of time. It is usually necessary to carry out all reactions under anhydrous conditions and with an excess of reagent as the deriva- tives are susceptible to hydrolytic degradation. Protection of samples from moisture, during and after silylation, is usually achieved by employing vessels with ground-glass stoppers or PTFE-faced screw-caps.Cork stoppers, which may induce decomposition of the deriva- tives,263 should be avoided. * For Part I, see Analyst, 1978, 103, 1. 193194 Analyst, VoZ. 103 The various silylating reagents in use will now be considered in turn, with attention being given to typical reaction conditions and to the groups of compounds to which each is applicable. NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Hexamethyldisilazane Hexamethyldisilazane (HMDS) is one of the original reagents used to prepare TMS deriva- tives and, although in comparison with subsequently introduced reagents it is not a strong TMS donor, its use continues. HMDS alone gives a good yield of derivative by direct reaction with acids, a poorer yield with alcohols andl no derivative with amines.264 However, the introduction of catalytic amounts of TMCS improves the yield significantly.1279264 When HMDS and TMCS are used together in pyridine, a classical silylating mixture is produced, This was originally introduced265 for the silylation of carbohydrates but has subsequently been used for many other compounds. The silylation of carbohydrates in aqueous solution is normally difficult to achieve but a reaction mixture consisting of HMDS, pyridine and tri- fluoroacetic acid can successfully silylate S O X , solids ~yrups.~66 Although HMDS, in carbon disulphide solution, can be used to prepare the TMS ester of fenoprofen for chromatographic analysis,267 it is more common to utilise the catalytic activity of TMCS in conjunction with a suitable solvent such as pyridine.Such reaction mixtures can be used to derivatise rapidly, for example, warfarin and other anticoagulants,268 chloram- phenic01~6~ and some ben~odiazepines.~~~ Pyriidine is perhaps the most frequently encountered reaction medium but use can be made, in certain instances, of dimethylformamide,2681270,271 dimethyl sulphoxide,268 carbon d i s ~ l p h i d e ~ ~ ~ and t0luene.~7~ The silylation of thiamphenicol is much slower in pyridine than in a~etonitrile.~~~ During derivatisation a fine precipitate of ammonium chloride is often produced, but this does not affect the reaction or the ~hromatography.l~~9~~~9~~~ If the reaction is carried out under reflux the ammonium chloride sublimes into the condenser and is separated from the reaction medi~m.~~2 Although it is normal practice to inject the reaction medium directly into the chromatograph, interference may occasionally be encountered from excess of reagent or by-products.To circumvent this problem, the derivative, when prepared in dimethylformamide solution, can be extracted using isooctane after prior saturation of the dimethylformamide with ammonium nitrate. 271 If suitable, HMDS provides a relatively cheap silylating reagent that is less extreme in its reactivity than newer reagents, a factor of imLportance in the derivatisation of, for example, chloramphenico1269 and fen~profen.~~' However, HMDS does not react readily with sterically hindered groups276 or the less reactive hydrogen atoms.N-Trimethylsilyldiethylamine N-Trimethylsilyldiethylamine (TMSDEA) is a more potent TMS donor than HMDS26* but does not appear to have gained popular acceptance for pharmaceutical analyses. However, by combining TMSDEA with other silylating reagents, it is possible to overcome problems encountered when the latter are used alone. Neomycin is completely silylated by a mixture of TMSDEA and trimethylsilylimidazole in p,yridine solution but the amounts used are criti~a1.2761~77 Ergonovine can be completely silylated by a similar mixture of reagents, neither of which gives satisfactory results when used alone.278 Agroclavine is silylated by using a complex mixture of TMSDEA, TMCS, pyridine and NO-bis(trimethylsily1) trifluoro- acetamide under comparatively mild temperature conditions.279 TMSDEA does offer the basis of a technique for introducing groups other than TMS into compounds. It is relatively easy to prepare the corresponding higher trialkylsilyldiethyl- amines, which allow, for example, triethyl-, tripropyl-, tributyl- and trihexylsilyl derivatives to be prepared.%O This approach can be used with cannabinoids, for which the separation of TMS derivatives is incomplete but that of tributylsilyl derivatives is complete.280 NO - Bis( trimet hylsilyl) acetamide NO-Bis(trimethylsily1)acetamide (BSA) was introducedB1 as a more potent TMS donor than either HMDS or TMSDEA and is now established as one of the most commonly used silylating reagents.BSA reacts with non-sterically hindered alcohols, amines, carboxylic acids, amides, amino acids and enols and, depending on the ease of transfer of the TMSMarch, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11.REVIEW 195 groups, will exchange one or both TMS groups. The resulting reaction by-products, N - trimethylsilylacetamide and acetamide, are sufficiently volatile not to interfere in most chromatographic separations. BSA can be used alone or in conjunction with l - l O % of TMCS as a ~ a t a l y s t . ~ ~ ~ , ~ ~ ~ The effect of the TMCS is to increase both the rate of reaction and the ability of the BSA to silylate hindered gr0ups.~74 It is also common practice to carry out reactions in a solvent whose properties favour silylation. Pyridine274s279s283-285 and dimethylf~rmamidel~ are popular solvents, with use also being made of acetonitrile,281s282 carbon tetrachloride,286 1,4-dioxan,B7 diethyl etheral and tetrahydrofuran.%* The efficiency of TMCS as a catalyst is demonstrated in the silylation of tryptamine, which forms a mono-TMS derivative with BSA in pyridine.283 In the presence of TMCS, the di-TMS derivative forms together with some substitution of the ring nitrogen.The choice of solvent can be dictated by the solubility of the drug being derivatised and that of the derivative, but interference in the chromatogram from tailing solvent peaks is also a factor that influences the choice of solvent. However, the solvent does, in some instances, influence the nature of the derivative formed, especially when there are several potentially reactive groups in the molecule being derivatised.This is well illustrated by chloramphenicol, which forms bis- and tris-TMS derivatives after reaction with BSA in acetonitrile or pyridine but only mono- and bis- derivatives in chloroform or ethyl acetate s0lution.~6~ Similar behaviour is noted in the silylation of thiarnphenic01.~~3 An interesting approach, not involv- ing solvents, is the direct reaction of BSA with compounds adsorbed on to alumina followed by chromatography of the BSA.289 BSA will also derivatise a large number of drugs when it is injected together with these drugs into a chromatograph using an “on-column derivatisa- t ion. ’ ’290 Reactions involving BSA are normally carried out under anhydrous conditions but it has been d e m o n ~ t r a t e d ~ ~ ~ ~ ~ 8 ~ that the presence of 1% of water dramatically increases the rate of reaction between BSA and noradrenaline in pyridine or acetonitrile solution.The catalytic activity may not be due directly to the water but to the trimethylsilanol that it forms by reaction with BSA.274 The presence of acids may also induce catalysis.274 Temperature is an important consideration in silylation reactions and for those less tractable substances derivatisation can often be assisted by a suitable increase in temperature. However, the stability to heat of the drug being derivatised must be taken into account. On completion of the reaction, it is usual to chromatograph the reaction mixture directly when excess of reagents and by-products will normally be eluted before the derivatives of interest. For separations carried out at relatively low temperatures, interfering peaks291-293 may be observed, which can be attributed to the age and condition of storage of the BSA.291 From fresh reagent chromatographic peaks due to acetonitrile, trimethylchlorosilane, hexa- methyldisiloxane and monotrimethylsilylacetamide may be detected in addition to BSA.Older reagent may give up to 13 peaks.291 In addition to giving increased rates of silylation and the derivatisation of less reactive groups, the high reactivity of BSA can present problems with some compounds that possess more than one reactive site. Thus, testosterone reacts with HMDS and TMCS in pyridine solution to give one derivative.275 With BSA, enolisation of the keto group occurs with the resultant formation of two isomeric derivatives, which appear as additional peaks on the chromatogram.Chloramphenicol shows similar behaviour to testosterone, forming only the bis-TMS ether with HMDS, but either a mixture of bis- and tris- derivatives or mono- and bis- derivatives with BSA, depending on the solvent used.269 Although in these instances the milder reagent is to be preferred, some keto compounds require the potency of BSA to give satisfactory derivatisation of hydroxy groups. In such instances where enolisation can occur, protection of the ketone group by oxime formation is then necessary.275 NO - Bis( trimethylsily1)trifluoroacetamide NO-Bis(trimethylsily1)trifluoroacetamide (BSTFA) was proposed294 as an alternative to BSA for two main reasons : BSTFA and its by-products monotrimethylsilyltrifluoroacetamide and trifluoroacetamide are more volatile than BSA or its b y - p r o d ~ c t s ~ ~ ~ , ~ ~ ~ and so cause less interference in chromatograms; and the presence of fluorine atoms results in less fouling of flame-ionisation detectors by deposits of silica.294196 AnaZyst, VoZ.I03 BSTFA possesses a TMS donor strength similar to that of BSA, although in some instances it shows slightly greater reactivity, e.g., towards the 11-/?-01 group in hydroxysteroids.274 It is used in the same way as BSA and for the same compounds with or without TMCS catalysis. In addition to TMCS, trifluoroacetic acid and pyridine are efficient catalysts for the silylation of phenols.32 BSTFA has gained wide acceptance and its satisfactory use has been described in the determination of , for example, ph~sphate,~g~ opiates,53 fl~phenazine,2~~ fluor0uracil,~98 pentazocine,14 salicylic acid299 and n~lidrin.~OO N- Methyl- N-trimethylsilyltrifluoroacetamide N-Methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) has similar reactive properties to BSA and BSTFA as a TMS donor but it and its reaction by-product, N-methyltrifluoroacet- amide, are even more volatile than BSA and BSTFA and their by-products.301 However, it is only when working with the more volatile TIMS derivatives or at low concentrations that the merits of MSTFA are appreciated.Although MSTFA is normally used in a similar manner to BSA and BSTFA, a novel approach can be used for bases such as catecholamines and ephedrine.302 The base, or its salt, is dissolved in trifluoroacetic acid and MSTFA is carefully added.A vigorous exothermic reaction ensues, derivative formation being complete within a few minutes. This technique has several advantages over conventional approaches : the trifluoroacetic acid acts as an excellent solvent for bases, which allows them to be derivatised in a minimum volume of solvent; the solvent, reagent and by-product are highly volatile; and the reaction does not require extended heating and is relatively mild. Under the described conditions, esterification of hydroxyl groups by the solvent does not occur. The use of MSTFA with a basic catalyst, such as potassium acetate, ensures the formation of the 20,21-enediol-TMS derivatives and of the 20,21-ene-17,20,21-triol-TMS derivatives of the glucocorticoids as well as the formation of thie enol-TMS ethers of the oxosteroids.303 Very stable N-trifluoroacetyl-0-TMS-phenolalkylamines can be prepared by reaction of the hydroxyamine or its salts first with a slight excess of MSTFA and then with N-methylbis- (trifluoroacetamide) .304p305 The resulting derivative chromatographs well without isolation from the reaction medium and can be detected at the femtomole level by using a mass spectro- meter in the single-ion mode as the detector.Trime thylsilylimidazole Trimethylsilylimidazole (TSIM) is a potent TMS donor but normally reacts only with hydroxyl g r 0 ~ p ~ , ~ ~ ~ s ~ 0 ~ although it does form derivatives with aniline, phthalimide and 9- toluenes~lphonamide.~~~ It is used in pyridiine262p273p282 or a ~ e t o n i t r i l e ~ ~ p ~ ~ ~ solution with the application of heat if neces~ary.*~s=~ If used in chloroform solution, as in the assay of etham- butol and pyridoxine, a much reduced solvent.peak is prod~ced.~~B A similar observation has been made in the assay of vitamin D2.307 TSIM can also be used as an “on-column” derivatis- ing reagent with an economy of reagent and time.308 TSIM is a valuable reagent for derivatising sterically hindered hydroxyl groups and, when using a series of reagents of increasing potency (BSA, BSA + TMCS and BSA + TMCS + TSIM) , it is possible to classify steroid hydroxyl groups as unhindered, moderately hindered or highly hindered, respectively.274 Although both BSA and TMCS increase the potency of TSIM in its reaction with sterols, such as cortol, at low temperatures they have little effect on reactions carried out above 159 0C.309 However, not all substances can tolerate such elevated reaction temperat~res.~O~ The presence of hydrogen chloride catalyses the reaction between hydroxysteroids and TSIM and can be used to advantage.310 A mixture of TSIM and TMSDEA is of value in silylating ergonovine as the latter’s poor solubility in TMSDEA hinders rapid reaction and TSIM alone will not react with the amino group.278 When the two reagents are combined, reaction is rapid.BSA does not give a unique derivative from ergonovine and so cannot be used. BSA also gives more than one derivative with thiamphenicol, whereas only TSIM in pyridine gives a single derivative without de- chlorinating the thiarnphenic~l.~~~ Although TSIM has the disadvantage of usually reacting only with hydroxyl groups, its potency, when employed in conjunction with a more general reagent such as BSA, gives a tool for derivatising the most intractable of compounds.An advantage of TSIM is its better NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC used in conjunction with BSA, gives a mixture capable of silylating most gr0ups.~7~s309 BYMarch, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 197 tolerance than other reagents towards the presence of moisture.306 The ability of TSIM to derivatise sugars, in the presence of up to 50% of water, is one of the major factors that resulted in the initial commercial availability of the reagent for gas chromatography.262 It is claimed that the TSIM peak tails badly,292 a factor that must be considered when working at low concentrations of drugs or with rapidly eluting derivatives.Other Trimethylsilyl Donors Other TMS donors have been investigated in attempts to produce reagents of greater potency and with a wider range of applicability than those already mentioned. Compounds examined include TMS-a~etanilide,~~~ TMS-~-etho~yacetanilide,29~ TMS-~iperidine,~~~ TMS- p y r r ~ l i d i n e ~ ~ ~ and TMS-m~rpholine.~~~ Claims favouring these reagents include their economy,293 their good solubility in organic solvent^,^^^^^^^ their high volatility2g29293 and their good solvent properties.292 There appears to be a relationship between pK, of the amine from which the reagent is derived and the TMS donor strength of the latter.292~293 As pK, increases the donor strength decreases.As yet, these reagents have not gained general acceptance. Halogenated Silylating Reagents Halogenated silylating reagents have been investigated with a view to producing silyl derivatives that can be sensitively detected by an electron-capture detector. The derivatives produced by these reagents tend to be less volatile and to have longer retention times than their TMS counterparts, a factor that can aid separations, as has been shown for the chloromethyl- dimethylsilyl (CDMS) ethers of the Al6-ster0ids.~~~ A further increase in the sensitivity of detection can be achieved by preparing the iodo analogues of the CDMS ethers, as with oestradiol and oestrone.71 CDMS ethers, and their corresponding bromo analogues, can be prepared by first reacting either chloro- or bromomethyldimethylchlorosilane with diethylamine and then centrifuging the precipitate.The product, in hexane solution, is then added to the drug dissolved in a suitable solvent, with heating if necessary, when derivative formation occurs. Excess of reagent can then be evaporated off before ~hromatography.7~~~~~ The electron-capturing properties of the bromo derivatives can be greatly enhanced by their conversion into the corresponding iodo derivatives by reaction with sodium iodide.93 Di(chloromethy1) tetramethyldisilazane (DCTMDS) reacts readily with phenols in pyridine solution to form chloromethylsilyl derivatives, which are sensitively detected by electron capture.312 Unfortunately, the derivatives of unsaturated phenolic acids and trihydroxy- benzoic acids give erratic results attributable to their thermal instability.As fluoroalkylsilanes with a- or /3-fluorine atoms relative to the silicon atom are thermally unstable, their use is limited.779313 However, pentafluoropentyl and trifluoropropyl derivatives of, for example, cholesterol are stable and volatile as well as invoking a strong electron- capturing response. Although the pentafluorophenyldimethylsilyl derivatives have lower volatility than those already mentioned, their stability and powerful electron-capturing properties make them favoured d e r i v a t i v e ~ . ~ 7 ~ ~ ~ ~ The substituting group has been named flophemesyl and a number of reagents have been investigated to allow its ready substitution into various gro~ps.~15 A mixture of equal amounts of flophemesyl chloride and flophemesyl- diethylamine is a potent reagent capable of reacting with sterically hindered hydroxyl groups and en0ls.~15 Unlike some other silylating reagents, flophemesylamine is selective and will silylate primary and secondary hydroxyl groups but will not react with ketones.314 However, the bulk of the fluophemesyl group makes the simultaneous silylation of proximal hydroxyl groups difficult.Despite the availability of these various reagents, the preparation and use of halogenated silyl derivatives is not yet a widely accepted technique in the gas chromatography of pharmaceuti- cals. Silylation of Various Functional Groups types of compounds and the optimum conditions to be used.various groups of compounds will, therefore , be summarised. It is of importance to consider the approaches available for the derivatisation of different The silylation techniques for198 NICHOLSON : DERIVATIVE FORMATION I N THE QUANTITATIVE GC Analyst, VOl. 103 Alcohols and phenols Alcohols and phenols react readily with all of the silylating reagents mentioned. Particular reference will be made later to steroids. For many carboxhydrates complete derivatisation is rapidly achieved by the classical appr0ach,~~5 using HMDS and TMCS in pyridine solution, to give stable, volatile derivatives. The same approach is applicable to other hydrloxy compounds such as chloramphenicol.269 To obtain a more rapid reaction with less reactive alcoholic groups, as in ethambutol and pyrid- oxine, TSIM proves an excellent reagent.306 I f increased sensitivity of detection of alcohols is required, the bromo- or iodomethyldimethylsilyl derivatives may prove of value when used in conjunction with electron-cap ture det e~tion.~" The silylation of phenols is similar to that of alcohols, BSA being a popular derivatising reagent.Most reactions occur at room temperature but heating is usually required with compounds exhibiting steric hindrance.281 A, typical derivatisation is illustrated for penta- zocine, which undergoes almost immediate d erivatisation when it is dissolved in dimethyl- formamide containing 10% of BSA.14 Amines With HMDS, primary amines form N-TMS derivatives while second\ary amines do not rea~t.~70 BSA reacts to give N-and "-derivatives with primary amines and N-derivatives with secondary amines, both reactions being non-quantitative.282 The diFficulty in silylating amines with a variety of reagents has been demonstrated264 as well as the instability of the derivatives in the presence of moisture264 and ketones.270 The non-quantitative silylation of amino groups presents difficulties in the complete silylation of catecholamines such as adrenalme and noradrenaline.However, it has been found282 that initial treatment with TSIM de:rivatises all of the hydroxyl groups. The sub- sequent addition of BSA and TMCS gives NN-silylation of primary amino groups but does not attack the secondary amine groups. The mechanism of such a reaction is uncertain but the product is a very stable derivative.B2 Other means of overcoming the inability to silylate catecholamines satisfactorily in one step involve silylation of the hydroxyl groups followed by acylation of the amino groups47,304,305 or carrying out the silylation in a medium of a volatile organic acid such as trifluoroacetic a ~ i d .~ * ~ A mixture of TMSDEA and TSIM is a satisfactory reagent for silylating both the hydroxyl and secondary amino groups of e r g o n ~ v i n e . ~ ~ ~ Although silylation of amine groups is possible, it appears that this is not a favoured reaction in pharmaceutical analysis, other better defined reactions, such as acylation, being preferred. Amines will undergo derivatisation to forin N-TMS derivatives. Carboxylic acids Carboxylic acids, which normally chromatograph poorly, undergo silylation to give volatile TMS esters, which chromatograph well but show instability in the presence of m o i ~ t u r e .~ ~ ~ J ~ 264,272 The TMS group is also readily replaced with a methyl group in the presence of diazo- methane.131 For these reasons, an alternative technique131 can be used to derivatise hydroxy- aromatic acids in which TMS ethers of the methyl esters are formed. However, for acids of low relative molecular mass, TMS derivatives are to be preferred to methyl derivatives as the formers' longer retention times allow better separation from the solvent peak.127 The TMS- benzenecarboxylic acid esters are preferable to the methyl esters as they allow a better separation from each other to be obtained.272 Reaction of the acids with HMDS and TMCS is usually satisfactory for the production of derivatives either at room t e r n p e r a t ~ r e l ~ ~ t ~ ~ ~ J l ~ or with heating if n e ~ e s s a r y .~ 6 ~ , ~ ~ ~ Such a reaction with benzenecarboxylic acids and fenoprofen267 is more efficient than that with the more reactive BSA.272 Unsatisfactory silylation of carboxylic acids using HMDS and TMCS has been reported,lg6 but this was considered to be due to the lack of purity of the reagents. As keto acids give thermally unstable derivatives, conversion of the keto group into its oxime before silylation is necessary in order to give a stable silyl e ~ t e r . 3 ~ ~ If increased sensitivity of detection is required, the use of chloromethylsilyl derivatives with electron-capture detection can provide the means of achieving this.312 However,March, 1978 ANALYSIS OF PHARMACEUTICALS.PART 11. REVIEW 199 unsaturated phenolic acids and trihydroxybenzoic acids give erratic results owing to the thermal instability of the derivatives. Steroids Steroids present major problems when attempts are made to silylate their hydroxyl groups. Firstly, another reactive group such as a ketone, which can react as an enol group, can potentially form two isomeric derivatives. Secondly, there is the possibility of steric hindrance, which prevents the milder reagents from reacting with some of the hydroxyl groups. For example, testosterone has only its sterically unhindered hydroxyl group readily silylated by HMDS and TMCS. More potent reagents, such as BSA and TMCS, also react with the keto grouping, forming two isomeric di-TMS derivatives.Prolonged exposure to this reagent results in a further derivative being formed by the introduction of a trimethylsiloxy group into a position adjacent to the ketone group through addition to the enol-ether structure.275 Such behaviour is typical of ketosteroids when they are exposed to the more potent silylating reagents.275 Although the reactivity of the keto groups can be suppressed to some extent by the choice of an appropriate solvent,276 it is common practice to convert the keto groups into inactive derivatives before silylation. The common derivatives that are formed are the oximes and alkoximes.317-323 In addition to preventing enolisation, methoxime formation also induces thermal stability into steroids containing the 17cc,21-diol-20-one structure.309 It should be noted that silylation of oximes, as opposed to alkoximes, can occur.323 When hydroxyl groups on a steroid molecule suffer steric hindrance, either from groups already present in the molecule or from adjacent hydroxyl groups that have been ~ i l y l a t e d , ~ ~ ~ the milder silylating reagents are ineffectual.Reagents such as BSA plus TMCS275 or TSIM plus TMCS274p275J09 are effective in overcoming steric hindrance. The reactivity of the latter reagent greatly increases with temperature. However, at temperatures greater than 150 "C the influence of BSA and TMCS is no longer n~ticeable.~O~ Due regard must be paid to the fact that some steroids cannot tolerate such elevated temperat~res.~O3 The use of varying degrees of silylation induced by a range of reagents has been considered as a means of structure analysis of steroids303 tert-Butyldimethylsilyl ethers (BDMS ethers) are steroid derivatives that are suitable for structure analysis using mass spectra or for sensitive detection using mass fragmento- graph^.^^^ The BDMS ethers have much simpler mass spectra than the corresponding TMS ethers and almost all show a strong M-57 peak, which is very useful in qualitative These derivatives also exhibit a stability towards moisture not usually found in TMS deriva- t i v e ~ . ~ ~ ~ ~ ~ ~ ~ Thus, they can be extracted with light petroleum from aqueous solution.324 However, filtration through Sephadex provides a simpler means of isolating the derivatives.323 For sensitivity of detection71J7J15 or to achieve a particular separation,71,295 recourse can be made to the halogenated silylating reagents, which have been mentioned above, and electron- capture detection. Use can also be made of these reagents in the gas-chromatographic- mass-spectrometric differentiation of ~ t e r o i d s .~ ~ , ~ ~ l Miscellaneous The thermally unstable members of the benzodiazepine family , lorazepam, oxazepam and tema- zepam, react readily with HMDS and TMCS in pyridine solution to give derivatives that can be sensitively detected by electron capture.153 Following the use of tributylsilyldiethylamine to prepare the tributylsilyl derivatives, complete separation of mono- and dihydroxy canna- binoids can be achieved.280 Attempts have been made to prepare suitable derivatives of various antibiotics, and these have been re~iewed.~~l Where applicable, gas chromatography was considered to be more precise, accurate, convenient and economical than microbiological assays.The anticoagulants warfarin, nicoumalone and phenprocoumon are chromato- graphed successfully following reaction with HMDS and TMCS, although ethyl biscoumacetate and dicoumarol show signs of decomposition even as their TMS derivatives268 Phenobarbi- tone, phenytoin and primidone can be silylated but the derivatives lack ~ t a b i l i t y . ~ ~ Some d o ~ b t l 7 ~ exists regarding the stability of silylated carbamazepine, although it has been considered ~atisfactory.~8~ Even if not yet attracting the attention of pharmaceutical Other groups of compounds have been successfully analysed following silylation.200 NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Analyst, vd.103 analysts, it is of interest to note that a significant number of inorganic anions form silyl derivatives capable of being chromatographed, e.g., phosphate, sulphate, thiosulphate, oxalate and pyrophosphate. 2969325,326 Summary It is obvious that silylation, and in particular trimethylsilylation, is a valuable derivatisa- tion technique that is applicable to a wide range of compounds of pharmaceutical interest. The derivatives are usually easily formed, thermally stable and show good chromatographic properties on a wide range of stationary phases. It is common practice to use phases of low polarity and those which contain hydroxyl grloups, such as the various Carbowaxes, must be avoided as excess of reagent will react with them.TMS groups, despite increasing the relative molecular masses of the compounds, do not greatly affect retention times in some instances and, because of the reduction in hydrogen bonding, actually decrease retention times in other instances. Disadvantages in this approach are usually centred on the need for anhydrous reaction conditions and the need to protect the derivative from moisture. The latter is usually achieved by maintaining the derivative in excess of reagent. As a result of the formation of silica during combustion of derivatives and excess of reagent in the flame-ionisation detector, silica deposits can form in the detector, causing instability of the detector and loss of sensitivity as well as jet b l o ~ k a g e .~ ~ ~ 9 ~ ~ 9 ~ ~ 7 This problein varies in magnitude with the design of the detector.291 Hydrolysis Certain compounds that do not readily chronnatograph can be converted, by hydrolysis, into simpler moieties that can either be chromatographed directly or following some other deriva- tisation step. As a systematic approach to this form of derivatisation is difficult, this section is divided into two parts, dealing first with acid hydrolysis and then with alkaline hydrolysis. Selected examples of each method of hydrolysis are given. Acid Hydrolysis A classical use of acid hydrolysis, which has recently been r e ~ i e w e d , ~ ~ * , ~ ~ ~ is found in the analysis of the 1,4-benzodiazepines. During heating in 6 M hydrochloric acid at about 100 "C for about 1 h, the benzodiazepines are hydrolysed to form substituted benzophenones.After neutralisation of the reaction medium the derivatives are extracted from it with ether and are sensitively detected by electron capture after ~hromatography.3~0 When clonazepam is hydrolysed by 6 M hydrochloric acid, there is a risk of chlorination of the benzophenone, which may be due to impure acid. The use of a mixture of equal volumes of 4 M hydrochloric acid and 2 M sulphuric acid overcomes this problem. The further addition of sodium sulphite may also be beneficial.331 At low pH, potassium chlorazepate is decarboxylated to nordiazepam, which is the substance actually undergoing hydrolysis to give 2-amino-5-chlorobenzophenone.332 However, oxaze- pam also gives this derivative on hydrolysis and this underlines the major disadvantage of this approach-its lack of specificity. This is further emphasised by reference to medazepam and diazepam and to oxazepam and chlordiazepoxide, each pair of which gives the same derivative.Although these drugs may not be administeredl in combination, there is still the possibility of metabolites or degradation products providing misleading results in assays that rely on hydrolysis. It is, therefore, of importance to establish the absence of possible sources of interference before applying this method of analysis. A different type of hydrolysis, also open to similar claims of non-specificity, involves the hydrolysis of a substituted amine to a primary amine followed by acylation of the product to introduce a grouping with electron-capturing properties.This procedure can be applied to practolol, which is deacylated by 1.5 M sulphslric acid. The product is then isolated and converted into its trifluoroacetate, which is detected with high sensitivity by electron ~apture.~33 Sulphonamides are hydrolysed by sulphuric acid to their corresponding amines, which can then be chromatographed However, this does not appear to have gained popu- larity as a method of analysis.March, 1978 ANALYSIS O F PHARMACEUTICALS. PART 11. REVIEW 201 Ethopabate is readily hydrolysed by heating in hydrochloric acid solution to give m- phenetidine.lg6 The latter can then be converted into its 2,4-dinitrophenyl derivative, which, when chromatographed with electron-capture detection, allows the sensitive and selective measurement of ethopabate.However, as most of the ethopabate metabolites also hydrolyse to give m-phenetidine, the method lacks specificity. On heating for 4 min at 100 "C in 1 M hydrochloric acid, the alkaloid pemoline is hydrolysed to 5-phenyl-2,4-dioxooxazolidine. The latter can be chromatographed directly or, if higher sensitivity is required, after methylation with dia~omethane.~~5 As the alteration in structure is minimal the hydrolysis does not greatly alter the selectivity of the assay. Alkaline Hydrolysis Various pharmaceuticals have been subjected to alkaline hydrolyses prior to chromatography and some of these will now be considered.Digoxin, which does not readily chromatograph, is converted quantitatively into digoxigenin by an aqueous pyridine - sodium hydroxide mixture. This derivative chromatographs well on low-polarity columns and does not require silylation of its hydroxyl groups.336 Direct chromatography of meprobamate results in its thermal degradation. This problem is circumvented by hydrolysing the meprobamate with aqueous potassium hydroxide to give 2,2-di(hydroxymethyl)pentane. The latter is then extracted from the reaction medium and assayed by chromatography after silylation of its hydroxyl groups. Silylation is necessary in order to prevent asymmetrical elution of the derivative.337 Meprobamate or mebutamate can also be determined following hydrolysis by 0.2 M sodium methoxide in place of aqueous alkali.338 A lengthy procedure is available for the sensitive measurement of chlorhexidine.339 A 24-h hydrolysis with 25% sodium hydroxide solution converts the chlorhexidine into j5-chloro- aniline.The latter is isolated and, following diazotisation and treatment with potassium iodide - iodine, is converted into 9-chloroiodobenzene. Chromatography of this derivative with the use of electron-capture detection gives a very sensitive means of determining chlor- hexidine .339 Camylofine is rapidly and quantitatively hydrolysed by aqueous methanolic sodium hydroxide into 3-methylbutan-1-01, which is readily chromatographed. This assay is rapid and simple, requiring no isolation of the derivative before chromat~graphy.~~O Measurement of the 2- and 3-chloro isomers of benoxaprofen, found as impurities in the latter, can be achieved after subjecting them to alkaline hydrolysis.The chlorobenzoic acids that are formed are isolated and esterified using diazomethane, after which the methyl chlorobenzoates can be determined after a chromatographic separation.341 However, a chromatographic separation of the methylated benoxaprofen isomers using a liquid-cryst a1 stationary phase may offer a suitable alternative to hydrolysis.342 Alkaline hydrolysis of guanido-containing drugs, which include guanethidine and debriso- quin, results in the formation of corresponding primary amines. The latter are then trifluoro- acetylated and chromatographed with flame-ionisation or mass-spectrometric detection.343 A two-phase hydrolysis system (aqueous alkali - cyclohexane) is necessary in order to prevent degradation of the product, which is extracted into the cyclohexane, by the alkali.Summary As both acid and alkaline hydrolyses reduce the specificity of analyses, by virtue of degrading a molecule to an entity that may also be produced from some other compound, great care must be taken in their application. Unless an analyst is certain that no sources of interference are present in the sample he is analysing, he should not choose hydrolysis as a means of derivatisa- tion. Interaction of Amino and Carbonyl Groups One of the first reports344 concerning the gas chromatography of products formed by the interaction of amines and acetone described two peaks on chromatograms of some sympatho- mimetic amines that were in acetone solution.These were probably peaks due to amines and amine - acetone derivatives. From this developed a simple peak shift technique to aid the identification of amines, particularly amphetamine and e ~ h e d r i n e . ~ ~ ~ - ~ ~ Since then, various reports have appeared in which this reaction has been used in the quantitative analyses of pharmaceuticals.202 Analyst, vol. 103 Derivative formation involves the reaction of primary and some secondary amines with aldehydes or ketones to give Schiff’s bases. Reaction conditions range from standing the amine in solution in the appropriate reagent for a suitable period of time with or without heating, to refluxing in the presence of a dehydrating reagent.348 Additional solvents may or may not be used.By dissolving the amine in, for example, acetone or butanone, derivatisation is achieved at a rate dependent on the amine. Thus, phenylpropanolamine reacts completely in 5 h whereas only a small amount of ephedrine reacts in 24 h.349 The reaction rate also depends on the ketone, acetone giving a more rapid rate than butanone. In general, the yield of derivative in a given time decreases with increasing relative molecular mass of the ketone.347 Amphetamine hydrochloride, when treated with acetone containing triet hylamine to liberate the free amphetamine, gives no reaction unless 10% of water is present. Reaction is then complete in 2 h. This is also true of othei- amines, but the amount of water necessary for reaction varies with the amine.350 In a similar reaction amphetamine can be determined after derivatisation with cyclohexanone in aqueous methanol by its extraction into hexane ready for chr~matography.~~~ Amine hydrochlorides also react with ketones in dimethylformamide solution using potassium carbonate to convert the salt into its free base.270 Reaction occurs in about 3 h at room temperature.Likewise, the hydrochlorides of amino acid methyl esters react with methanolic benzaldehyde,218 and primary amine hydrochlorides will react with methanolic he~ane-2,5-dione,~~~ both in the presence of pyridine. Amines are readily derivatised by pentafluorobenzaldehyde in acetonitrile solution in 1 h at 60 0C.739353 This choice of solvent has the advantage of allowing the subsequent addition of a silylating reagent to the reaction mixture to silylate any free hydroxyl groups prior to ~hrornatography.~~~ The use of halogenated reagents such as pentafluorobenzaldehyde introduces the possibility of using the sensitivity of the electron-capture detector to assay the amine derivatives.Extreme sensitivity of dletection can be obtained by using perfluoro- octanaldehyde but the presence of extraneous peaks precludes its use in quantitative work.202 However, pentafluorobenzaldehyde gives a rnore than acceptable derivative for electron- capturing purposes ; thus, with phenylethylaniine, the derivative formed invokes a response about three times greater than does the heptafluorobutyramide derivative.202 Acetonitrile, a favoured solvent for these reagents, cannot be used with benzaldehyde and heptafluoro- butyraldehyde, with which it gives rise to side-products.202 Pyridine, on the other hand, is suitable only for reactions that do not involve fluorine-containing reagents.202 When excess of pentafluorohenzaldehyde causes an unacceptable background response from the electron-capture detector, the reagent can be removed by washing with a suitable alkali.354 Although pentafluorobenzaldehyde gives adequate results when fresh, it is found that on standing it gives rise to extraneous peaks on ~hromatograms.~55 This problem can be over- come by converting it into its hydrogen sulphite addition compound, which can be recrystal- lised and which is stable on storage.This reagent can be used to derivatise phenylpropanol- amine in aqueous solution in the presence of potassium hydroxide.By shaking with hexane, the derivative is simultaneously formed and extracted ready for ~liromatography.~~5 If ammonia is present in any of the reaction mixtures, excess of reagent must be added to overcome its utilisation by the ammonia to form “hydroamides.”356 The converse of determining amines by derivatisation with various carbonyl compounds is demonstrated by an ingenious assay for benzaldehyde. In this assay the benzaldehyde is made to react with 3-methylthioaniline to give a sulphur-containing derivative, which is selectively detected by a flame-photometric detector.357 The gas-chromatographic determination of ephedrine and pseudoephedrine in admixture presents problems owing to the difficulty in separating them.Although the separation of their acetone derivatives can be achieved, the slow rate of reaction of ephedrine relative to pseudoephedrine, under normal conditions, makes quantitative derivatisation of the former with acetone almost impossible. This trouble can be overcome by the addition of a suitable dehydrating reagent (e.g., silica gel) to the amines in anhydrous acetone and refluxing for 6 h. The derivatives so formed can then be isolated and separated by gas ~hrornatography.3~8 When hydroxyl groups are present in a molecule and are also required to be derivatised, this can be achieved by adding silylating reagent to the reaction rnixt~re.~709353 Acetone, or other ketones, reverses any N-silylation to give eneamines as with, for example, the catecholamines.NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GCMarch, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 203 The nature of derivatives formed following the interaction of amino and carbonyl groups varies. According to elementary considerations, only primary amines should react with aldehydes and ketones to form eneamines: This has been confirmed by mass spectrometry.270 However, the situation is complicated when a secondary amine, such as ephedrine, also when there is a hvdroxyl group suitably positioned R.NH, + O=C.R’, +- R.N=C.R’, + H20 gives a derivative. This situatign arises to allow the formation of an o~azolidine~~~ as with ephedrine : y 3 CH3 The differing rates stereochemistrv. of reaction of ephedrine and pseudoephedrine arise from their different While the ab&e reactions, involving eneamine and oxazolidine formation, are those normally encountered, other reactions do occur and some of these are recorded below.Tertiary amines do not react. Adrenaline, in the presence of pentafluorobenzaldehyde in acetonitrile solution, forms two isomeric substituted tetrahydroisoquinolines.353 Although pentafluorobenzaldehyde reacts with phenylethylamine to form the expected eneamine, its reaction with N-methylphenylethylamine is ~nexpected.~~ This is caused by the elimination of hydrogen fluoride to give two isomeric derivatives with structures ‘0;’ F and laF F F CHO A similar reaction is observed between N-methylphenylethylamine and methyl pentafluoro- benzoate. When pentane-2,4-dione reacts with aniines, it does so to form eneamines that are stabilised by hydrogen bonding21* : CH3 CH3 R I R I I I C- N-CH.COOH H,N-CH.COOH+ iH :H + H20 /--OH \ c=o ‘C=O‘ CH I I CH3 CH3 However, hexane-2,5-dione reacts with primary amines to form highly stable N-substituted 2,5-dimet hylpyrrole~~~~ : c=o / y-42 CH.L c=o \ +HZNR + + H20204 Analyst, VoZ.103 1 , 1,1,5,5,5-Hexafluoropentane-2,4-dione reacts with mono-substituted guanido-containing drugs (e.g., debrisoquin) to form cyclic derivatives.358 This reaction is considered in the section describing the cyclisation of biguanides and guanidines. As expected, N-methyl-/i?-phenylethylamine: does not react with acetone or cyclobutanone but an anomalous reaction does occur with bmoth cyclopentanone and cyclohexanone270 : NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC This reaction would obviously lead to confiising results if attempts were being made to differentiate primary and secondary amines.Some primary amines, such as phentermine and chlorphentermine, do not react with ketones. This is attributed to steric hindrance of, in the examples mentioned, the two a- me t hyl groups. 347 In a reaction similar to that giving oxazolidine formation, cc-amino acids react with dichloro- tetrafluoroacetone to form o~azolidinones~~~~~~~~~0: Such a reaction has the advantage of derivatising both acidic and amino groups in one step. Other reactive groups in the oxazolidinone can subsequently be derivatised. Thus, tyrosine can be converted into its 0-heptafluorobutyryloxazolidinone, which possesses powerful electron- capturing proper tie^.^^ Summary Derivative formation utilising the reaction between amino and carbonyl groups offers advantages and disadvantages that must be considered before deciding on its use, On the credit side are mild reaction conditions, which are easily achieved, a degree of selectivity towards primary amines and a limited number of secondary amines, the facility to produce derivatives capable of being detected with high sensitivity by electron capture and the production of stable derivatives with superior chromatographic behaviour to the original amines.On the debit side are the slowness of reaction, the inability to derivatise other functional groups, such as phenolic, in a one-step reaction as in acylation and the unpredict- ability of some reactions.Despite its advantages, this form of derivatisation has gained only limited attention in comparison with acylation, but it would appear to be well worth greater consideration, Use of NN-Dimethylformamide Dialkyl Acetals Since their i n t r o d ~ c t i o n ~ ~ l 9 ~ ~ ~ as derivatising reagents, NN-dimethylformamide dialkyl acetals (DMF dialkyl acetals) have been applied to a wide range of compounds of pharmaceu- tical interest. Although the various approaches that are used are similar, the end products are not always of the same type, as will be discussed below. However, the derivatisation involves the replacement of an active hydrogen atom by an inactive grouping. The basic reaction involves treating the compound of interest with the appropriate DMF dialkyl acetal either as the pure compound or in a suitable solvent.Generally, the application of heat is necessary for reaction, after which the mixture is chromatographed although prior removal of reagent may be carried out. Various DMF dialkyl acetals are commercially available either as the pure compound or in a suitable solvent. However, it may prove advantageous to re-distil the reagent before use in order to avoid spurious peaks on chr~matograrns.~~~ The alkyl groups commonly encountered are methyl, ethyl, propyl and butyl and the decision of which to use may be dictated by the particular separation required.363March, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 205 Although the DMF dialkyl acetals will react readily with compounds that are soluble in them361~364-366 reaction is slow for poorly soluble compound~.~6~,~67 In the latter instance, the addition of about an equal volume of a mutual solvent greatly expedites reaction.Solvents that are used include pyridine,361,363,367,368 benzene,361 methanol,361 chloroform,361 dichloro- methane,361 tetrahydr~furan,~~~ NN-dimethylf~rmarnide,~~~ a ~ e t o n i t r i l e , ~ ~ ~ ~ ~ ~ ~ acetone368--370 and diethyl ether.371 As the efficiency of derivatisation depends both on the compound undergoing the reaction362J67 and on the solvent, it is necessary to establish which solvent is the most suitable for a given compound if there is no literature information available. In all instances water must be excluded from the reaction m e d i ~ r n .~ ~ ~ , ~ ~ ~ Reaction is usually achieved by heating at 60-100 0C361,362,364-366,370,371 for about 10 min3*1~364-366,370 before chromatography. Reaction is often complete when a solution of the compound in the reagent is ~ b t a i n e d . ~ ~ ~ , ~ ~ ~ With poorly soluble compounds (e.g., aspartic acid) reaction for up to 1 h may be necessary.362 With thiabenda~ole~~~ and sulphonamides368 a reaction time of 40-60 min at 120 "C is advocated but, in comparison with other methods, these appear to be extreme conditions. Although theophylline and probenecid have been described as undergoing quantitative reaction immediately at room temperature,36g it appears more likely that the derivatisation only occurs following injection of the reaction mixture into the chromatograph.370 It is also possible to carry out derivatisation by simultaneously injecting a solution of the compound of interest together with the reagent into a chromato- graph.361 However, this does not appear to be a favoured technique.When reaction mixtures are chromatographed without prior heating, there is a risk that numerous extraneous chromato- graphic peaks will be found that are not produced by the heated mixture.366 After the derivatisation reaction is complete the reaction mixtures can, in many instances, be chromatographed dire~tly.~61-36~,~~6-~71 If the solution is too concentrated for direct chromatography, prior dilution with chloroform, as with barbiturates,364 can be carried out. If the concentration of the derivatised compound is low, the excess of DMF dialkyl acetal can be removed in a stream of nitrogen and the residue dissolved in carbon d i s ~ l p h i d e .~ ~ ~ As the latter invokes a poor response from the flame-ionisation detector, minimum solvent inter- ference is experienced on chromatograms. The derivatives of barbiturates appear to be stable and can have their reaction mixtures diluted with chloroform. However, the glutethimide derivative cannot be treated in this way as it decomposes when d i l ~ t e d . ~ 6 ~ Carbamazepine and cyheptamide form derivatives that are stable for at least 17 h in the reagent but, in carbon disulphide solution, the cyheptamide derivative degrades ra~idly.36~ Sulphonamide derivatives start to decompose gradually after about 6 h.368 With carboxylic acids the DMF dialkyl acetals react to give the corresponding alkyl esters.361,362,369,371 The lower alkyl esters are produced with equal facility, including the tert- butyl esters, which are normally difficult to ~ r e p a r e .~ ~ l Primary amines, on derivatisation, are converted into N-dimethylaminomethylene derivatives, R.N=CH.N. (R')2.362,366,368 Such a reaction has advantages over other derivatisations. It allows the simultaneous derivatisa- tion of carboxylic acid and amino groups in amino acids, which is not readily achieved by many other methods.362 The reaction is easier to carry out than silylation, and it avoids producing a secondary amine, as occurs in acylation, which can interact with the column substrate on chromatography.366 Heterocyclic =NH groups, as in t h e ~ p h y l l i n e , ~ ~ ~ , ~ ~ ~ thiabendazole367 and s~lphonamides,~~8 undergo alkylation.Barbiturates, which might be expected to react similarly, undergo a different reaction with the introduction of two alkyloxy groups a t position 2,364 although some NN-dialkylation may also 0ccur.~*0 Although carbamazepine and cyheptamide form derivatives, their structures are unknown.365 Derivatisation of hydroxyl groups does not appear to occur.362 Little information is available on the stability of the derivatives that are produced. Summary reaction and its relative mildness. bility that more than one product will be produced. requiring the derivative to be isolated from the reaction mixture before chromatography. The advantages of the DMF dialkyl acetals as derivatising reagents lie in their rapid Unlike flash alkylation there is no worry about the possi- This approach also has the benefit of not206 NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Analyst, VoZ.103 Oxime Formation Although the keto group presents relatively few problems in gas chromatography as far as adsorption or decomposition is concerned,372 derivatisation of that group is used to allow other objectives to be attained. The latter can be demonstrated by reference to the field of keto- steroid chromatography, to which this technique has largely been confined. When hydroxyl groups are silylated in the presence of a keto group there is the possibility of enolisation of the keto group. This can result in the formation of two different alcohols, which can silylate to form two separate silyl derivatives. By converting the keto group, prior to the introduction of the silylating reagent, into an oxime derivative the enolisation reaction is prevented.3099373 In this section oxime is considered to include oximes and alkoximes unless otherwise stated.A significant increase in retention time, dependent on the particular oxime formed, results from derivatisation. This increase can be used to aid the separation of ketosteroids from non- ketosteroids and steroids with unreactive ketone groups,374 as well as monoketosteroids from diketosteroids.375 By the introduction of halogenated groups, via oxime formation, the sensitivity of electron- capture detection can be used with the ~ t e r ~ i d ~ .~ ~ ~ s ~ ~ ~ , ~ ~ ~ Also, the introduction of the nitrogen atom in the oxime group allows the sensitivity of the nitrogen alkali flame-ionisation detector to be utilised,377 but this does not yet appear to be in common use. The mass spectra of oximes are used in the structural analyses of steroids and the coupled gas chromatograph - mass spectrometer is of considerable use in this connection.320,322,374 Derivatisation is achieved by dissolving the ketone in a suitable solvent and adding the appropriately substituted hydroxylammonium chloride. Reaction can be achieved, in some instances, at room temperature but in other instances the application of heat is necessary. Following the reaction the solvent, if required, can be removed by evaporation and the residue dissolved in a convenient solvent ready for chromatography.Although 0-methoximes are usually the simplest oximes that are used, use can be made of unsubstituted oximes, which are prepared in a similar way by using hydroxylammonium chloride.318~3~9 0-Methoximes were the first oxjmes to be prepared for the gas-chromatographic analyses of keto~teroids~'~ and have become established derivatives. The reagent employed in their preparation is methoxyammonium chloride and is regularly used in pyridine solu- reaction, which is slow (k, often overnight) even with the application of heat. At room temperature the reaction may, in extreme instances, require up to 100 h.373 Polar steroids containing several hydroxyl groups are normally derivatised without heating.3099317,373 When methoximes do not allow a desired separation to be achieved, recourse can be made to higher alkoximes. Thus, isopentyloximes give a useful separation of mono- and diketosteroids from related hydroxy~teroids.~~~ Both sec-butyl and isopentyl oximes have retention times conveniently placed between those of methoximes and ben~yloximes.~~~ 0-Benzyloximes, which are prepared in the same way as methoximes, have much longer retention times than the methoximes on chrornat~graphy.~~~ This property can be used to separate hydroxysteroids and hydroxyket.oster~ids.~~~~~~~ The fluorinated analogues, 2,3,4,5,6-0-pentafluorobenzyloximes, can be prepared by using the appropriate recently synthesised reagent321,m0 and have the additj onal advantage of possessing strong electron- capturing properties, which make them suitable for use with electron-capture detection.However, the current commercial unavailability of the reagent presents problems. The relative involatility of the pentafluorobenzyloximes requires chromatography to be carried out on short, low-loaded columns.376 A number of w-haloalkoxyamines have been examined as potential reagents for producing the corresponding oximes for use with electron-capture detection.322 Of those examined, only 2-chloroethoxyamine was sufficiently reactive to be considered of practical value. As yet little interest has been shown in this reagent. Derivatisation is normally carried out in pyridine solution using a large excess of the desired reagent. On completion of the reaction various procedures are available for preparing the derivative for chromatography.The simplest involves removal of the pyridine in a stream of nitrogen and the extraction of the oxime from the residue by a suitable solvent such as benzene,372 diethyl ether,378 cyclohexane376 or ethyl acetate.309 Further purification of the derivative can be achieved by subjecting the residue to a water - organic extractant partition fion.309,310,317,372,378 Heat may310W or may not309,317,372,373,378,3'9 be applied in order to achieveMarch, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 207 if necessary.309J76 If hydroxyl groups are to be silylated before chromatography, this can be carried out on the residue left after evaporating the pyridine310J17,320,374,379 or on the pyridine s0lution.~l8 To avoid evaporating the pyridine, the reaction mixture can be diluted with water and the oxime extracted with a convenient solvent such as benzene,317 ethyl acetate319 or he~ane.3~~93~5 The extract can then be washed with a ~ i d , ~ 0 ~ , ~ ~ 7 , ~ ~ 1 J ~ ~ dried and, if necessary, concentrated before chromatography.If silylation is required the extract is taken to dryness first.319,321~3~2 Particular care must be taken to remove all of the reagent in those instances when electron-capture detection is to be used.321 A one-stage oxime formation - silylation reaction gives only limited success and the above procedure is to be preferred.310 It is of interest to note that oximes, as opposed to alkoximes, will undergo silylation by replacement of the oxime proton with a silyl Although most ketosteroids will form oximes in good yield, the 11-ketosteroids resist all but extreme310 attempts at derivatisation owing to steric h i n d r a n ~ e .~ ~ ~ , ~ ~ 0 , ~ ~ ~ , ~ ~ ~ Reaction will occur in the absence of the 18-methyl gro~p.~lO If derivatisation is carried out in ether solu- tion, rather than pyridine, 3-ketosteroids react but 17-ketosteroids do not.372 Typically, of oximes, the problem of syw and anti-isomer formation is encountered in the oxime derivatives of ketosteroids. A C-3 ketone group appears to be susceptible to oxime isomer formation, with 4-en-3-one structures in particular readily giving rise to isomers.317, 374-3769378 However, not all C-3 ketosteroids form isomeric derivatives,317 ,375 e.g., ketones of the 5/3-H series.317 Isomer formation may also be found with C-2,322 C-6373 and C-16 unsub- stituted at C-17375 ketosteroids.C-16, C-17 and C-20 carbonyl groups do not usually appear to form isomers317,319,320,374,375 although isomers of a C-16 ketosteroid have been recorded.317 The separation of syn- and anti-isomers on polar columns presents difficulties in making quantitative interpretations of chromatograms. It is easier, therefore, to chromatograph the isomers on a non-polar column when their peaks are superimposed and can be measured as The oxime derivatives normally show good thermal stability and undergo chromatography without decomposition. However, cortisone and hydrocortisone methoximes break down to give 17-keto analogues as a result of the loss of the 0-methoxime dihydroxyacetone side- chain .372,381 However, the use of a methoxyamine - hexamethyldisilazane reagent gives a thermally stable derivative that can be chr~matographed.~sl Some oximes, e.g., benzaldoxime and salicylaldoxime, undergo thermal dehydration to form the corresponding nitrile~.~s~ Such a reaction depends on the column packing and temperature but at 250 "C can be quanti- tative.Too great an excess of methoxyamine can cause Beckmann fission of the oximes of 17-ketosteroids on top of the chromatographic column unless a silylating reagent (used to silylate hydroxyl groups) is present.310 onea317,376,379 Summary The major importance of oxime formation, as a derivatising reaction, is to protect keto groups in hydroxysteroids from enolisation during silylation of hydroxyl groups.For this there is no alternative reaction, but the stability of most oxime derivative makes this no obstacle. The use of the increased retention times of ketosteroids on derivatisation is of considerable value in aiding their separation from non-ketosteroids. Although heating can be used to accelerate the reaction, care must be taken to ensure the stability of the steroid during heating. The production of syn- and nnti-isomers is an unwanted reaction but the formation of two peaks on chromatograms can be suppressed by the use of non-polar columns. Detracting from the benefits of oxime formation is the slowness of reaction. Carbamate Formation The preparation of derivatives of tertiary amines presents difficulties if they have no other suitable functional groups available for derivatisation.These difficulties have recently been circumvented by utilising the reaction between tertiary amines and chloroformate esters to form carbamates.208 formate ester reacts with a tertiary amine group according to the equation NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Analyst, VoZ. 103 In the presence of a basic catalyst, the highly reactive chlorine of the appropriate chloro- ? -c - '0 R where R, and R3 are usually short-chain alkyl groups. The resultant NN-disubstituted carbamate, which is formed via a quaternary ammonium intermediate,383 can then be chroma- tographed directly using flame-ionisation det e ~ t i o n ~ ~ ~ or following a suitable extraction stage.3859386 The carbamates chromatograph well with no signs of adsorption.The reaction is carried out in h e ~ a n e , ~ ~ ~ heptane384-388 or toluene,389 using a suitable chloro- formate ester in the presence of anhydrous sodium ~arbonate~8~--385,388--390 or another base such as trimeth~lamine~~~ or diis~propylethylamine~~~~ as a catalyst. Heating at about 100 "C for up to 2 h is required in order to complete the reaction. Various chloroformate esters are used in this derivatisation and include methy1,386,388 &hy1,383,385-388 propyl,386y388 i ~ o p r o p y l , ~ ~ ~ bu ty1,3869388 is0 bu t yl ,388 ben~y1,~85,3*~ ,388 nitro- benzy1,385,386 pentaflu~robenzyl,~~~-~~~ trichloro~ethy1389~3go and bromoethyl chloroformates.385 Most of the carbamates that are formed are stable, except the bromoethyl derivative, which is unstable in s0lution.~8~ Isopropyl and benzyl chloroformates react slowly with, for example, Recipavrin, compared with other chloroformates, which makes the former less acceptable.388 Pentafluoroben~yl~8~-~~~ and trichloroethyl c h l o r o f o r n i a t e ~ ~ ~ ~ ~ ~ ~ ~ have the obvious attraction of forming derivatives with potentially good electron-capturing properties.On the other hand, hexafluoroisopropyl and pentafluoropropyl chloroformates give derivatives of poor electron- capturing ability. 390 The presence of a catalyst often shortens the reaction time considerably without affecting the products of the derivati~ation.~~~,~~~~~~~ One role of the catalyst is described as scavenging hydrochloric acid present in the chloroformate.Unless this is done it is considered that the acid inhibits the formation of the intermediate quaternary ammonium compound.383 For a given chloroformate the reaction rate depends on the amine being derivatised388; the converse also applies.389~390 Thus, the catalytic action is minimal in the reaction between pethidine and trichloroethyl chlorof~rmate~~~ but is considerable in the reaction between pethidine and pentafluorobenzyl chlorof ~ r m a t e . ~ ~ ~ Chloroform and ethyl acetate have been examined as alternative solvents to heptane or hexane but gave lower yields of car barn ate^.^^' Benzene is an acceptable solvent in that it gives the same yield of derivative as does h e p t : ~ n e . ~ ~ Initial reports described the carbamate formation as being achieved on a boiling water bath386388 and that lower temperatures reduce the yield of derivative.387 However, NN- dimethyldibenzo [b,f]thiepin-10-methylamine (DTM) undergoes complete reaction in 30 min at 60 OC.383 More recently, the derivatisation of iinipramine and similar compounds was carried out at 105 OC3S4 and derivatisation of pethidine at 125 "C.The reaction rate is governed by the nature of the nitrogen substituents. For compounds that possess substituents larger than NN-dimethyl, e.g., NN-diethyl, the rate of carloamate formation is sl0we.r.~88 On completion of the carbamate formation there are various approaches that can be adopted for measuring the tertiary amine derivative. The reaction mixture can be chromatographed directly,384 or it can first be shaken with aqueous sodium hydroxide to remove excess of reagent before chromatographing the carbamate contained in the organic pha~e.~~5,386 This approach is limited by the need to use flame-ionisation detection, even when the pentafluorobenzyl carbamates are formed, owing to the high background experienced when electron-capture detection is ~ ~ e d .~ 8 ~ , ~ ~ ~ With the latter reagent the background originates from impurities in the pentafluorobenzyl chloroformate (PFBC) , k., pentafluorobenzyl chloride (PFBCl) , pentafluorobenzyl alcohol (PFBA) and dipentafluorobenzyl carbonate (di-PFBCO,) .384 Aqueous alkali washes fail to remove di-PFBCO 3, which causes interference with electron- capture detection. Recently,384 the introduction of an alcoholic potassium hydroxide wash has overcome this problem as it removes all but the PFBCl from the organic extract of the carbamate. Provided that the PFBCl concentration is small, its high volatility prevents it from causing any serious interference.384 The concentrations of PFBA and di-PFBCO, inMarch, 1978 ANALYSIS OF PHARMACEUTICALS.PART 11. REVIEW 209 the PFBC can be prevented from increasing by storing the latter in cool and dry conditions. The use of trichloroethyl chloroformate instead of PFBC has been claimed389~390 to give a much cleaner reaction mixture more suitable for use with electron-capture detection. Instead of determining the carbamate directly by chromatography, its formation can be used as an intermediate stage in the analysis. Thus, after forming the carbamate, the reaction mixture can be evaporated to dryness and subjected to hydrolysis with hydrobromic acid.The secondary amine that is formed is then acylated with heptafluorobutyric anhydride and ~hromatographed,~~~9~~8 using electron-capture detection. Owing to the rapidity of the hydrolysis of the methyl carbamates, these are the preferred dri~atives.,~~ Using 30% hydrogen bromide in acetic acid, hydrolysis is complete in less than 10 min at 100 OC.388 This process is more tedious than that available now using the direct chromatography of penta- fluorobenzyl carbarnates3g4 and the latter will probably be preferable other than in special circumstances. From the equation already given above for the formation of carbamates, it would appear that the by-product R3Cl offers a suitable indirect means of measuring the original amount of tertiary amine.Further, being a halogen-containing compound, it should possess electron- capturing properties. Unfortunately, as R, is often a simple alkyl group such as methyl, this is not always a practical proposition. Nevertheless, one such determination has been reported for DTM.,S3 The two N-methyl groups of the latter are removed by ethyl chloroformate to form dimethyl carbarnate, leaving the remainder of the molecule to form dibenzo [b,fJ thiepin- 10-methyl chloride. As this molecule is neutral it can readily be separated from excess of reagent and sensitively measured using electron-capture detection. Such a reaction is essen- tially one of deamination of the major part of the molecule.The above reaction illustrates one of the major problems in this technique, namely the inability, as yet, to predict which substituent of the nitrogen atom will be eliminated during carbamate formation. There are other instances in which deamination rather than carba- mate formation occurs, e.g., with methad0ne,~~~9~~8 p r o m e t h a ~ i n e ~ ~ ~ ~ ~ ~ ~ and compounds that contain a piperazine ri11g.~90 The low yield~38~9388 obtained during some derivatisations suggest that this also may be a side-reaction that occurs together with carbamate formation. A further side-product, an olefin386-388 formed by deamination, is produced during carbamate formation but in amounts that are sufficiently small and consistent not to detract from quantitative analyses.,g6 With PFBC side-product formation occurs to a total extent of about 5-10%.386 It should be noted that, if the secondary amine corresponding to the tertiary amine is present, they may both form the same carbamate, which would give anomalous results, e.g., amitriptyline and nortriptyline388 and pethidine and n~rpethidine.~~~ This is not always true, as can be illustrated by DTM and ~ o ~ - D T M .~ ~ ~ Summary Although carbamate formation is basically a useful reaction for aiding the sensitive detection of tertiary amines, its use is, as yet, in its infancy and further work is required to determine how widely applicable it will become. Pyrolysis Although the term pyrolysis is often coupled with qualitative analytical techniques, many instances of quantitative analyses of pharmaceuticals utilising this approach are recorded.Pyrolysis may be fortuitous or be deliberately induced. With the former it is only by the application of techniques such as gas chromatography - mass spectrometry that such a reaction can be confirmed, although a significant departure from an expected retention time may indicate that such a reaction has occurred. Induced pyrolysis involves the introduction of a sample into a chromatograph, with a suitably elevated injection port temperature, where pyrolysis occurs. In this section the application of pyrolysis to similar groups of compounds and to single “odd” compounds is considered. 1,4- Benzodiazepines A number of benzodiazepines undergo quantitative elimination of water, at the tempera- tures required for chromatography, to form quinazolinecarboxaldehydes, and it is these compounds which are actually detected and determined.Drugs that are involved include210 Analyst, Val. 103 oxazepam,189,391~392 l o r a ~ e p a m , ~ ~ ~ o-chlorooxazepam189 and nitroo~azepam.~~g N-Methyloxa- zepam is, however, fairly stable under these c0nditions.l8~ On chromatography ketazolam undergoes a different pyrolytic reaction to form diazepam.394 This approach is acceptable provided that the analyst is sure that there is no possibility of two drugs giving the same derivative. Obviously, ketazolam cannot be determined in the presence of diazepam, an analysis that is further complicated by the fact that diazepam is a precursor of k e t a ~ o l a m . ~ ~ ~ NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Quaternary Ammonium Compounds Quaternary ammonium compounds cannot be chromatographed intact but, by inducing pyrolysis, simpler entities are produced, which can be used as a measurement of the original compound.The products usually determined are tertiary amines and the actual products of pyrolysis relate to the original quaternary ammonium i.e. : Quaternary ammonium com9ound Pyrolysis products Alkylbenzyldimethyl . . .. . . Benzyl chloride Dimethylalkylamine Methylbenzylalkylamine Methyl chloride Methyl chloride Methyl chloride Dialkyldimethyl . . .. . . . . Methyldialkylamine Alkyltrimethyl . . .. . . . . Dimethylalkylamine This approach is applicable to neostigmine and pyridostigmine bromides but not to edro- phonium chloride, isopropamide iodide and meperizolate bromide, which give more than the one expected peak on chroma tog ram^.^^^ Using very high pyrolysis temperatures (730 "C) , acetylcholine and choline can also be accurately determined.397 Thiazinamium methyl- sulphate can be determined after conversion into its iodide, which is then pyrolysed at 300 "C to give promethazine.The latter can then be chromatographed and is sensitively detected by the alkali flame-ionisation detect0r.~O8 A very sensitive determination of neostigminie and pyridostigmine is achieved by measuring the alkyl halide, using electron-c.apture detection, rather than the tertiary amine. The drugs are converted into their iodides, which are then quantitatively pyrolysed a t about 320 "C to give methyl iodide, which is determined with great sensitivity.399 An extension of this technique to tertiary amines is made possible through a Hofmann degradation reaction.The tertiary amines are converted into their N-methyl iodides by their reaction with methyl iodide. Following treatment of these products with moist silver oxide, the corresponding quaternary ammonium hyclroxides are produ~ed.~OO-~O~ Pyrolysis of the latter compounds at about 300 "C results in the formation of substituted ethylenes following the elimination of N-methyl secondary amines. In place of silver oxide, treatment with a suitable anion-exchange resin can be used,397 and Hofmann degradation of benzalkonium chlorides using aqueous potassium hydroxide has been described in earlier works as giving superior results to silver oxide.403 9 404 This approach can readily be used as an aid to identifying tertiary amines, such as amitripty- line, by the peak shift techniq~e~8~,~00 as other methods of derivative formation for tertiary amines are more complex. However, the method is applicable to quantitative analyses as with butyrophenones, e.g., droperidol, haloperidol and penflurid01.~~~ The resultant deriva- tives chromatograph with shorter retention times and sharper peaks than do the parent drugs. Although this approach is applicable to choline, it is unsatisfactory for acetyl~holine.~~7 This derivatisation is simpler than carbamate formation but as yet has received scant attention. Compounds with sterically hindered nitrogen atoms (e.g., benzphetamines) do not react.401 Some drugs, such as the phenothiazines, can give two derivatives which may be isomers.401 Sulphonylureas Sulphonylureas are thermally unstable and c,annot be chromatographed intact.Tolazamide can be quantitatively determined following pyrolysis at 236 "C to give p-toluenesulphonamide, There are some drawbacks that are already known.March, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 21 1 which chromatographs to give a sharp ~ e a k . ~ O ~ However, the efficiency of the conversion is only about 30%. In an attempt to chromatograph sulphonylureas successfully, they were first methylated and the resultant products chromatographed to give single peaks, which were attributed to the methyl derivative~.~S However, although the methyl derivatives are first formed, they under- go quantitative pyrolysis at 285 "C to form N-methyls~lphonamides.~~ Methylation can be achieved by using dimethyl sulphate in a base-catalysed r e a c t i ~ n ~ ~ s ~ ~ ~ or by employing a flash alkylation procedure.405 The latter approach avoids the need to isolate the methyl derivative, which is necessary in the former method.Methylation with dimethyl sulphate gives the N- methylsulphonamide derivative^^^ as only one N-hydrogen is available for methylation. Flash methylation gives the NN-dimethylsulphonamide derivatives of tolbutamide, tolaz- amide and chlorpropamide. This is attributed to pyrolysis of the sulphonylureas to the sulphonamides, which are then NN-dime t h ylat ed. 406 Although pyrolysis coupled with alkylation offers a means of assaying sulphonylureas, there is one disadvantage in that, for example, tolbutamide and tolazamide give the same derivative and each can be assayed only in the absence of the other.Similarly, any impurities in the drug being assayed will give the same derivative as the drug if both contain the same sulphonamide group. This problem can be overcome by utilising a double derivatisation approach in which the methylated sulphonylurea is acylated to give an N-methyl-"-per- fluoroacyl derivative, which is thermally stable and consequently chromatographs intact .1169407 Miscellaneous Reactions Floropipamide, which contains an amide group, readily undergoes the well known thermal dehydration of the amide to give a nitrile, which is the compound actually measured.lS9 When introduced into a chromatograph with an injection port at 310 "C, lasalocid undergoes immediate quantitative pyrolysis to produce a ketone and two dihydronaphthol~.~~~ By measuring the sharp peak produced by the ketone an accurate analysis of the antibiotic can be made. Various ergot alkaloids, which do not react favourably to chromatography, can be quantita- tively pyrolysed under catalytic conditions in a metal vapouriser at 235 0C.409 Although the exact nature of the derivatives is uncertain they give rise to chromatographic peaks that can be used in the quantitative analyses of the alkaloids.When menadione sodium hydrogen sulphite is pyrolysed in an injector port heated to 250" C, it is converted into menadi~ne.~lO By utilising this approach, decomposition of the menadione during chloroform extraction from aqueous solution can be avoided.Buformin and phenformin, when pyrolysed at 276 "C as their hydrochlorides in pyridine solution, form substituted 1,3,5-triazines whose chromatographic peaks can be used to assay the original drug.411 Although buformin gives two products a suitable choice of chromato- graphic conditions allows them to be assayed as if they were a single compound. Summary Pyrolysis is a derivatisation technique that requires a prior knowledge of its results before it can be applied successfully to the determination of a drug. Its use requires a careful appraisal of the reaction conditions if these are unknown. Generally, this approach should not be attempted without prior knowledge of the compound's behaviour on pyrolysis or unless it falls into one of the classes of compounds for which such information is available.Reduction Reduction has been used to a limited extent as a derivatisation technique in pharmaceutical analyses. Three classes of compounds to which it can be applied will be considered. Esters Esters are not readily derivatised, but one of the simplest approaches, reduction to alcohols, introduces the possibility of varied subsequent derivatisation through the hydroxyl groups. Although direct chromatography of the alcohols produced by the reduction of pentapiperide methylsulphate allows the determination of the two isomers of the latter,412 it is the subsequent212 Analyst, VoZ. 103 acylation of the alcohol group with a suitable electron-capturing grouping that is usually de~ired.~OY413,414 By this means sensitive and selective detection of esters is achieved.Rapid reduction under mild conditions is readily carried out using aluminium lithium hydride. The latter can be added as a solid to a solution of the drug in tetrahydrofuran412 or as a saturated solution in diethyl ether to the drug in c y c l ~ h e x a n e . ~ ~ , ~ ~ ~ , ~ ~ ~ After standing for a short time, excess of reagent is removed by its reaction with water and the alcohol is acylated with, for example, pentafluoropropionic anhydride before chromatography with electron-capture This approach can be applied successfully to the determination of atr0pine,~l3 co~aine,~0,~~~1414 ind0methacin,~l3 methylphenidate413 and benz~ylecgonine.~~~ The simplicity of the alu- minium lithium hydride reduction, compared with the more exacting transesterification pro- cedures,206 makes this technique attractive, especially if sensitivity of detection is required.Other compounds, such as amines, are unaffected by the reduction and can be determined concurrently as acyl derivative~.~O?~~~ NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Ketones and Aldehydes These compounds, such as methadone, haloperidol and diethylpropion, are readily reduced by sodium borohydride to the corresponding alcohols, which can be chr~matographed.~~ Although there appears to be no record of such a reaction, there is no reason why the alcohols produced by the reduction could not be ,acylated as has been described above for alcohols produced by the reduction of esters.The reduction is readily achieved in about 10 min following the addition of solid sodium borohydride to an ethanolic solution of the drug. Unsaturated Compounds Such compounds can be reduced to their saturated analogues by catalytic hydrogenation, as in the case of vitamin A, and its acetate.415 'The reaction occurs readily overnight at room temperature in the presence of a platinum oxide catalyst. The only problem encountered in this example is the partial reduction of the alcohol group of vitamin A, to an alkyl group. In quantitative analyses, the sum of the areas of the peaks due to the alcohol and the alkyl derivative is taken as a measure of the vitamin A,. Reduction of vitamin A, gives products that are stable to the effects of light or oxygen, unlike the parent compound.Summary Reduction, as a derivatisation technique, appears to have much to offer for the indirect preparation of electron-capturing derivatives of esters, ketones and aldehydes. However, this potential has still to be applied to a wider range of compounds to see if the promise will hold. Oxidation One group that has received much attention is the substituted diphenylmethanes, which comprise a wide range of important drugs. However, the technique is applicable to various other drugs and amitriptyline in particular has attracted much interest. Oxidation is used to produce derivatives from a range of drugs. Substituted Diphenylmethanes These compounds are oxidised by a range of oxidising agents to give benzophenones or similar ketones that often possess more desirable chromatographic properties than the parent drug.In alkaline permanganate, brompheniramine and chlorpheniramine are oxidised in 15 min at 100" C to give ketones, which are then readily extracted into isooctane ready for chromato- gra~hy.~16 Although the yields are not quantitative they allow quantitative measurements to be made by virtue of their reproducibility. However, for compounds that require a longer period of oxidation this approach leads to decomposition of the derivative by the reaction mixture. This difficulty can be circumvented by adding, say, heptane to the reaction mixture before oxidation. Then, as the derivative is formed, it is continuously removedMarch, 1978 ANALYSIS OF PHARMACEUTICALS.PART 11. REVIEW 213 from the aqueous oxidation phase into the organic phase, where it is less susceptible to degradation. 417 Acidic (e.g., phenytoin), quaternary ammonium (e.g., emepronium) and non-extractable compounds give good yields of derivatives in the two-phase oxidation system. However, this is not so with basic (e.g., amines) or neutral (e.g., amides and esters) drugs, which can be extracted into the organic phase and consequently undergo slow oxidation.417 With amines and some neutral drugs the single-phase system, first described, must be used. The neutral drugs to which this applies include benzhydrol and analogues, with diphenhydramine and diphenylpyraline being examples of amines. With amides and esters a prior alkaline hydroly- sis, to produce the corresponding acids, allows the two-phase oxidation to be sequentially applied.In general, the concentration of permanganate (typically 0.6%) and the amount of drug used are not critical. However, the concentration of sodium hydroxide solution should be at least 2.5 M for optimum oxidation. Yields of derivatives vary from poor with quaternary ammo- nium drugs to quantitative for phenytoin, benzhydrol and br~modiphenhydramine.~~~ Alkaline permanganate gives superior oxidation yields from diphenylmethyleneamino com- pounds (73-92y0) than does acidic dichromate and is to be preferred.41s Acidic dichromate is a more efficient oxidant for diphenylmethylene ethers (e.g., diphenhy- dramine) than alkaline permanganate and has been investigated as a suitable oxidant.418 It is used in the same way as described above for alkaline permanganate in a two-phase system.418 Yields are usually greater than goyo, with chlorphenoxamine (89%) and embramine (71%) proving exceptions.Basic drugs are retained in the acidic aqueous phase and can be oxidised in a two-phase system, unlike the oxidation by alkaline permanagnate. An alternative to aqueous systems for oxidation is a non-aqueous system consisting of 0.02-0.04 mol of chromium trioxide in 1 1 of acetic acid containing 6% m/V of sulphuric acid.419 The oxidation is carried out at room temperature, after which the benzophenones are extracted into hexane from the basified reaction medium. Yields, which depend on the reagent batch, vary but typical values are diphenylalkanes 85-S9y0, benzhydrols >97y0, benzhydryl ethers 92-98% and diphenylalkylamines 50-80y0.Normally anhydrous reaction conditions are essential and the presence of sulphuric acid is required for complete reaction.419 However, diphenadione and chlorophacinone can be determined following oxidation by a mixture containing 1.5 g of chromium trioxide in 1 ml of water plus 59 ml of acetic acid.420 After heating under reflux for 30 min the derivatives are extracted into hexane ready for chroma- Chromic acid gives very low yields with some compounds to which an ester or amide group is attached through the a-carbon at0m.4~1 This can be overcome by using the two-phase alkaline permanganate system with the inherent problems discussed above. An alternative is to use a two-phase acidic barium peroxide reagent which retains drugs, which are usually basic, in the aqueous phase until they are oxidised, when they pass into a heptane layer.421 The oxidation involves heating a solution of the drug in sulphuric acid in the presence of barium peroxide and under a layer of heptane.The concentration of acid and the period of heating depend on the drug. Yields from the substituted diphenylmethanes are variable with only compounds that have an amino group attached to the a-carbon atom giving anything approaching quantitative yields, e.g., meclozine and chlorocyclizine. However, the precision is good and allows the method to be used. Thus, methadone, which does not chromatograph well in small amounts and is not sensitively detected by flame-ionisation detection, can be sensitively determined, after oxidation, by electron-capture detection.422 Under the conditions used a constant yield of 79.1% of the appropriate benzophenone is formed.If alkaline permanganate is used instead of barium peroxide, degradation of the derivative is incurred. Chromic acid does not produce a benzophenone from methadone. tography: Amitriptyline Amitriptyline, a tertiary amine, can be successfully chromatographed following oxidation and several reports on this topic have been p ~ b l i s h e d . ~ ~ ~ ~ ~ 6 On oxidation with potassium permanganate, amitriptyline is converted into anthraquinone, which possesses good chromato- graphic p r o p e r t i e ~ ~ ~ ~ ~ ~ ~ 6 and is sensitively detected by electron c a p t ~ r e .~ ~ ~ - ~ ~ s The method is214 Analyst, VOZ. 103 not specific as other similar dibenzocycloheptenic compounds react in the same way. Also, amitriptyline undergoes autoxidation to dibenzosuberone, which could give the same deriva- t i ~ e . ~ ~ ~ Alkaline permanganate is, however, considered unsuitable as an oxidising medium in this instance. It gives low yields4N9426 (e.g., 44%) that are dependent on the alkali concentration426 and a two-phase system is necessary in order to prevent degradation of the anthraq~inone.4~6 This results in extraction of amitriptyline from the aqueous phase, with an inevitable increase in reaction time.426 Although chromic acid and barium peroxide, both in acidic media, might be expected to give an improved performance, this is not so, with yields of less than 20% being obtained.426 Acid cerium(1V) sulphate has proved the only acceptable oxidant although yields of anthraquinone are low (56%).424-426 A suitable oxidising medium is 5.5 M sulphuric acid containing 25 mg ml-l of cerium(1V) s ~ l p h a t e : .~ ~ ~ ~ ~ ~ ~ In a two-phase system with heptane, amitriptyline is oxidised under reflux conditiclns in 25 min when the two phases are stirred vigorously during the reaction.424 However, the reaction is more rapid (5 min) in the absence of h e ~ t a n e , ~ ~ ~ but temperature, acid concentration and length of oxidation are critica1.4259426 NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC Phenelzine Phenelzine, which is unstable under basic conditions and cannot be extracted directly from aqueous solution, is readily determined following its oxidation by potassium iodate to 2- ~henylethanol.~~' The oxidation occurs very rapidly at room temperature using potassium iodate in dilute sulphuric acid as the oxidant.Although the yield is low (less than 60%) the selectivity of the method is high and allows th,e sensitive measurement of phenelzine. Reac- tion conditions are not critical. Citrate Citrate is not readily derivatised but a simple oxidation - bromination reaction followed by alkaline hydrolysis converts it into bromoform, which can be detected with very high sensi- tivity using electron-capture detection.428 The first stage of the reaction is achieved, at room temperature, using potassium bromide - sulphu.ric acid - potassium pennanganate reagent and is complete in 10 min.After extracting the product into heptane, the former is converted into bromoform by shaking for 30 s with 1 M sodiu:m hydroxide solution. The assay is reasonably selective with interference being encountered only from 3-0x0- pentanedioic acid and, at low levels of citrate, by organomercury salts in the potassium permanganate reagent. Mandelic Acid Mandelic acid undergoes ready oxidation whLen steam distilled from acidic potassium perio- date solution.429 The resulting product, benzaldehyde, is then extracted from the distillate using ether and chromatographed. No interference is experienced from structurally related compounds and the need to isolate mandelic acid before its derivatisation is avoided. YC-93 This compound undergoes partial oxidation of its 1,4-dihydropyridine ring to a pyridine ring at temperatures necessary for its chromatography.By using a sodium nitrite - hydrochloric acid oxidation prior to chromatography, quantitative oxidation is obtained leading to quanti- tative analysis of the Summary Various advantages accrue following oxidation of the substituted diphenylmethanes. The ketonic products usually possess superior chromatographic properties to the original drug4199 4219422 and can be chromatographed at much lower temperat~res.4169~~~ This is an advantage when using a tritium electron-capture detector of low maximum operating temperature.416 Sensitive detection using electron capture can be a c h i e ~ e d ~ l ~ - ~ ~ ~ as the derivatives possess electron-capturing properties similar to those of heptafluorobutyryl derivatives.419 As tertiary amines and quaternary ammonium compounds do not readily allow themselves to be derivatised, oxidation provides a relatively simple means of derivatisation.421March, 1978 ANALYSIS OF PHARMACEUTICALS.PART 11. REVIEW 215 The major disadvantage of this oxidation is that more than one drug can give the same benzophenone on oxidation, which leads to a loss of selectivity in an analysis.419 However, such a problem can be recognised by using qualitative analysis if the nature of the sample is unknown. Likewise, although oxidation of amitriptyline gives a derivative with desirable properties, there is a resultant loss of specificity with nortriptyline, protriptyline and cyproheptadine, all of which give rise to anthraquinone on oxidation.424+425 Care must, therefore, be exercised in the application of this method.For other compounds a detailed knowledge of their oxidation chemistry is required before this form of derivatisation can be carried out on them. Isothiocyanate and Isocyanate Formation Isothiocyante Formation This is a seldom used technique that results from the treatment of primary amines with carbon disulphide. Secondary amines form thiocarbamic acids, which can readily be separated from isothiocyanates by a buffer wash.431 Derivative formation is achieved by treating the free amine in ethyl acetate solution with carbon disulphide. Reaction occurs in about 30 min at room t e m ~ e r a t u r e . ~ ~ ~ , ~ ~ ~ Before chromatography the reaction medium is evaporated and the residue dissolved in ethyl a ~ e t a t e .~ ~ ~ ~ ~ ~ ~ Amphetamine can be derivatised by direct extraction from aqueous alkali by carbon disulphide followed by chromatography of the e~tract.43~ If necessary, any hydroxyl alternatively, acylation can be Although compounds such as amphetamine, tranylcypromine and various tryptamines react as expected,434 an anomalous reaction is found with phenylpr~panolamine.~~~ The latter reacts to form 4-methyl-5-phenyloxazolidine-2-thione in a reaction parallel to that which it undergoes with acetone : Tertiary amines do not react. groups in the derivatised molecule are readily silylated before c h r o m a t ~ g r a p h y ~ ~ ~ ~ ~ ~ ~ or, L S Summary Isothiocyanate derivatives of primary amines have several advantages over their silyl counterpart~~~l: only one derivative can form as opposed to the possibilities of N - or NN- silylation ; isothiocyanates are more volatile than the corresponding silyl derivatives ; and the relatively small increase in relative molecular mass allows amines of high relative molecular mass to be examined by mass spectrometers of low a.m.u.limits. Additionally, the high solvent front often associated with silylation mixtures is avoided. Although not reported, the use of a flame-photometric detector combined with the isothiocyanates offers interesting possibilities for the selective detection of primary amines. Isocyanate Formation Isocyanate formation (Lossen rearrangement) allows hydroxamic acids, which are thermally unstable, to be readily chromatographed.The hydroxamic acid is first a~etylated~~s or ~ i l y l a t e d ~ ~ ~ before being introduced into a chromatograph with a suitably elevated injector temperature. Quantitative conversion of the hydroxamic acid derivative into its isocyanate analogue then occurs. Although with, for example, bufexamac the isocyanate conversion occurs without silylation, the rearrangement is slow and in~omplete.~~7 After silylation the desired conversion is achieved as required. Summary This approach is important in that it allows a small number of drugs possessing the hydrox- amic acid group to be chromatographed. However, owing to the limited information available, this reaction requires investigation before it is applied to compounds for which few data are available.216 NICHOLSON DERIVATIVE FORMATION IN THE QUANTITATIVE GC Analyst, VOl.103 Cyclisation of Biguanides and Guanidines The gas-chromatographic properties of the biguanides can be greatly improved by their conversion into cyclic 1,3,5-triazines. This atpproach can be applied to phenf0rmin,~38-~~0 buf omin4399440 and metf ~ r m i n . ~ ~ ~ 9440 The drug is treated with a suitable organic acid anhydride in a solvent such as dichloro- methane439s440 or in the absence of a solvent.438 Reaction occurs rapidly and is complete by the time the reactants are evaporated off either in Z U Z C U O ~ ~ ~ or at 50 0C.439 If the derivative is to be chromatographed using flame-ionisation detection, the residue can be taken up in, say, acetone ready for analysis.438 However, for electron-capture detection, a distribution between aqueous alkali and an organic extractant serves to separate the derivative from the acid anh~dride.~~9 A typical reaction can be demonstrated by considering phenformin : I CCI F2 PI1 en f or in i n 1,3,5-Triazine A range of halogenated organic acid anhydrides can be used for the cyclisation reaction with equal facility.A yield of 50-90% of theory is to be expected, depending on the biguanide.440 When trifluoroacetic anhydride is used to derivatise phenformin it should be freshly distilled and stored in a vacuum de~iccator.~~8 If this is not done a second product, produced by a side- reaction, is detected. Storage of the derivative for up to 24 h in contact with the reagent also leads to the formation of this side-product.Monochlorodifluoroacetic anhydride is to be preferred if easy identification of the derivative by mass spectrometry, using the single chlorine atom, is required.439 The guanidines debrisoquin and guanethidine, which are chemically similar to the biguanides, also f o m cyclic derivatives. Derivatisation is achieved by refluxing an aqueous solution of the drug at pH 9 with a benzene solution of hexafluoroacetylacetone. After washing the benzene solution free from excess of reagent with alkali it is then subjected to chromatography in conjunction with electron-~apture~~~ or mass-spectr~metric~~~ detection. The reaction with debrisoquin is NH Debrisoquin If acetylacetone is used the derivative is not halogenated but high sensitivity of detection can still be obtained by using a nitrogen detector, as with debrisoquin and guanoxan.442 Summary Although limited work has been reported or1 cyclisation reactions, they do offer a means of chromatographing compounds that would otherwise present difficulties.The derivatives, in many instances, also possess strong electron-capturing properties that can be utilised as a means of achieving sensitive detection. As far as the biguanides are concerned, chemical cyclisation offers a more satisfactory and widely applicable approach than does pyrolytic cy~lisation.~~3March, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 217 Phosphorus-containing Derivatives A number of derivatives that contain phosphorus have been investigated with a view to utilising the sensitivity and selectivity of the alkali flame-ionisation detector (AFID) to aid the analyses of various compounds.However, the use of such derivatives has yet to gain accep- tance in pharmaceutical analyses. Some of the reagents, and the compounds to which they have been applied, will now be considered. (Dimethy1amino)dimethylphosphine This compound (DMADMP) can be used to prepare dimethylphosphinous or dimethyl- phosphinic esters, depending on the conditions used. Thus, oestrone reacts with DMADMP, in anhydrous acetonitrile under nitrogen in 2 h at 90 "C, to give a dimethylphosphinous ester. This derivative is separated from excess of reagent, by thin-layer chromatography, before gas chromatography when an AFID is ~ ~ e d . ~ ~ ~ y ~ ~ 4 If flame-ionisation detection is used this separation is unnecessary but the use of the more sensitive AFID requires complete removal of the reagent.Dimethylphosphinic esters of hydroxysteroids are prepared in a similar manner, but with the reaction mixture being exposed to sufficient oxygen to allow the required derivatives to be The reaction rate depends on the reaction solvent, acetonitrile giving the most rapid rate.445 The yields are about 95%. Although most derivatives are stable to hydroly- sis,444s445 moisture should be excluded from the reaction m i ~ t u r e . ~ ~ 5 As with dimethyl- phosphinous esters, a prior separation of the reagent is necessary if chromatography with an AFID is to be ~ s e d . ~ ~ 5 Many derivatives show good thermal stability but the phosphinic esters of secondary monohydroxyketosteroids show an increasing tendency to eliminate dimethylphosphinic acid as the injection port temperatures are increased above 280 0C.444 2- Chloro- 1,3,2-dioxaphospholanene and 2-chloro-l,3,2-dioxaphosphorinanene These compounds react almost instantaneously with lower aliphatic alcohols in benzene solution, in the presence of triethylamine, to give 2-alkoxy-l,3,2-dioxaphospholanes and 2- alkoxy-l,3,2-dioxaphosphorinanes, re~pectively.~~' No isolation of the derivatives is necessary before chromatography as the reagents are retained on the column and so do not upset the AFID base line.Detection is very sensitive. 0 0'-Dimethyl-a-hydroxyphosphonate This compound reacts readily with carboxylic acids to give the corresponding esters.The reaction is carried out in benzene solution in the presence of pyridine and dicyclohexyl- carbodiimide, the mixture being heated at 50 "C for 1 h.g48 Sensitive detection using an AFID is possible but a valve switching system is necessary in order to bleed off excess of reagent following injection of the reaction mixture. Summary Although phosphorus-containing derivatives of hydroxy compounds and carboxylic acids can be prepared, much basic work remains to be done before they can be accepted as deriva- tives for routine use. The complication incurred by the injection of excess of reagents, when using an AFID, is a major barrier to the ready acceptance of some of the reagents. Amide Formation The formation of amides of carboxylic acids can be used in the determination of the latter. In particular, isomeric acids can be separated following their derivatisation with (-)-a- methylbenzylamine to give the corresponding diastereoisomers. The acids are first either converted into the acid chlorides by reaction with thionyl chloride449 or into an imidazolide intermediate by reaction with 1 ,1'-carbonyldiimidazole.450 The product is then made to react with (-)-a-methylbenzylamine at room temperature to give the required derivative ready for chromatography. Satisfactory separation of the d- and I-isomers of ibuprofen can be achieved by adopting this te~hnique.~SO218 NICHOLSON : DERIVATIVE FORMATION I N THE QUANTITATIVE GC Analyst, Vd.103 Summary Amide formation is a more tedious method flor derivatising acids than some of those already recorded, which does not make it an attractive procedure.However, it does have the advan- tages of allowing the separation of isomeric acids to be carried out and introduces the possi- bility of using the nitrogen detector, although flame-ionisation detection appears to have been satisfactory for the examples given a b ~ ~ e . ~ ~ ~ , ~ ~ ~ ~ Phenylhydrazone Formation Substituted phenylhydrazones are used as derivatives of carbonyl compounds but reports on their application to compounds of pharmaceutical interest are few. Derivatisation is simple, an acidic solution of the appropriate phenylhydrazine being added to a solution of the carbonyl compound. On standing, the derivative is formed and can then be isolated for chro- matography.One of the first reagents to be used was 2,4Ldinitrophenylhydrazine, which gives suitable derivatives for the determination of compouiids of low relative molecular mass. Despite investigations on other reagents, only pentafluorophenylhydrazine has proved usefu1451,452 although a report recording the use of 2,6-dinitrophenylhydrazine is available.453 A number of chloro-substituted phenylhydrazines have been examined for use with ketosteroids but, although the yields were high, the products were ~ n s t a b l e . ~ ~ l When the carbonyl compound is soluble in water, derivative formation is achieved by treating it with a solution of the phenylhydrazine in dilute sulphuric a~id45~9~5~ or dilute hydrochloric a ~ i d . ~ ~ ~ - ~ ~ 8 On standing, usually ~ v e r n i g h t ~ ~ ~ - ~ ~ ~ but it can be for as little as 10 min,454 derivative formation takes place with precipitation.For non-water-soluble com- pounds a solution of the phenylhydrazine in methanol, acidified with acetic acid, is used.4519452 Under these conditions hydrocodone reacts in 2 h at room temperature4s2 but oestrone requires an overnight reaction.451 Detection of 2,4-dinitrophenylhydrazones and pentafluorophenylhydrazones using electron capture is more sensitive than flame-ionisation detection so that it is, therefore, necessary to isolate them from excess of reagent before Chromatography. With those prepared in aqueous solution, and when there are sufficient amounts, the derivative can be separated by filtra- ti0114559457 or ~entrifugation.~~4 When this is not suitable, the derivative can be extracted with chloroform456 or carbon tetrachloride458 and the extract, after washing and drying, concentra- ted to a convenient v0lume.~~6 When flame-ionisation detection is used the final solvent can be ethyl acetate,457 carbon d i s ~ l p h i d e ~ ~ ~ or acetone458 but for electron-capture detection the use of benzene4529457 or h e ~ a n e ~ ~ l is preferable.Derivatives prepared in non-aqueous solution are isolated by benzene452 or hexane451 extraction a.fter the addition of aqueous alkali. If neces- sary, a thin-layer chromatographic stage can be employed prior to gas chromatography if sensitivity is at a premium.451 The chromatography of derivatives requires the use of high column temperatures as the derivatives are relatively involatile.This is particularly true of compounds with a benzene ring to which a carbonyl group is in the a - p o ~ i t i o n . ~ ~ ~ As a result of these high temperatures, some column packings (e.g., SE-30 and F-60) have a short life if used continuously for such analyses457 The injection of excess of reagent also reduces the column life.460 Such problems can be overcome by using on-column stopped-flow injection and short column-to-detector joints.460 Derivative decomposition is also considered to be related to the nature of the support, the amount of liquid coating and the length and material of the column.459 A 1-m stainless-steel column has been recommended as causing minimum decomposition.459 Even with such precautions frequent standardisations are r e q ~ i r e d .~ ~ ~ s ~ ~ O Loss of derivatives by adsorption on to PTFE bushings in syringes is also a factor that apparently influences the reproducibility of chroma tog ram^.^^^ Many derivatives chromatograph to give a double peak on chromatograms, which has been attributed to isomer formation.453,456-458 As with oxime isomers, suitable chromatographic conditions can be employed to make these peaks elute as one, thus permitting the easy quantitative assessment of chromatogram~.45~,~*~~ Although the relative sizes of the double peaks have been described as being affected by the Reports have described the instability of derivatives on chromatography. this has been ref~ted.~58March, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 219 Summary The relative involatility of phenylhydrazone derivatives makes their use in pharmaceutical analyses unattractive.Potential thermal degradation coupled with possible isomer formation detracts further from the technique. Oxime formation appears to be preferable, with the advent of pentafluorobenzyloximes allowing the replacement of the nitro- and fluoro-sub- stituted phenylhydrazones for use with electron-capture detection. The permission of the Head of the Medicines Testing Laboratory, Dr. C. Daglish, to publish this review is acknowledged. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305.306. 307. 308. 309. 310. 311. 312. 313. 314. 315. References * Brittain, G., Am. Lab., 1969, May, 57. Tischio, J. P., J . Pharm. Sci., 1976, 65, 1530. Mason, P. S., and Smith, E. D., J . Gas Chromat., 1966, 4, 398. Sweeley, C. C., Bentley, R., Makita, M., and Wells, W. W., J . Am. Chem. Soc., 1963, 85, 2497. Brobst, K. M., and Lott, C. E., Cereal Chem., 1966, 43, 35, Nash. J. F., Bopp, R. J., and Rubin, A., J . Pharm. Sci., 1971, 60, 1062. Dickert, F. W., J . Chromat., 1972, 64, 355. Janssen, G., and Vanderhaeghe, H., J . Chromat., 1973, 82, 297. Capella, P., and Horning, E. C., Analyt. Chem., 1966, 38, 316. Kagi, D. A., Hildebrand, R. P., and Clarke, B. J., J . Chromat., 1975, 115, 79. Kaufman, M. L., Friedman, S., and Wender, I., Analyt. Chem., 1967, 39, 1011.Janssen, G., and Vanderhaeghe, H., J . Chromat., 1977, 134, 365. Chambaz, E. M., and Horning, E. C., Analyt. Biochem., 1969, 30, 7. Chambaz, E. M., Maume, G., Maume, B., and Horning, E. C., Analyt. Lett., 1968, 1, 749. Van Giessen, B., and Tsuji, K., J . Pharm. Sci., 1971, 60, 1068. Margosis, M., and Tsuji, K., J . Pharm. Sci., 1973, 62, 1836. Sondack, D. L., J . Pharm. Sci., 1974, 63, 584. Barrow, K. D., and QuigIey, F. R., J . Chromat., 1975, 105, 393. Harvey, D. J.. and Paton, W. D. M., J . Chromat., 1975, 109, 73. Klebe, J. F., Finkbeiner, H., and White, D. M., J . Am. Chem. Soc., 1966, 88, 3390. Horning, M. G., Moss, A. M., and Horning, E. C., Biochim. Biophys. A d a , 1967, 148, 597. VandenHeuvel, W. J. A., J . Chromat., 1968, 36, 354. Kupferberg, J. J., J .Pharm. Sci., 1972, 61, 284. Windheuser, J. J., Sutter, J . L., and Auen, E., J . Pharm. Sci., 1972, 61, 301. Fish, F., and Wilson, W. D. C., J . Chromat., 1969, 40, 164. Roncucci, R., Simon, M.-J., and Lambelin, G., J . Chromat., 1971, 57, 410. Lovelady, H. G., and Foster, L. L., J . Chromat., 1975, 108, 43. Haeberer, A. F., and Chortyk, 0. T., J . Agric. Fd Chern., 1974, 22, 1135. Street, H. V., J . Chromat., 1969, 41, 358. Margosis, M., J. Chromat. Sci., 1974, 12, 549. Piekos, R., Osmialowski, K., Kobylezyk, K., and Grzylowski, J., J . Chromat., 1976, 116, 315. Piekos, R., Teodorcyzk, J., Gryzylowski, J., Kobylezyk, K., and Osmialowski, K., J . Chromat., 1976, Stalling, D. L., Gehrke, C. W., and Zumwalt, R. W., Biochem. Biophys. Res. Commun., 1968, 31, 616.Gower, D. B., and Thomas, B. S., J . Chromat., 1968, 36, 338. Wiese, P. M., and Hanson, R. H., Analyt. Chem., 1972, 44, 2393. Kelsey, M. I., Keskiner, A., and Moscatelli, E. A., J . Chromat., 1973, 75, 294. Cohen, J. L., and Brennan, P. B., J . Pharm. Sci., 1973, 62, 572. Ali, S. L., Chromatographia, 1975, 8, 33. Li, H., and Cervoni, P., J . Pharm. Sci., 1976, 65, 1352. Donike, M., J. Chromat., 1969, 42, 103. Donike, M., J. Chromat., 1973, 85, 1. Gleispach, J.. J. Chromat., 1974, 91, 407. Donike, M., J. Chromat., 1975, 115, 591. Donike, M., J. Chromat., 1975, 103, 91. Richard, B. M., Manno, J. E., and Manno, B. R., J . Chromat., 1974, 89, 80. Edlund, D. O., and Alfinsen, J . R., J . Ass. Off. Analyt. Chern., 1970, 53, 287. Rasmussen, K. E., J . Chromat., 1976, 120, 491.Sakauchi, N., and Horning, E. C., Analyt. Lett., 1971, 4, 41. Thenot, J.-P., and Horning, E. C., Analyt. Lett., 1972, 5, 21. Chapman, J. R., and Bailey, E., Analyt. Chem., 1973, 45, 1636. Morita, J., and Montgomery, W. G., J . Chromat., 1976, 123, 454. Bevan, W. I., Haszeldine, R. N., and Young, J . C., Chemy Ind., 1961, 789. Burkinshaw, P. M., Morgan, E. D., and Poole, C. F., J . Chromat., 1977, 132, 548. Morgan, E. D., and Poole, C. F., J . Chromat., 1975, 104, 351. 117, 431. * References 1-261 appeared in Part I.220 NICHOLSON : DERIVATIVE FORMATION IN THE QUANTITATIVE GC 316. Horii. Z., Makita, M., and Tamura, Y., Chemy Ind., 1965, 1494. 317. Horning, M. G., Moss, A. M., and Horning, E. C., Analyt. Biochem., 1968, 22, 284. 318. Nambara, T., Kudo, T., and Ikeda, H., J .Chromat., 1968, 34, 526. 319. Brooks, C. J. W., and Harvey, D. J., Steroids, 1970, 15, 283. 320. Devaux, P. G.. Horning, M. G., Hill, R. M., and Horning, E. C., Analyt. Biochem., 1971, 41, 70. 321. Nambara, T., Kigasawa, K., Iwata, T., and Ibuki, M., J . Chromat., 1975, 114, 81. 322. Nambara, T., Iwata, T., and Kigasawa, K., J . Chromat., 1976, 118, 127. 323. Kelly, R. W., and Taylor, P. L., Analyt. Chem., 1976, 48, 465. 324. Phillipou, G., Bigham, D. A., and Seamark, R. F., Steroids, 1975, 26, 516. 325. Mathews, D. R., Schults, W. D., Guerin, M. R., and Dean, J. A., Analyt. Chem., 1971, 43, 1582. 326. Butts, W. C., and Rainey, W. T., Analyt. Clzenz., 1971, 43, 538. 327. Alexander, R. J., and Garbutt, J. T., Analyt. Chem., 1965, 37, 303.328. Hailey, D. M., J . Chromat., 1974, 98, 527. 329. Clifford, J. M., and Franklin Smyth, W., Analyst, 1974, 99, 241. 330. Beharrell, G. P., Hailey, D. M., and McLaurin, M. K., J . Chromat., 1972, 70, 45. 331. Knop, H. J., Van der Kleijn, E., and Edmunds, L. C., Pharm. Weekbl., 1975, 110, 297. 332. Hoffman, D. J., and Chun, A. H. C., J . Pharm. Sci., 1975, 64, 1668. 333. Desager, J. P., and Harvengt, C., J . Pharm. Pharmac., 1975, 27, 52. 334. Fravolini, A., and Begliomini, A., J . Ass. Off. 14nalyt. Chem., 1969, 52, 767. 335. Van Boven, M., and Daenens, P., J . Chromat., 1977, 134, 415. 336. Kibbe, A. H., and Araujo, 0. E., J . Pharm. Sci., 1973, 62, 1703. 337. Martis, L., and Levy, R. H., J . Pharm. Sci., 1974, 63, 834. 338. Cerri, O., Boll. Chim.-Farm., 1969, 108, 217, 339.Seifert, K., Casagrande, D., and Silberman, H., J . Chromat., 1975, 109, 193. 340. Crombez, E., Van Den Bossche, W., and de Moerloose, P., J . Chromat., 1976, 117, 161. 341. Browner, S. J., Cockerill, A. F., Maidment, R. J., Rackham, D. J., and Snook, G. F., J . Pharm. Sci., 342. Hall, M., and Mallen, D. N. B., J . Chromat., 1976, 118, 268. 343. Hengstmann, J. H., Falkner, I;. C., Watson, J. T., and Oates, J., Analyt. Chem., 1974, 46, 34. 344. Parker, K. D., Fontan, C. R., and Kirk, P. L., Analyt. Chem., 1962, 34, 1345. 345. Beckett, A. H., and Wilkinson, G. R., J . Pharm. Pharmac., 1965, 17, 104s. 346. Beckett, A. H., and Rowland, J., J . Pharm. Pharmac., 1965, 17, 59. 347. Beckett, A. H., Tucker, G. T., and Moffat, A. C., J . Pharm. Pharmac., 1967, 19, 273.348. Yamasaki, K., Tujita, K., Sakamoto, M., Okada, K., Yoshida, M., and Tanaka, O., Chem. Pharm. 349. Brochmann-Hanssen, E., and Svendson, A. B., J . Pharm. Sci., 1962, 51, 938. 350. Beckett, A. H., and Rowland, M., J . Pharm. Pharmac., 1964, 16, 27T. 351. Clark, C. C., J . Ass. Off. Analyt. Chem., 1975, 5'6, 1174. 352. Walle, T., Acta Pharm. Suec., 1968, 5, 353. 353. Moffat, A. C., and Horning, E. C., Biochim. Biophys. Acta, 1970, 222, 248. 354. Hoshika, Y., Analyt, Chem.. 1977, 49, 541. 355. Neelakantan, L., and Kostenbauder, H. B., J . Pharm. Sci., 1976, 65, 740. 356. Hoshika, T., Analyt. Chem., 1976, 48, 1716. 357. Hoshika, T., J , Chromat., 1976, 129, 436. 358. Erdtmansky, P., and Goehl, T. J., Analyt. Clzem., 1975, 47, 750. 359. Husek, P., J.Chromat., 1974, 91, 475. 360. Husek, P., Ergebn. Ex$. Med., 1975, 20, 24. 361. Thenot, J.-P., Horning, E. C., Stafford, M., and Horning, M. G., Analyt. Lett., 1972, 5, 217. 362. Thenot, J.-P., and Horning, E. C., Analyt. Lett., 1972, 5, 519. 363. Ziudema, J., Licht, J. E. C. P. M., Prins, J., and Merkus, F. W. H. M., Pharm. Weekbl., 1976, 111, 570. 364. Venturella, V. S., Gualario, V. M., and Lang, R. E., J . Pharm. Sci., 1973, 62, 662. 365. Perchalski, R. J., and Wilder, B. J., Clin. Chem., 1974, 20, 492. 366. Scoggins, M. W., J . Chromat. Sci., 1975, 12, 146. 367. Tanaka, A., and Fujimoto, Y., J . Chromat., 1976, 117, 149, 368. Nose, N., Kobayashi, S., Hirose, A., and Watanabe, A., J . Chromat., 1976, 123, 167. 369. Arbin, A. A., and Edlund, P.-O., Acta Pharm.Suec., 1974, 11, 249. 370. Bailey, D. G., Davis, H. L., and Johnson, G. E., J . Chromat., 1976, 121, 263. 371. Karashima, D., Takahashi, S., Shigematsu, A., Furukawa, T., Kuhara, T., and Matsumoto, I., Bio- 372. Fales, H. M., and Luukkainen, T., Analyt. Chem., 1965, 37, 955. 373. Morgan, E. D., and Woodbridge, A. P., Chem. Commun., 1971, 475. 374. Devaux, P. G., Horning, M. G., and Horning, M. C., Analyt. Lett., 1971, 4, 151. 375. Baillie, T, A., Brooks, C. J. W., and Horning, E. C., Analyt. Lett., 1972, 5, 351. 376. Koshy, K, T., Kaiser, D. G., and VanDerSlik, A. L., J . Chromat. Sci., 1975, 13, 97. 377. Novotny, M., Makarinec, M. P., Steverink, A. T. G., and Farlow, R., Analyt. Chem., 1976, 48, 468. 378. Allen, J. G., Thomas, G. H., Brooks, C. J. W., and Knights, B.A., Steroids, 1969, 13, 133. 379. Horning, E. C., and Maume, B. F., J . Chromat. Sci., 1969, 7, 411. 380. Youngdale, G. A., J . Pharm. Sci., 1976, 65, 625. 381. Gardiner, W. L., and Horning, E. C., Biochim. Biophys. Acta, 1966, 115, 524. 382. Lohr, L. J., and Warren, R. W., J . Chromat., 1962, 8, 129. 383. Degen, P. H., and Riess, W., J . Chromat., 1973, 85, 53. 384. Hartvig, P., Handl, W., Vessman, J., and Svahn, C. M., Analyt. Chem., 1976, 48, 390. A nabst, Vol. 103 1976, 65, 1305. Bull., Tokyo, 1974, 22, 2898. med. Mass Spectrom., 1976, 3, 41.March, 1978 ANALYSIS OF PHARMACEUTICALS. PART 11. REVIEW 221 385. 386. 387. 388. 389. 390. 391. 392. 393. 394. 395. 396. 397. 398. 399. 400. 401. 402. 403. 404. 405. 406. 407. 408. 409. 410. 411. 412. 413. 414.415. 416. 417. 418. 419. 420. 421. 422. 423. 424. 425. 426. 427. 428. 429. 430. 431. 432. 433. 434. 435. 436. 437. 438. 439. 440. 441. 442. 443. 444. 445. 446. 447. 448. 449. 450. 451. 452. 453. Hartvig, P., and Vessman, J., Analyt. Lett., 1974, 7, 223. Hartvig, P., and Vessman, J., J . Chromat. Sci., 1974, 12, 722. Vessman, J., Hartvig, P.. and Molander, M., Analyt. Lett., 1973, 6, 699. Hartvig, P., and Vessman, J., Acta Pharm. Suec., 1974, 11, 115. Hartvig, P., Karlsson, K.-E., Johansson, L., and Lindberg, C., J . Chromat., 1976, 121, 235. Hartvig, P., Ahnfelt, N. O., and Karlsson, K.-E., Acta Pharm. Suec., 1976, 13, 181. Sadee, W., and Van Der Kleijn, E., J . Pharm. Sci., 1971, 60, 135. Forgione, A., Martelli, P., Marcucci, F., Fanelli, R., Mussini, E., and Jommi, G.C., J . Chromat., 1971, Marcucci, F., Mussini, E., Airoldi, L., Guaitani, A., Garattini, S., J . Pharm. Pharmac., 1972, 24, 63. Weber, D. J., J . Pharm. Sci., 1972, 61, 1797. Metcalfe, L. D., J . Chromat. Sci., 1975, 13, 516. Chan, K., Williams, N. E., Baty, J. D., and Calvey, T. N., J . Chromat., 1976, 120, 349. Szilagyi, P. I. A., Schmidt, D. E., and Green, J. P., Analyt. Chem., 1968, 40, 2009. Jonkman, J . H. G., Wijsbeek, J., Brouwer-De Boer, S. H., Zeeuw, R. A,, Van Bork, L. E., and Orie, Pohlmann, J. L. W., and Cohan, S. L., J . Chromat., 1977, 131, 297. Hucker, H. B., and Miller, J. K., J . Chromat., 1968, 32, 408. Street, H. V., J . Chromat., 1972, 73, 73. Rosenfeld, J., Kawai, M., Rigg, J. R. A., and Khandelwal, J. K., J . Chromat., 1976, 129, 387. Jennings, E. C., and Mitchner, H., J . Pharm. Sci., 1967, 56, 1590. Mitchner, H., and Jennings, E. C., J . Pharm. Sci., 1967, 56, 1595. Wickramasinghe, J . A. F., and Shaw, S. R., J . Pharm. Sci., 1971, 60, 1699. Aggarwal, V., and Sunshine, I., Clin. Chem., 1974, 20, 200. Braselton, W. E., Bransome, E. D., Ashline, H. C., Stewart, J. T., and Honigberg, I. L., Analyt. Westley, J . W., Evans, R. H., and Stempel, A., Analyt. Biochem., 1974, 59, 574. Szepesi, G., and Gazadag, M., J . Chromat., 1976, 122, 479. Winkler, V. W., Regnier, F. E., Yoder, J. M., and Macy, L. R., J . Pharm. Sci., 1972, 61, 1463. Wickramasinghe, J , A. F., and Shaw, S. R., J . Chromat., 1972, 71, 265. Warner, C. R., Prue, D. G., Orzech, C. E., and Kho. B. T., J . Chromat., 1971, 57, 413. Huffman, R., Blake, J. W., Ray, R., Noonan, J., and Murdick, P. W., J . Chromat. Sci., 1974, 12, 382. Javaid, J , I., Dekirmenjian, H., Brunngraber, E. G., and Davis, J. M., J . Chromat., 1975, 110, 141. Fenton, T. W., J . Chromat., 1973, 79, 410. Bruce, R. B., Pitts, J . E., and Pinchbeck, F. M., Analyt. Chem., 1968, 40, 1246. Hartwig, P., Sundin, H., and Vessman, J., Acta Pharm. Suec., 1972, 9, 269. Caddy, B., Fish, F., and Tranter, J., ANalyst, 1974, 99, 555. Vessman, J., Acta Pharm. Suec., 1971, 8, 251. Bullard, R. W., Holguin, G., Peterson, J. E., J . Agric. Fd Chem., 1975, 23, 72. Hartvig, P., and Handl, W., Acta Pharm. Suec., 1975, 12, 349. Hartvig, P., and Naslund, B., J . Chromat., 1975, 111, 347. Bouche, R., J. Pharm. Sci., 1972, 61, 986. Hamilton, H. E., Wallace, J. E., and Blum, K., Analyt. Chern., 1975, 47, 1139. Wallace, J. E., Hamilton, H. E., Coggin, L. K., and Blum, K., Amzlyt. Chem., 1975, 47, 1516. Hartvig, P., Strandberg, S., and Naslund, B., J . Chromat., 1976, 118, 65. Caddy, B., and Stead, A. H., Analyst, 1977, 102, 42. Bjork, R. G., Analyt. Biochem., 1975, 63, 80. Nicholson, J. D., Analyst, 1969, 94, 413. Higuchi, S., Sasaki, H., and Sado, T., J . Chromat., 1975, 110, 301. Narasimhachari, N., and Vouros, P., Analyt. Biochem., 1972, 45, 154. Narasimhachari, N., and Vouros, P., J . Chromat., 1972, 70, 135. Hall, C. R., Cordova, V., and Rieders, F., Pharmacologist, 1965, 7, 148. Narasimhachari, N., and Vouros, P., Biomed. Mass Spectrom., 1974, 1, 367. Santoro, R., Warren, R., and Roberts, G., J . Chromat., 1976, 117, 375. Vagelos, P. R., VandenHeuvel, W. J. A., and Horning, M. G., Analyt. Biochem., 1961, 2, 50, Bell, D., Boreham, D. R., and Martin, B. K., J . Pharm. Sci., 1971, 60, 1368. Mottale, M., and Stewart, C. J., J . Chromat., 1975, 106, 263. Matin, S. B., Karam, J. H., and Forsham, P. H., Analyt. Chem., 1975, 47, 545. Alkalay, D., Volk, J., and Bartlett, M. F., J . Pharm. Sci., 1976, 65, 525. Malcolm, S. L., and Marten, T. R., Analyt. Chem., 1976, 48, 807. Lennard, M. S., Silas, J. H., Smith, A. J., and Tucker, G. T., J . Chromat., 1977, 133, 161. Vogt, W., Fischer, I., and Knedel, M., 2. Analyt. Chem., 1973, 267, 28. Jacob, K., Vogt, W., Knedel, M., and Schafer, W., Biomed. Mass Spectrom., 1976, 3, 64. Vogt, W., Jacob, K., and Knedel, M., J . Chromat. Sci., 1974, 12, 658. Jacob, K., Vogt, W., Fischer, I., and Knedel, M., Tetrahedron Lett., 1975, 1927. Viceanu, R., and Schulz, P., J . Chromat., 1973, 82, 279. Schulz, P., and Vilceanu, R., J . Chromat., 1975, 111, 105. Tosolini, G. P., Moro, E., Forgione, A., Ranghieri, M., and Mandelli, V., J . Pharm. Sci., 1974,63, 1072. Van Giessen, G. J., and Kaiser, D. G., J . Pharm. Sci., 1975, 64, 798. Attal, J., Hendeles, S. M., and Eik-Nes, K. B., Analyt. Biochem., 1967, 20, 394. Barnhart, J. W., and Caldwell, W. J., J . Chromat., 1977, 130, 243. Bassette, R., Brewington, C. R., and Schwartz, D. P., Microchem. J., 1968, 13, 297. 59, 163. N. G. M., J . Pharm. Pharmac., 1975, 27, 849. Chem., 1976, 48, 1386.222 NICHOLSON Galetto, W. G., Kepner, R. E., and Webb, A. D., Analyt. Chem., 1966, 38, 34. Fracchia, M. F., Schuette, F. J., and Mueller, P. K., Envir. S G ~ . Technol., 1967, 1, 915. Leonard, R. E., and Kiefer, J. E., J . Gas Chromat., 1966, 4, 142. Kallio, H., Linko, R. R., and Kaitaranta, J., J . Chromat., 1972, 65, 355. Hoshika, Y., and Takata, Y., J . Chromat., 1976, 120, 379. Pias, J. B., and Gasc6, L., Chromatographia, 11975, 8, 270. Papa, L. J., and Turner, L. P., J . Chromat. Sci., 1972, 10, 744. Received July 14th, 1977 Accepted August 9th, 1977 454. 455. 456. 457. 458. 459. 460.

ISSN:0003-2654    

DOI:10.1039/AN9780300193

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

Afialyst, March, 1978, Vol. 103, fip. 223-226 223 Determination of Metiamide (SKF 92058) by Differential = pu I se Pol a rog ra p hy P. E. Anderson Analytical DeveloFment Laboratory, Smith, Kline and French Laboratories Ltd., Welwyn Garden City, Hertfordshire, AL7 1EY and W. Franklin Smyth Department of Chemistry, Chelsea College, University of London, Manresa Road, London, S W3 6LX Metiamide (SKF 92058) undergoes anodic oxidation a t the dropping-mercury electrode with the formation of mercury salt (s). Differential-pulse polaro- graphic analysis has been carried out on several formulations in 0.1 M sodium hydroxide solution as supporting electrolyte using the peak corresponding to this oxidation process. The effects of degradation products and anti- oxidants on this determination have been investigated in addition to com- paring the polarographic method with one involving high-performance liquid chromatography.Keywords : Metiamide determination; di3erential-pulse potarogvaphy ; anti- oxidants ; degradation products Metiamide (SKF 92058) (I) is a compound developed by Smith, Kline and French which possesses H,-antagonist properties. It is unstable in solution at both extremes of the pH scale, so liquid formulations are manufactured at approximately pH 5. Even at this pH, it is unstable at high temperatures such as may be achieved during sterilisation in an auto- clave, Degradation under these conditions has been proved to occur via cyclisation of the side-chain to produce 2-mercaptothiazoline (11). CH2. S.CH2. CH2. NH. C. NH.CH3 /-7 kNH II S S N e N I It Other experiments have shown that metiamide is also unstable in the presence of oxygen when exposed to bright light.In this instance, the exact mechanism of the degradation is unknown. A complex mixture of products is obtained, including precipitated sulphur, hydrogen sulphide and carbonyl sulphide. Therefore, it has been found necessary to include an antioxidant in liquid formulations, the two under investigation being ascorbic acid and sodium disulphite. In the context of investigating the stability of pharmaceutical formulations, the chosen analytical method should ideally be capable of determining the compound of interest in the presence of excipients and degradation products. This paper reports an investigation of whether metiamide can be determined polaro- graphically in the presence of the potential interferents 2-mercaptothiazoline, ascorbic acid and sodium disulphite (all other excipients were found to be electroinactive and did not interfere), and compares the results with the present method of high-performance liquid chromatography on a strong anion-exchange column.Experiment a1 Apparatus Polarographic curves were recorded in the sampled d.c. and differential-pulse modes with a PAR, Model 174A, analyser in conjunction with a three-electrode cell system. The dropping- mercury electrode had an outflow velocity of 2.571 mg s-l and a drop time of 3.46 s at the potential of the saturated calomel electrode and a mercury head of 55 cm in 1 M potassium chloride solution.224 ANDERSON AND SMYTH: DETERMINATION OF METIAMIDE Analyst, vat.103 Reagents A stock solution of 1 0 - 3 ~ metiamide was prepared in AnalaR methanol and stored in darkness under refrigeration. The following supporting electrolytes were used: 0.1 M hydrochloric acid, 0.1 M phosphate buffer (p1-I 7.4), 0.1 M tetramethylammonium chloride solution (pH 9.0) and 0.1 M sodium hydroxide solution. Stock solutions of 10-3 M 2-mercapto- thiazoline, ascorbic acid and sodium disulphite were also prepared. Experimental Techniques A solution of 1 0 - 4 ~ metiamide in each of the supporting electrolytes was subjected to polarography using both anodic and cathodic scans commencing at -0.6 V, using the sampled d.c. mode of operation. No stable, useful reduction wave was obtained in any of the supporting electrolytes. However, each solution did produce a well defined anodic wave.Of the four supporting electrolytes tried, the wave in 0.1 M sodium hydroxide solution was the most clearly defined and was selected for further analytical study using diff erential-pulse polarography. In all subsequent work using diff erential-pulse polarography, peak heights were measured on recorded polarograms by subtracting the peak height at -0.54 V from that at -0.33 V. The relationship between concentration and peak height was investigated by diluting a 10-3 M stock solution of metiamide in AnalaR methanol appropriately with the same solvent and subjecting the resulting solutions to polarcgraphy. The precision of the techniques was investigated by taking 10.0 ml of the methanolic M stock solution and diluting to 100.0 ml with 0.1 M sodium hydroxide solution.The resulting solution was subjected to polarography in ten 10-ml portions. In this manner, statistical analysis of the ten measured peak heights gave a true indication of the precision of the analytical technique, as random errors caused by dilution errors were eliminated. Results and Discussion Differential-pulse polarography of lov4 M metiamide in the analytically optimum medium of 0.1 M sodium hydroxide solution produced the graph shown in Fig. 1, which corresponds to mercury salt formation, as has been observed for other sulphur-containing compounds containing the structural -NH-C( =S)-NH-. A graph of peak height zleysztus concentration .was plotted in the concentration range 0.029- 0.203 mg ml-l and gave a straight line passing through the origin.The full extent of this linear range was not investigated as the above-mentioned range was adequate for the working 0.6 0.54 0.33 Potent i a I /V Fig. 1. Diff erential-pulse polarogram of metiamide as a M solution in 0.1 in sodium hydroxide. A, Metiamide ; B, ascorbic acid ; C, sodium disulphite ; and D, 2-mercaptothiazoline.March, 1978 BY DIFFERENTIAL-PULSE POLAROGRAPHY 225 concentrations encountered. Statistical analysis in the middle of the chosen concentration range which corresponded to a peak height of 15.12 cm gave a standard deviation (0) of 0.209 59 cm (k, 1.39%). Differential-pul se polarography of M 2-mercaptothiazoline also showed anodic behaviour in the form of two much smaller waves when compared with metiamide and suggested that the presence of 2-mercaptothiazoline at relatively low levels should not significantly interfere in the determination of metiamide. Confirmation of this suggestion was sought by performing the following experiment. Four approximately 1 0 - 4 ~ solutions of metiamide were prepared and known amounts of 2-mercaptothiazoline were added to each.The solutions were polarographed alongside standard solutions of slightly lower and higher concentrations. The results obtained are given in Table I. TABLE I DETERMINATION OF METIAMIDE IN PRESENCE OF 2-MERCAPTOTHIAZOLINE Mass of metiamide/mg* 78.3 71.2 80.2 74.6 64.9 94.1 Mass of X-mercapto- Apparent Peak height/ thiazoline/mg* degradation, yo cm Recovery, yo 0.46 1.2 8.78 99.7 0.92 2.6 7.94 99.2 2.30 5.6 9.08 100.7 4.60 11.2 8.35 99.5 - - 7.20 - - - 10.68 - * The masses tabulated were diluted to 100 ml with methanol and then 2 ml of the resulting solution were diluted to 20 ml with 0.1 M sodium hydroxide solution.These results appear to confirm that metiamide can be determined accurately in samples that have undergone up to 10% degradation without interference from 2-mercaptothiazoline. In order to test for possible interference from the antioxidants, ascorbic acid or sodium disulphite, M solutions of each in 0.1 M sodium hydroxide solution were subjected to polarography, and the polarograms obtained are shown in Fig. 1. It is obvious that sodium disulphite will not interfere but that ascorbic acid, which also undergoes anodic oxidation, would be expected to interfere.As ascorbic acid is present in the liquid formulation at a level of only one twentieth of the concentration of metiamide, the degree of interference will be insignificant. Polarography of Samples The technique was applied to the determination of metiamide in three types of pharrna- ceutical formulations : tablets, injections and oral liquid containing ascorbic acid. The results obtained are given in Table 11, together with the results obtained by the high- performance liquid chromatographic method. High-performance Liquid Chromatography Metiamide is determined by ion-exchange chromatography on Zipax SAX columns (E. I. du Pont de Nemours Ltd.). The mobile phase is 0.1 M disodium tetraborate, adjusted to pH 8.5 with formic acid.All of these results compare favourably with those obtained by high-performance liquid chromatography except for the sterilised injections, for which the polarographic results were significantly higher than those obtained by high-perf ormance liquid chromatography. A comparison between the two techniques can conveniently be divided into two parts, dealing with the techniques themselves and their application to analytical problems. Being instrumental, both techniques are indirect, and in order to determine any compound the instrumental response must be compared with that of a standard sample. There is not a great difference between them as far as the time factor is concerned. In the present chromatographic method, injections can be made at 16min intervals, whereas polarography Benzoic acid is employed as an internal standard.226 ANDERSON .AND SMYTH TABLE I1 DETERMINATION OF METIAMIDE IN PHARMACEUTICAL FORMULATIONS BY POLAROGRAPHIC AND HIGH-PERFORMANCE LIQUID CHROMATOGRAPHIC (HPLC) METHODS Metiamide content Type of formulation Polarographic method HPLC method Tablets .. .. . . 201.4 mg per tablet 200.2 mg per tablet 200.2 mg per tablet Liquid , . .. . . 4.27% m/V 4.12% m/V 4.22% m/V 4.11% m/V Injection* sterilised : by filtration . . .. 10.29% m/V 9.96% m/V 10.22% m/V 9.80% m/V in an autoclave . . .. 9.88% m/V 9.59% m/V 10.12% m/V 9.68% m/V 198.8 mg per tablet * A sample from one batch of material was divided and sterilised by two different methods. requires a total turn-round time (determination plus cell cleaning) of about 10 min.This difference probably would not be considered important. Another difference, though again not very important, is in the precisions of the two methods. The polarographic method has been shown to have a standard deviation of 1.39%, whereas experiments carried out during the development of the chromatographic method have shown the technique to have a standard deviation of 1.09%. It is reasonable to compare these two figures, as both represent the error when one operator carried out repeated determinations on one solution on one day, thus eliminating dilution errors, operator error and day-to-day instrumental variation. Also, the internal standard method is used in high-performance iquid chromatography, so two peak heights have to be measured compared with one in polarography.So far, therefore, there has not been a great deal to choose between the methods. When one considers the application of the tvvo techniques to analytical problems, and the determination of metiamide in particular, high-performance liquid chromatography does have one distinct advantage. This difference ljLes in the natures of the two techniques in the sense that polarography, which in this instance is the determination step for the parent compound, discriminates against but does not measure interferences, whereas high- performance liquid chromatography consists of a separation stage followed by a determina- tion step. In the present chromatographic method, metiamide, ascorbic acid and 2-mercaptothiazoline are all resolved from each other and so the method can really be described as stability indicating. The conclusion must be that high-performance liquid chromatography remains the superior method for the determination of metiamide. However, polarography is recommended as an excellent method for its rapid determination for product control purposes. We thank the management of Smith, Kline and French Laboratories Ltd. for permission to publish this paper. References 1. 2. 3. 4. Smyth, W. F., Svehla, G., and Zuman, P., Analytica Chim. Acta, 1970, 51, 463. Smyth, W. F., Svehla, G., and Zuman, P., Analytica Chim. Acta, 1970, 52, 129. Smyth, W. F., Zuman, P., and Svehla, G., J . EJectroanaZyt. Chem., 1971, 30, 101. Smyth, M. R., and Osteryoung, J., to be published. Received August 25dh, 1977 Accepted October loth, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300223

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

Analyst, March, 1978, Vol. 103, pp. 227-232 227 Simultaneous Kinetic Determination of Mixtures of Ions by Digital Analysis of Amperometric or Potentiometric Data Anne- Marie Albrecht-Gary, Jean- Paul Collin, Pierre Jost, Philippe Lagrange and Jean-Paul Schwing €hole Nationale Sztp6rieure de Chimie, 1 Rue Blaise Pascal, 67000 Strasbourg, France Electroanalytical techniques allow the kinetic determination of concentrations of ions in their mixtures when the reactions used proceed at sufficiently slow velocity. A4mperometry and potentiometry with digital data acquisition and treatment have been used for determinations of calcium and magnesium ions based on the exchange reactions between lead(I1) and alkaline earth metal complexes of (trans- 1,2-cyclohexylenedinitrilo) tetraacetic acid.Over a concentration range of -5 x lo-‘ M precision and accuracy are about 5-10yo in amperometry and 5% in potentiometry. Keywords : Calcium determination ; magnesium determination ; electroanalysis ; kinetics ; automation Chemical analysis can be based on reaction-rate methods, and books by Mark and Rechnitz,l Guilbault2 and Yatsimirskii3 and reviews by Mottola4 and Greinke and Mark5 show the con- siderable popularity of these kinetic methods. Differential reaction rate methods allow the quantitative determination of mixtures without separation. Spectrophotometry is the most widely used technique to follow the reaction rate, but electroanalytical techniques can also be used to follow changes in amounts of reactants or products. In this paper we discuss the advantages of computer-controlled amperometric and potentio- metric studies in kinetic methods of analysis, which we have applied to the determination of calcium and magnesium in mixtures of the two.Pausch and Margerum6 have described the determination of alkaline earth metal ions, using stopped-flow spectrophotometry. They were able to analyse 10-5 M solutions with a precision of &lo%. The reactions between lead(I1) and either CaCDTA2- or MgCDTA2- complexes proceed a t sufficiently different velocities to permit the quantitative determination of calcium(I1) and magnesium( 11) [CDTA4- is the (trans-l,2-cyclohexylenedinitrilo) tetraacetate anion]. The exchange reaction rate between lead and a metal (M) is first order in chelate and in H+ if lead(I1) is in excess: slow MCDTA2- + H+ -+ HCDTA3- + M2+ fast HCDTA3- + Pb2+ 3 PbCDTA2- + H+ The response time of the apparatus allows a choice of the pH of the solution.We chose pH 7.6 in 0.5 M sodium acetate solution; the ionic strength of the solution was maintained constant a t 0.5 and the activity factors of the reagents remained constant. The ampero- metric measurements were made with a dropping-mercury electrode and the potentiometric measurements with a lead-sensitive cationic electrode. These electrodes give a signal that varies with the concentration of free lead and thus we can calculate [PbCDTA2-] versus time. The solution of the integrated equations gives the concentrations of calcium(I1) and mag- nesium (11) : [PbCDTA2-] = [PbCDTA2-1, - [Ca(II)] exp(-kobE1,Ca t ) - [Mg(II)]eXp(--kobs,~gt) .. (1) where [PbCDTA2-1, is the concentration of PbCDTA2- at infinite time.228 Apparatus ALBRECHT-GARY et al. : KINETIC DETERMINATION OF IONS BY Analyst, VoZ. 103 Experimental A schematic diagram of the apparatus is shown in Fig. 1. Injector N2 N2 Injector Fig. 1. Schematic diagram of apparatus Amperometric measurements The amperometric measurements were carried out with a three-electrode Beckman Electro- scan 30 potentiostat (POT). The potential of the dropping-mercury electrode (DME) was -0.70 V veYsus a silver - silver chloride electrode in 3 M sodium chloride reference electrode (RE) (reduction of lead acetate). The electromagnetic hammer for tapping the capillary, its electronic generator and the cell (A) were a Metrohm E354 unit.The drop time was con- trolled to Q s. The graph of polarographic current v e m w time was plotted on a Metrohm E261 recorder, which was more sensitive than -the Electroscan 30, and was equipped with a linear potentiometer giving a voltage signal that was proportional to the chart ordinates. This voltage and the related time were punched on paper tape with a Chauvin-Arnoux EBP 50 data acquisition system. The punched tape was processed on an Olivetti P 652 microcomputer. Potentiornetric measurements protected from the light : The potentiometric measurements were made with the cell as shown below, which was Test solution 1 0.5 M j 0.4 M CH,COONa pH 7.6 PbE I in 0.5 M CH,COONa CH,COONa j 0.1 M NaCl 1 AgCl I Ag I I A lead ion-sensitive electrode of the type Tac:ussel PBl (PbE) was employed. The e.m.f.of the cell (P) was determined with an accuracy of kO.01 mV over the range rt200.00 mV by employing a Tacussel Aries 20 000 digital voltmeter (DVM). This instrument possesses an input impedance greater than 10l2 a, and so the leakage current of the digital voltmeter has no effect on the e.m.f. of the cell (P). This voltmeter was interfaced to the Olivetti micro- computer equipped with a standard interface (ICU), a tape puncher (PN), a tape reader (LN), a teleprinter (TTY) and a magnetic tape memory (MLU). A Schneider DN 2000 clock controlled the time between the data points.March, 1978 DIGITAL ANALYSIS OF AMPEROMETRIC OR POTENTIOMETRIC DATA 229 A stream of purified nitrogen was employed to flush out oxygen (amperometry) and to keep the solution protected from carbon dioxide of the atmosphere (amperometry and poten- tiometry). The temperature was maintained at 25 & 0.1 “C by circulation of thermo- statically controlled water through jacketed cells.Reagents and Solutions Sodium acetate, lead acetate, calcium acetate, magnesium acetate, D-mannitol, sodium borate and CDTA4- (as Titriplex IV) were Merck pro analisi or Suprapur products. Mercury was 99.999 99% from Mercure-Industrie. Distilled, de-ionised and boiled water was used for preparing all solutions. The calcium and magnesium acetate solutions were prepared in 0.5 M sodium acetate solution and the excess of CDTA4- was lO-SO%. The pH was adjusted to 10-11 by adding 50% m/V sodium hydroxide solution. The lead acetate solution containing 0 .5 ~ sodium acetate was buffered at pH 7.6 with 5 x M sodium borate and 3 x M mannitol. All glassware was cleaned with nitric acid. Procedure Forty millilitres of the lead acetate solution were introduced into the cell and allowed to stand for 20 min with passage of nitrogen to remove oxygen and to attain temperature equili- brium. A 2-in1 sample of MCDTA2- was then rapidly injected with a syringe into the stirred lead acetate solution. When making amperometric measurements the stirrer was auto- matically stopped 4 s after the injection. In this solution the excess of lead was 20-50y0 versus CDTA4- and the ionic strength was constant. One hundred and eighty data points were generally recorded during 20 min. Towards the end of the reaction period a standardi- sation procedure was begun in order to determine the correlation between experimental measurements and lead acetate concentration.This standardisation was carried out by successive additions of known lead acetate solutions. The computation of calcium and magnesium concentrations started “on line” in potentiometry and “off line” in amperometry. Results and Discussion Determination of Concentration of PbCDTA2- We used the relation Equation (1) shows that it is necessary to evaluate [PbCDTA2-] as a function of time. [PbCDTA2-] = [Pb(II)] + [Pbfree] . . .. .. ‘ * (2) where [Pb(II)] is the known concentration of lead and Pbfree represents all species of lead except PbCDTA2-. [Pbfree] is measured experimentally. At pH 7.6 we postulate the pre- valence of the fast equilibria shown below whose constants, K,, K2 andK3, are found in the literature’ : where Ac- Thus or fast Pb2+ + Ac- +- PbAc+ is the acetate ion; Pb2+ + OH- + PbOH+ fast [PbAc+] K1 = [Pbz+] {Ac-] [PbOH+] K2 = [Pb2+] [OH-] fast [PbCDTA2-] [H+] Pb2+ + HCDTA3- + PbCDTA2- + H’ K3 = [Pb2+] [HCDTA3-]230 ALBRECHT-GARY et d.: KINETIC DETERMINATION OF IONS BY Analyst, Vd. 103 as [Ac-I~constant and [OH-] = constant, [IPbfree] = K[Pb2+] where K is an apparent constant dependent upon the pH and the acetate-ion concentration. The amperometric current intensity measured at -0.7 V versus silver - silver chloride in 3 M sodium chloride solution is proportional to [Pbfree]. The e.m.f. measured with the lead-sensitive electrode is given by E = e, + ln[Pb2+] where e, is an apparent constant including the standard potential of the lead electrode, the reference half-cell potential, the junction potentials and activity factors.We can also write RT RT 2F 2F E = e, - - InK + - In[Pbfree] The standardisation procedure gives e,’. For low concentrations of lead, where [Pb(II)] < 5 x 10-6 M, we replace 1 2F [pbfreel = exp [ [Pbfree] = exp ( A +. BE + CE2) (E - e i ) by the expression where A , B and C are constants determined by standardisation for each kinetic analysis. Calculations The method employed for the calculations of [Ca(II)] and [Mg(II)] has been described elsewhere. The n determinations of [PbCDTA2-] are the n elements of a column vector P. The problem is to find by linear least-squares treatment the column vector P containing the elements [PbCDTA2-] [Ca(II)] and [Mg(II)] such that .... - * (5) IIP - G * P‘ll = minimum . . where G is the d-row and three-column matrix: : System Performances When making polarographic measurements we observed, towards the end of the reaction, slow variations in intensity (1-3%), perhaps due to a local decrease in lead concentration giving rise to an irregular streaming of the solution around the mercury drops. When making potentiometric measurements we kept the lead electrode connected to the voltmeter and immersed in a solution containing a low concentration of lead acetate. This electrode was also protected from light to avoid instabilities in its potential. During the experiments a drift of about 0.1 mV h-l was observed. No correction was introduced for this drift.Another factor is the lag in response caused by rapid and significant changes in lead con- centration such as those brought about by injection of a concentrated buffered lead acetate solution. The establishment of the new equilibrium took generally 15 s in the amperometric system and 9 s in the potentiometric system; the time is longer in amperometry because theMarch, 1978 stirring must be stopped before taking the measurements. to the first 20 s of the kinetic analysis were not taken into account in the calculation. [Pbfree] was about 2% in amperometry and about 1.5% in potentiometry. DIGITAL ANALYSIS OF AMPEROMETRIC OR POTENTIOMETRIC DATA 23 1 The readings corresponding M, the mean relative standard deviation of the measurement of In the range 10-6-5 x Quantitative Analysis Typical analyses of mixtures of calcium(I1) and magnesium(I1) are given in Tables I and 11.The concentration ranges are 2 x 10-6-10-5 M with amperometric detection and 5 x 10-7-10-5 M with potentiometric detection. The determination of small concentrations of calcium(I1) and magnesium(I1) is better by potentiometry than by amperometry because the electrode response versus the concentration is logarithmic in the first instance and linear in the second; also, in potentiometry the solution can be stirred throughout the experiment. TABLE I DETERMINATION OF CALCIUM(II) IONS IN MIXTURES WITH MAGNESIUM(II) IONS The concentrations of magnesium(I1) ions determined simultaneously are given in the corresponding rows and columns in Table 11.Detection by amperometry Detection by potentiometry A A r Known concen- tration/ 15.22 10.15 20.3 5.07 7.61 2.54 10.15 3.81 M X 10' Determined concen- tration/ 15.27 10.04 20.93 4.24 7.36 2.35 9.58 3.2 M X lo6 Relative error, yo f0.3 - 1.0 f3.1 - 16.4 -3.3 - 7.5 -5.6 - 16.0 Relative standard deviation, 0.2 1.2 0.4 0.4 1.0 3.4 0.8 1.2 % Known concen- tration/ 11.95 2.38 23.68 4.76 4.51 2.37 0.92 0.47 M X lo6 Determined concen- tration/ 11.43 2.37 23.8 4.86 4.53 2.53 0.87 0.51 M X 10' Relative error, yo - 4.4 - 0.4 + 0.5 $2.1 +0.5 +6.8 - 5.4 +8.5 Relative standard deviation, 1.0 2.4 0.5 3.4 1.3 1 .o 0.8 2.0 % The precision is adequate if a precision and accuracy of 5-107, are required but the error may reach 15-20% in amperometry in exceptional instances.The relative standard de- viation given in Tables I and I1 describes the best fit for the solution of equation (5) and confirms the validity of the kinetic law used. The relative error on calcium(I1) and mag- nesium(I1) is much greater (5-10y0) and may be due to adsorption or desorption of reactants at the cell wall or accidental introduction of oxygen during standardisation. TABLE I1 DETERMINATION OF MAGNESIUM(II) IONS IN MIXTURE WITH CALCIUM(II) IONS The concentrations of calcium(I1) ions determined simultaneously are given in the corresponding rows and columns in Table I. 7 Known concen- tration/ 9.45 14.17 4.72 7.09 4.72 9.45 2.36 2.36 M X lo6 Detection by amperometry Determined concen- tration/ 10.03 14.0 5.14 8.08 4.79 10.28 2.83 2.47 M X 10' Relative error, yo + 5.8 - 1.2 + 8.9 + 13.9 + 0.2 + 8.8 + 20.0 +4.7 l r Relative standard deviation, % 2.0 1.0 2.5 0.4 1.5 0.9 2.7 1.6 Detection by potentiometry Known concentra- tion/M x los 11.9 2.39 4.73 4.8 2.26 4.77 0.93 0.47 Determined concentra- tion/M x 10' 11.4 2.37 4.5 4.87 2.18 4.93 0.93 0.41 Relative error, yo - 4.2 - 0.9 -4.5 + 1.9 -3.5 + 3.4 0.0 - 12.8 Relative standard deviation, 1.9 3.0 1.9 1.8 3.1 1.0 1.0 2.6 % Table I11 shows the relative precision at the 99% confidence level as a function of the range of metal concentration and of the technique of detection.The stopped-flow results are taken from the papers of Pausch and Margerum6 and Willis et aL9 The interesting feature of232 ALBRECHT-GARY, COLLIN, JOST, LAGRANGE AND SCHWING TABLE I11 COMPARISON BETWEEN DIFFERENT DETECTION TECHNIQUES Concentration Detection technique rang el^ Precision, yo Reference Amperometry .. .. .. . . 10-4-10-5 < 5 This work 10-5-2 x 10-6 10 This work Potentiometry . . .. .. . . 10-5-5 x lo-’ 5 This work Stopped-flow spectrophotometry . . 2 x 10-4-10-6 5-10 5, 9 the stopped-flow spectrophotometric detection is the time needed for the measurement, 1-10 s compared with 20 min for electroanalytical techniques if two cells are used (nitrogen being bubbled through the solution in one cell and measurements made in the other). The precision and the sensitivity are better, however, in electrochemistry than in fast spectro- photometric detection and the apparatus required for the electrochemistry is cheaper than that for spectrophotometry.The time needed for computation is the same for all of the techniques, but it is directly dependent on the number of data points. For example, a treatment of 180 data points needs about 1 h with the Olivetti P652 small computer, but only 10 s of real time with a Univac 1110 computer. The electroanalytical techniques used in differential kinetic analysis of calcium(I1) and magnesium(I1) give more accurate results than spectrophotometry. The limitations of this method, which has a good selectivity, are the ratio of the rate constants of the two parallel exchange reactions and the ratio of the two concentrations of metal. The first ratio must be below 0.2 and above 5 and the second between 0.1 and 10. This work was financially supported in part by the CNRS (France) and the DGRST (France). 1. 2. 3. 4. 5. 6. 7. 8. 9. References Mark, H. B., and Rechnitz, G. A., “Kinetics in AnalyticalChemistry,” Wiley-Interscience, New York, Guilbault, G. G., “Enzymatic Methods of Analysis,” Pergamon Press, Oxford, 1970. Yatsimirskii, K. B., “Kinetic Methods of Analysis,” Pergamon Press, Oxford, 1966. Mottola, H. A., Talanta, 1969, 16, 1267. Greinke, R. A., and Mark, H. B., Analyt. Chern., 1976, 48, 87R. Pausch, J. B., and Margerum, D. W., Awalyt. Chem., 1969, 41, 226. SillCn, L. G., and Martell, A. E., Comflilers, “Stability Constants of Metal-ion Complexes,” Special Collin, J. P., and Lagrange, P., Bull. SOC. Chint. Fr., 1976, 9-10, 1309. Willis, B. G., Woodruff, W. H., Frysinger, J. R., Margerum, D. W., and Pardue, H. L., AIzaZyt. Received July 14th, 1977 Accepted September 23rd, 1977 1968. Publications Nos. 17 and 25, Chemical Society, London, 1964 and 1971. Chern., 1970, 42, 1350.

ISSN:0003-2654    

DOI:10.1039/AN9780300227

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

Analyst, March, 1978, Vol. 103, pp. 233-237 233 Simple Semi-automated Procedure for Determining the Lipophilic Nature of Organic Compounds G. Voss Agrochemicals Division, CIBA -GEIGY Ltd., Basle, Switzerland This paper describes a semi-automated procedure for determination of the relative lipophilicities of organic compounds within a homologous series. The principle of the method is based on the ability of flexible polymer tubing to absorb highly lipophilic materials more readily than less lipophilic materials. The experimental assembly consists of a Technicon proportioning pump that generates two continuously flowing streams of an aqueous solution of the test compound and an ultraviolet spectrophotometer equipped with flow cell and recorder. The first stream passes through the flow cell without being pumped through polymer tubing, thus representing the reference sample.The second stream, however, is “tube-partitioned.’’ It arrives in the flow cell after the first stream and is recorded as a second peak shortly after the refer- ence sample. The ratio of the peak heights was found to be a measure of the lipophilic nature of the compounds investigated. Substituted phenyl- ureas and sulphonamides were screened by the new procedure and the results compared with those obtained by more conventional approaches using linear regression analysis. Keywords : Lipophilicity ; automated analysis ; Phenylwea herbicides ; sul- phonainides The lipophilic nature of organic compounds is known to represent one of the most important parameters for determining the behaviour of biologically active substances.Thus lipophili- city may govern the reactivity of organic molecules with active sites of enzymes and receptors or determine uptake, transport, excretion and metabolic rates. Therefore, a knowledge of the lipophilic character of compounds in a homologous series often appears to be a valuable tool for chemists engaged in the design of bioactive molecules and in quantitative structure - activity evaluations. The degree of lipophilicity of compounds is usually expressed by partition coefficients between organic solvents, mostly octanol, and water. Because of its greater simplicity thin-layer chromatography (reversed phase on paraffin-coated silica gel plates) has also been used.l The present method, however, takes advantage of the ability of flexible plastic tubing to absorb organosoluble compounds, a phenomenon that has been described earlier in connection with automated procedures for pesticide residue analysis., We have investi- gated the absorption effect more quantitatively and find a satisfactory correlation between absorptivity and lipophilicity data.On account of its speed of operation, its simplicity and its reproducibility the new procedure may be of interest to those working in fields of research indicated above. Experimental The instrumental assembly is shown in Fig. 1. The all-glass injection chamber (IC) and its mode of operation have been described previously.2 The principle feature of the flow diagram is a splitting of the stream shortly after the injection chamber. Both streams carry the same solutions, but one of them is being re-pumped through a flexible pump tube (Tygon, supplied by Technicon) in which absorption takes place.At the same time this process causes a time delay, so that the sample passing through this tube (the partitioned stream) arrives in the flow cell after the reference stream. Two peaks appear on the recorder from one injection, and by comparing their sizes a measure for the absorbed fraction is obtained. The tube partition coefficient, Pt, is calculated from the expression where h, and h, are the heights of reference and partition peaks, respectively.234 voss : SIMPLE SEMI-AUTOMATED PROCEDURE FOR DETERMINING Analyst, VoZ. 103 Stream I 71 mlmin-I IC I- Water or buffer . 2.0 n I 1 Waste 0 Stream 1 1 A 1.2 L Partitioning tube r t 0 Proportioning 0 twrn~ of Auto- Analyzer O I DO l- coil Waste 4 Spectrop h ot& meter Fig.1. Flow diagram of semi-automated system for determining lipophilicities of organic compounds, IC = Injection chamber with magnetic stirrer bar (magnetic stirrer underneath IC is not shown), volume of IC approximately 1 ml. The spectrophotometer is equipped with a silica flow cell, volume 80-90 pl, path length 10 mm. All tubing except pump tubes are of glass or PTFE. H1, DO and C5 denote types of Technicon glass fittings. With highly lipophilic materials the two peaks can be considerably different in size, which results in reduced accuracy. To overcome this problem it is often advisable not to use equal flow-rates for both streams but to divide the sample in such a way that an increased flow-rate is obtained for the sample to be partitioned.The examples presented in this paper are based on a flow through the injection chamber of 2 ml min-l. As 1.2 ml min-1 are re- pumped for partitioning the reference stream is characterised by a flow-rate of 0.8 ml min-1. They were dissolved in methanol in such concentrations that a portion of 5-10 pl injected into water (phenylureas) or phos- phate buffer pH 5.0 (sulphonamides) flowing through the injection chamber resulted in peaks of convenient sizes (Figs. 2, 3 and 4). All necessary calculations, including linear regression analysis, were carried out with a Hewlett-Packard HP-65 calculator. The chemicals used were of analytical-reagent grade.Results Fig. 2 presents the original peaks obtained with six substituted phenylurea herbicides, and Table I summarises their Pt values together with figures derived from them. Water solubilities and 7r values derived from partition coefficients (for the octanol - water system) were taken from the l i t e r a t ~ r e . ~ , ~ The intercorrelations between 7rt on the one hand and l/log p.p.m. and 7r on the other were very satisfactory. The equations are as follows: ]/log p.p.m. = 0.25nt + 0.25 7r = 1 . 0 7 ~ t + 0.01 n Y S 6 0.97 0.03 6 0.99 0.07 where 'YG is the number of results, Y the correlation coefficient and s the standard deviation. Fig. 3 represents the original recordings obtained with eight sulphonamides, the structures, tube partitioning, octanol - water partitioning and Rm values of which are summarised in Table 11.Linear regression analysis resulted in the following equations : 7r = 1.127rt + 0.01 AR, = 0.76~t + 0.12 T = 1.38ARm + 0.14 n Y S 8 0.96 0.18 8 0.94 0.16 8 0.95 0.20March, 1978 THE LIPOPHILIC NATURE OF ORGANIC COMPOUNDS 235 a, t m e v) 2 -30 min-i 2 I 3 c . * 4 U 5 I 6 c Time Fig. 2. Recordings at 236 nm obtained with six substituted urea herbicides. 1, Fenuron; 2, monuron; 3, diuron; 4, metoxur- on; 5, chlortoluron; and 6, fluometuron. 6, V C m -e a 9 3 t-30min-i 4 5 Time Fig. 3. Recordings at 266 nm ob- tained with eight substituted sulphon- amides injected into Sorensen phos- phate buffer, pH 5. Substituents: 1, unsubstituted; 2, 4-CH,; 3, 4-C1; 4, NH,; and 8, 2-C1. 4-CN ; 5, &NO,; 6, 4-N(CH3), ; 7, 4-S02- Time Fig.4. Reproducibility test show- ing 10 injections of lop1 of monuron solution.236 voss : SIMPLE SEMI-AUTOMATED PROCEDURE FOR DETERMINING Analyst, VoZ. 103 TABLE I COMPARISON BETWEEN TUBE PARTITIONING VALUES, WATER SOLUBILITIES AND T VALUES DERIVED FROM OCTANOL - WATER PARTITION COEFFICIENTS OF SUBSTITUTED PHENYLUREA HERBICIDES N H--0-[I\] (CH?,) Water solubility, Compound R Pt LogPt rt* p.p.m. I? Fenuron .. .. H 0.08 - 1.12 0 3 850 Monuron .. . . 4-C1 0.42 -0.38 0.74 230 Diuron . . .. . . 3,4-Di-C1 1.77 0.25 1.37 42 Metoxuron . . . . 3-Cl-4-OCH3 0.29 -0.53 0.59 678 Chlortoluron . , . . 3-C1-4-CHa 1.07 0.03 1.15 70 Fluometuron . . . . 3-CF3 0.75 -0.13 0.99 90 * rt = log Pt (substituted)-log Pt (unsubstituted). l/log p.p.m.0.28 0.42 0.62 0.35 0.54 0.51 71 0 0.70 1.46 0.72 1.28 1.07 If the regression analysis was carried out without considering the 2-chloro derivative (ortho- effect in the T and AR, but not in the nt series) the following equations were obtained: n = 1 . 2 3 ~ t + 0.05 AR, = 0.86nt + 0.15 9% Y S 7 0.98 0.13 7 0.98 0.10 The question of whether the ortho-effect can or cannot be analysed by tube partitioning awaits further elucidation. TABLE I1 COMPARISON BETWEEN TUBE PARTITIONING VALUES, OCTANOL - WATER PARTITION COEFFICIENTS AND R, VALUES O F SUBSTITUTED SULPHONAMIDES All values determined at pH 5.0 in Sorensen phosphate buffer. R Pt H .. . . 0.37 4-CH,. . . . 0.83 4-C1 . . . . 1.78 4-CN . . . . 0.41 4-NO2.. . . 0.89 4-N(CH3)2 . . 0.26 4-SOzNH2 . . 0.04 2-c1 . . .. 1.72 log Pt -0.43 - 0.08 0.25 - 0.39 - 0.05 -0.59 - 1.36 0.24 r t 0 0.35 0.68 0.04 0.38 - 0.16 -0.93 0.67 Poet.* 35.19 101.18 188.43 67.27 137.38 29.70 2.23 106.02 log P" 1.55 2.01 2.28 1.83 2.14 1.47 0.35 2.03 r* 0 0.46 0.73 0.28 0.59 - 0.08 - 1.19 0.48 R m * t ARm*t 1.21 0 1.65 0.44 2.05 0.83 1.38 0.16 1.62 0.40 1.35 0.13 1.58 0.37 0.59 -0.63 * Data from Professor Dr. J. K. Seydel and co-workers (Forschungsinstitut Borstel, Germany). t Thin-layer chromatography on paraffin-coated silica gel plates. Discussion From the results presented in this paper and from additional work in our laboratories we conclude that the ability of polymer tubing to absorb organic compounds according to their lipophilic nature can be utilised for routine and research work. Although the exact mech- anism of the tube partitioning process is not yet completely understood, we feel we have sufficient evidence for the usefulness of the procedure. This is particularly true for investi-March, 1978 THE LIPOPHILIC NATURE OF ORGANIC COMPOUNDS 237 gations on homologous series of compounds of which the relative lipophilicities are of interest to chemists and biologists. The conditions under which tube partitioning is carried out can be varied to a great extent according to the particular requirements of the series under investigation. Thus, tube diameters and lengths can be altered or Tygon tubing can be replaced by other materials such as Acidflex, which show different absorption characteristics.The speed of operation can be increased by using a two-channel recording system or, if an entire series cannot be measured at the same fixed wavelength, partitioning can be carried out manually and the solution pumped through the plastic tube, scanned over the appropriate wavelength range and compared with an unpartitioned sample.According to our experience a polymer partitioning tube can be used for a great number of experiments until its quality suffers. Variations between batches of tubes or the influence of tube ageing on partitioning properties, however, are still to be investigated. I am indebted to Professor Dr. J. K. Seydel (Forschungsinstitut Borstel, Germany) for permission to present his unpublished results, and for the samples of sulphonamides listed in Table 11. The technical assistance of Miss E. Grund is gratefully acknowledged. References 1 . Draber, W., Buchel, K. H., and Dickore, K., “Proceedings of the Second International IUPAC Congress on Pesticide Chemistry, 1971,” Volume 5, Gordon and Breach, New York, London, Paris, 1972, pp. 153-175. 2. 3. Voss, G., J . Ass. Off. Analyt. Chem., 1973, 56, 1506. Geissbuhler, H., Martin, H., and Voss, G., in Kearney, P. C., and Kaufman, D. D., Editors “Herbi- cides. Chemistry, Degradation and Mode of Action,” Volume I, Marcel Dekker, New York and Basle, 1975, pp. 209-291. Hansch, C., and Deutsch, E. W., Biochirn. Biofihys. Acta, 1966, 112, 381. Received August 4th, 1977 Accepted September 19th, 1977 4.

ISSN:0003-2654    

DOI:10.1039/AN9780300233

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

238 Analyst, March, 1978, Vol. 103, $9. 238-245 Determination of Phosphorus in the Presence of Iron(ll1) and Iron(l1) S. McLeod and (the late) A. R. P. Clarke Commonwealth Scientific and Industrial Research Organisation, Division of Soils, Private Bag No. 2, Glen OsmoNd, South Australia 6064 Both manual and automated techniques have been developed for the determination of phosphorus in the presence of iron(II1). Interference from iron(II1) has been substantially eliminated by its reduction with hydroxyl- ammonium chloride prior to the formation of the molybdophosphate blue complex using ascorbic acid. The small but reproducible error due to the formation of iron(I1) is dependent on the iron to phosphorus ratio present in the digest solution but is significant only when this ratio is more than 500.Both methods have been applied successfully to the determination of phosphorus in digests and extracts of soil materials where the ratio is far below the value of 500, thus making a correction unnecessary. Keywords : Phosphorus determination ; soil analysis ; spectrophotometry ; iron interference The interference of iron(II1) in the spectrophotometric determination of phosphorus using the reduced molybdophosphate blue complex has been described for various reductants. Boltz and Mellonl used hydrazine sulphate and reported that more than 200 p.p.m. of iron(II1) in an aliquot caused significant errors. Ging2 used hydroquinone and reported that up to 540 pg of iron(II1) per 25 ml can be tolerated, and Fogg and Wilkinson3 reported that up to 20mg of iron(II1) could be tolerated in the aliquot when using ascorbic acid as the reductant.Chalmers4 used ammonium iron(I1) sulphate and reported that an error of 1 pg in 50 pg of phosphorus was caused by 1 mg of iron(II1) after its prior reduction with silver. Hingston5 used ascorbic acid and reported that up to 50mg of iron(II1) could be tolerated in the aliquot without error after its prior reduction with sodium disulphite. The above data are difficult to compare but, if they are re-calculated to common units of iron(II1) that could be tolerated in solution, the values obtained are approximately 100, 22, 400, 20 and 1 000 pg ml-1 of iron(III), respectively. It appears, therefore, that the magnitude of the interference by iron(II1) is dependent on the reductant used to form the molybdophosphate complex.Similarly, the error reported after prior reduction of iron(II1) also appears to depend on the reductant used for the colour formation or the conditions used for prior reduction. The aim of this investigation was to evaluate the effects of iron(II1) and iron(I1) on the recovery of phosphorus and to develop both manual and automated procedures for the routine determination of phosphorus in soil. Manual Procedure Experimental Efects of iron(III) on the determination of phosphorus A series of solutions containing 10 pg of phosphorus was prepared, containing 0, 5, 15, 25, 50, 75 and 100mg of iron(II1) added as an aqueous solution of iron(II1) chloride. Phosphorus was determined in a final volume of 50 ml by the manual method recommended below but without the prior reduction step.Reduction of iron(III) with sodium disulphite A series of solutions containing 0, 5, 15 and 25 pg of phosphorus was prepared, each containing 0, 5, 15 and 25 mg of iron(II1). The series was replicated ten times and analysed in random order. Reduction of iron(II1) was carried out by adding 1 g of sodium disulphite to each flask and allowing the flasks to stand overnight before developing the molybdo- phosphate blue complex as described in the proposed method. The solutions were made up to a final volume of 50 ml with water.MCLEOD AND CLARKE 239 Reduction of iron (111) with hydroxylammonium chloride This experiment included a larger number of phosphorus levels and five-fold replication was the largest that would enable all of the analyses to be completed in 1 d.Phosphorus was determined as described in the proposed procedure using a 10% m/V solution of hydroxylammonium chloride as the prior reductant. Efect of iron(I1) on the recovery of phosphor.us A series of solutions containing 0, 5, 15 and 25 pg of phosphorus, each containing 0, 5, 15 and 25mg of iron(I1) added as an aqueous solution of ammonium iron(1I) sulphate, was prepared and replicated five times. Phosphorus was determined by the proposed procedure, including the prior reduction step. Several randomly selected phosphorus - iron(I1) solutions were analysed concurrently, without the prior reduction step. 0 Iron ( I I I ) added/mg Fig. 1. Effect of iron(II1) on the recovery of phosphorus by (a) manual and (b) automated procedures. Results and Discussion When no prior reduction step was used, there was a marked decrease in the recovery of phosphorus with increasing concentrations of iron(III), as can be seen in Fig.1 (a). Little if any iron(II1) can be tolerated without error when the method is used without this step. Prior reduction of iron(II1) with sodium disulphite resulted in over-estimation of phos- phorus. A graph of the mean error in micrograms of phosphorus against milligrams of iron(II1) added for each level of phosphorus showed that there was a linear relationship between the measured error in the recovery of phosphorus and the concentration of iron(III), as can be seen in Fig. 2 (a). The addition of more sodium disulphite had no further effect on the error in the recovery of phosphorus. Similarly, there was a linear relationship between the over-estimation of phosphorus and the concentration of iron(II1) when prior reduction of iron(II1) was effected with hydroxyl- ammonium chloride [Fig.2 (b)]. The addition of excess of hydroxylammonium chloride imparted a greenish colour to the molybdophosphate complex, as did heating the solutions for longer than the prescribed period. However, neither of these effects was found to interfere with the sensitivity or reproducibility of the analyses and neither had any further effect on the error in the determination of phosphorus over the range used. There was no significant difference between the two methods of prior reduction of iron(II1) except at low levels of phosphorus, where the error in micrograms of phosphorus was different for the 0 and 5 p g of phosphorus levels when reducing with sodium disulphite but not different when reducing with hydroxylammonium chloride.At all other levels of phosphorus the error was proportional to the amount of added iron(II1) and to the level of phosphorus. The effect of iron(I1) on the determination of phosphorus was similar to that of iron(II1). There was a linear relationship between the measured error for phosphorus and the concentration of iron(II), with and without reduction by hydroxylammonium chloride, except at the 0 pg of phosphorus level, as can be seen in Fig. 3.MCLEOD AND CLARKE: DETERMINATION OF PHOSPHORUS Anahst, VoZ. 103 5 15 25 5 15 25 +S.E. k0.014 k0.029 i0.038 k0.016 +-0.021 50.028 Iron(l1l) added/mg Fig.2. Reduction of iron(II1) with (a) sodium disulphite and (b) hydroxylammonium chloride. Symbols S.E. is the standard error of the mean for each level of indicate mean value and range of observations. iron. The variance of the over-estimation of phosphorus calculated from the replicate observa- tions made at each level of phosphorus showed a significant increase (P (0.05 using Bartlett’s test6) with increasing iron(II1) concentration but no change with increasing phosphorus levels. In contrast to the reduced iron(II1) results, there was no change in the +S.E. I Iron(ll1 added/mg Fig. 3. Effect of iron(I1) on the recovery of phosphorus. Symbols indicate mean value and range of observations with (0) and without ( x ) hydroxylammonium chloride.S.E. is the standard error of the mean for each level of phosphorus added.March, 1978 IN THE PRESENCE OF IRON(III) AND IRON(II) 241 variance with increasing iron(I1) concentrations but a significant increase in the variance with increasing phosphorus level. These changes are indicated in Figs. 2 (a), 2 (b) and 3, where the appropriate standard errors (S.E.) of the means are given. The change in the pattern of variance from iron(I1) to reduced iron(II1) may be caused by some difference between iron(I1) formed by the reduction of iron(II1) and iron(I1) added as a solution of ammonium iron(I1) sulphate. These results show that there is otherwise little difference between the reduction of iron(II1) and the addition of iron(I1) and thus indicate the minimum error in the determination of phosphorus that can be attained in the presence of iron(II1) after its complete reduction to iron(I1).For most practical purposes this error is negligible for solutions that contain a ratio of iron to phosphorus of up to 500, which is greater than that normally found in soil extracts and digests, and can be disregarded. When necessary, more precise analyses can be obtained on solutions with known iron concentrations by applying a correction factor calculated from graphs such as those in Figs. 2 and 3. Reduction of iron(II1) with hydroxylammonium chloride was found to be rapid even in the cold. The reduction was faster, however, with standard solutions (1-2 min) than with some soil extracts (1-10 min). When added to hot solutions, complete reduction occurred in standard solu- tions and soil extracts within 5 min and therefore heating was adopted for the recommended met hod.The type of reducing agent is relatively unimportant, provided that its presence does not interfere with the subsequent formation of the molybdophosphate complex and that complete reduction of the iron(II1) is attained. Hydroxylammonium chloride has many advantages in that it is very stable in solution, effects complete reduction in a short time, can be adapted for use in an automatic analyser and is more acceptable for use on a routine basis. Sodium disulphite is unstable in solution, effects complete reduction only after several hours, has a disagreeable odour that could be toxic to the analyst and cannot be used in an automated system owing both to its instability in solution and the large and continuous release of sulphur dioxide.The proposed manual procedure has been applied to the routine determination of soil phosphorus in hydrochloric acid and other extracts of soil materials' and has proved very satisfactory . Recommended Manual Method for the Determination of Phosphorus in Soil Reagents Ammonium molybdate solution. Dissolve 48 g of (NH,),Mo,0,,.4H,O in 200 ml of water. Add carefully 495 ml of concentrated sulphuric acid (95-97y0) to 305 ml of water and allow the mixture to cool. Add the molybdate to the acid slowly with stirring and dilute to 11 with water. Transfer the solution to a dark-glass reagent bottle and store it in a refrigerator. Add 50ml of the above molybdate solution to about 150ml of water.Mix the two solutions and dilute to 250ml with water. This mixed reagent must be prepared freshly each day immediately before use, as slow decomposition of the ascorbic acid occurs. Sodium hydyoxide solution, appvoximately 4 M. Dissolve 160 g of sodium hydroxide in about 600 ml of water and dilute to 1 1 with water when cool. SuZphuric acid, approximately 3 M. Add 18 ml of concentrated sulphuric acid (95-97y0) to about 60 ml of water and dilute to 100 ml with water. p-Nitrophenol solution. Dissolve 0.5g of p-nitrophenol in about 60 ml of water and dilute to 100 ml with water. HydroxyZammonium chloride solution. Dissolve 100 g of reagent in about 600 ml of water and dilute to 1 1 with water. Mixed reagent solution.Dissolve 2.1 g of reagent-grade ascorbic acid in about 20 ml of water. Procedure Pipette an aliquot of extract containing less than 25 pg of phosphorus into a 50-ml Cali- brated flask, add 1 drop of p-nitrophenol indicator and neutralise with 4 M sodium hydroxide solution to a straw-yellow colour or the first permanent appearance of a precipitate. Add 2 ml of 3 M sulphuric acid and dilute to about 30 ml with water. Mix and transfer the flasks to a boiling water bath. After 5 min, add 2 ml of hydroxylammonium chloride reagent and mix. Add an additional 2 ml of reagent, mix and leave for a further 5 min. By dip-pipette242 MCLEOD AND CLARKE : DETERMINATION OF PHOSPHORUS Analyst, Yol. 103 add 10 ml of the mixed reagent, wash down with water, mix and leave on the water-bath for 10min.Remove the flasks from the water-bath, cool and dilute the solution to 50ml with water. Read the resulting colours in a spectrophotometer at 820 nm using 2-cm curvet tes. A range of phosphorus standards containing 0-25pg of phosphorus are measured con- curently. After use, it is advisable to leave the calibrated flasks filled with dilute sulphuric acid (1 + 1). Automatic Analyser Procedure Experimental As the conditions and sensitivities of automated systems are completely different from those of manual procedures, the effects of iron(II1) and iron(I1) were re-evaluated. The manifold used in the following experiments was originally designed for the determination of phosphorus in sulphuric acid digests of plant samples with a working range of 0-10 pg ml-l of phosphorus. I n order to avoid undue alterations to this existing system, the prior reduction treatment with hydroxylammonium chloride was carried out manually in the analyser sample tubes as described for the manual method.Efect of iron(I1I) and iron(I1) on the determination of fihosphorzcs Automation of the colorimetric procedure enabled the effects of iron(II1) and reduction of iron(II1) and the effect of iron(I1) to be determined in one experiment. Generally four levels, of 0, 2, 4 and 10 pg ml-l of phosphorus, were used throughout with iron levels of 0, 1, 5, 10, 15 and 20 mg ml-l of iron(II1) or iron(I1) as necessary. The series was replicated five times and, after reduction of the appropriate solutions with 20% m/V hydroxyl- ammonium chloride, phosphorus was determined on the complete series in random order using a 1-min sample - 1-min wash cycle on the analyser.The solutions requiring reduction of the iron(II1) were repeated using 30% m/ V hydroxylammonium chloride and analysed concurrently. Prior reduction of the iron(I1) series was omitted. Results and Discussion The effect of iron(II1) on the recovery of phosphorus is shown in Fig. 1 (b) for three levels of phosphorus. As expected, there is a marked decrease in recovery of phosphorus with increasing iron(II1) concentrations. A comparison between Fig. 1 (a) and (b) shows clearly that this automated method is more sensitive than the manual method to the presence of iron(II1) and that only 4mg of iron(II1) need be present to suppress completely the forma- tion of the molybdophosphate complex.A sharp double peak resulted on the recorder chart with all solutions that contained more than 2 mg of iron(III), as can be seen in Fig. 4. As in all continuous flow systems, the front of the sample is mixed with the previous wash and, in turn, the front of the incoming wash is contaminated with the previous sample, as shown in Fig. 5. Because iron(II1) acts as a complete suppressant at higher concentrations, the molybdophosphate complex is formed only in the diluted leading and trailing fractions of these samples and so produces these distinctive double peaks. Although no result can be calculated from this type of output, it provides a useful indication of the presence of any interfering ion that acts as a suppressant in any colorimetric determination. The results in Table I show the effect of reducing iron(II1) with both 20 and 30% m/V hydroxylammonium chloride.These results are the means of five replicates and, as before, show an increased recovery of phosphorus when the iron(II1) concentration is less than 10 mg ml-1. The higher additions of iron(II1) obviously were not fully reduced by 20% m/V hydroxyl- ammonium chloride and even with the stronger reduction treatment enough iron(II1) remained unreduced from the addition of 20mg of iron(II1) to suppress completely the molybdophosphate complex. Statistical analysis of these results shows that the standard error increases with increasing phosphorus concentration, and that the error in the recovery of phosphorus is directly proportional to the concentration of iron(II1) and not to the ratioMarch, 1978 IN THE PRESENCE OF IRON(III) and IRON(II) 243 A 1 3 I J I F 1 -Flow Fig.4. Recorder trace showing typical “double peaking” caused by iron(II1) interference. Peaks: A, P, + 10 mg of iron(II1) reduced; B, P,, + l m g of iron(II1) unreduced; C, Pz standard; D, PI, standard; E, P, + 2 mg of iron(II1) unreduced; F, P, standard; G, P, + 4mg of iron(II1) unreduced; H, iron(II1) unreduced; I, P, + 15mg of iron(II1) reduced; J , P, + 15mg of iron(I1); K, PI, + 0.5 mg of iron(II1) unreduced; L, P, + 0.5 mg of iron(II1) unreduced; and M, P2 + 15 mg of iron(I1). P,, P, and PI, represent phosphorus concentrations of 2, 4 and 1 Opg ml-l, respectively. of phosphorus to iron(II1) as was found in the manual procedure.Similarly, as can be seen in Table 11, the effects of iron(I1) added as ammonium iron(I1) sulphate agree very closely with those in Table I except that there is no significant difference in the standard error with increasing phosphorus concentrations. These results are different to those Recorder trace Sample Wash Sample , Wash I & ‘ ’ & ‘ ‘ & I Wash I wash I Sample I I Wash ’ 1 I -Flow Fig. 6. Schematic diagram of “double peaking” effect.244 MCLEOD AND CLARKE : DETERMINATION OF PHOSPHORUS Analyst, Vol. 103 obtained by the manual procedure, which may be due partly to the automated method being more sensitive to interference from iron(II1) and partly to different conditions pertaining to the formation of the molybdophosphate complex, the latter being the pre- dominant effect. TABLE I EFFECT OF PRIOR REDUCTION OF IRON(II1) ON THE DETERMINATION OF PHOSPHORUS BY AN AUTOMATIC ANALYSER Recovery of phosphoruslpg ml-l A 7 7 Phosphorus Hydroxylamine Iron(I1I) concentration/mg ml-l concentration/ concentration, A 3 pg ml-1 % 1 5 10 15 20 S.E.* 2 20 1.87 2.05 2.09 1.91 0 30 1.89 2.07 2.11 2.21 0 &O.OlS 4 10 20 3.85 4.06 4.14 3.90 o 30 3.85 4.05 4.13 4.25 0 k0.026 20 9.90 10.09 9.27 8.27 0 30 9.94 10.12 10.16 10.25 0 f 0.058 * S.E.is the standard error of the mean of five determinations. Recommended Method for the Determination of Phosphorus in Soil with an Auto- matic Analyser Reagents Dissolve 5 g of reagent-grade ammonium molybdate in about 100ml of water. Add carefully 56ml of concentrated sulphuric acid to 300ml of water, mix and cool.This reagent is stable for several weeks if stored in a dark-glass reagent bottle. Ascorbic acid solution. Dissolve 2 g of reagent-grade ascorbic acid in 200 ml of water and keep the solution in a dark-glass reagent bottle. This reagent decomposes slowly and must be prepared freshly each day. Hydroxylammonium chloride solution, 20% m/V. Dissolve 20 g of hydroxylammonium chloride in about 60 ml of water and dilute to 100ml with water. This solution is very stable and can be stored for long periods of time. (This solution is equivalent to the 30% solution used in the experimental work described.) Add 90 ml of hydrochloric acid (32%) to about 600 ml of water and dilute to 1 1 with water. The concentration of this solution is equivalent to the acid concentration in the soil digests. Ammonium molybdate solution.Combine the two solutions and dilute to 500ml with water. Hydrochloric acid, 9% V/V. TABLE I1 EFFECT OF IRON(II) ON THE DETERMINATION OF PHOSPHORUS BY AN AUTOMATIC ANALYSER Recovery of phosphorus/ pg ml-' r h 7 Phosphorus Iron(I1) concentration/mg ml-1 concentration/ r- A 1 pg ml-l 1 5 10 15 S.E.* 2 1.95 2.09 2.07 2.20 f0.027 4 3.91 4.05 4.14 4.23 f0.027 10 9.98 10.12 10.15 10.23 f0.027 *S.E. is the standard error of the mean of five determinations. Procedure Fig. 6. chloride. The manifold and flow sequence for the determination of phosphorus in soil are shown in The system includes the prior reduction of iron(II1) present with hydroxylammonium For most routine determinations, the hydrochloric acid extracts are dispensedMarch, 1978 IN THE PRESENCE OF IRON(III) AND IRON(II) 245 directly into the analyser and a time cycle of l-min sample - l-min wash is used.The absorbance is measured at 820nm with a 15-mm flow-through cell, Standard phosphorus solutions in the range 0-10 pg ml-l prepared in 9% V/V hydrochloric acid are analysed after every 30 samples and the results are calculated from the recorder chart by computer using the program INTERPS.~ Hydrochloric acid at the concentration in the soil extracts corrodes the sampling needle and the sampling head of the analyser over a period of time. Hence the use of this method is now limited to those occasions when a large number of analyses warrants the automation of the colorimetric procedure.Pump mi min -’ Hydroxylammonium chloride Ascorbic acid I 1 ; Speccophotometer 820 nrn Fig. 6. Diagram of manifold for the determination of phosphorus in soil. SMC = single-mixing coil. Conclusion The reduction of iron(II1) prior to the formation of the molybdophosphate blue complex decreases the error in the determination of phosphorus to an acceptable level. The type of reductant is not critical, provided that complete reduction to iron(I1) is obtained. There is a small, reproducible error resulting from the presence of iron(II), which is proportional t o the concentration of iron in the aliquot. The magnitude of this error depends on the conditions of the formation of the molybdophosphate complex, and individual methods must be evaluated in order to determine whether or not the error is acceptable. Both the recom- mended manual and automated methods have been applied successfully to the routine determination of phosphorus in soil extracts, for which the remaining error can be disregarded for most soils. The authors thank Mr. R. D. Bond of the Division of Soils for his many helpful suggestions throughout the experimental work and Mr. K. Cellier of the Division of Mathematics and Statistics for the statistical analyses. References 1. 2. 3. 4. 5. 6. 7. 8. Boltz, D. F., and Mellon, M. G., Analyt. Chem., 1947, 19, 873. Ging, N. S., Analyt. Chew., 1956, 28, 1330. Fogg, D. N., and Wilkinson, N. T., Analyst, 1958, 83, 406. Chalmers, R. A., Analyst, 1953, 78, 32. Hingston, F. J., Tech. Memo. Div. Soils, CSIRO Aust., No. 9, 1959. Bartlett, M. S., J l R. Statist. SOG., Suppl., 1937, 4, 137. Taylor, R. M., and Schwertmann, U., Aust. J . Soil Res., 1974, 12, 133. Colwell, J. D., Tech. Pap. Div. Soils, CSIRO Aust., No. 24, 1974. Received February 3rd, 1977 Amended September 26th, 1977 Accepted October loth, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300238

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

246 Analyst, March, 1978, Vol. 103, $9. 246-251 Relationship Between Activity and Concentration Measurements of Plasma Potassium* D. M. Band, J. Kratochvil, P. A. Poole Wilson and Tom Treasure Sherrington School of Physiology and the Intensive Care Unit, St. Thomas's Hospital, London, SE1 7EH An ion-selective electrode incorporating valinomycin in a poly (vinyl chloride) membrane was used for examining potassium in blood and plasma. Using an electrode cell thermostated at 37 "C and a saturated potassium chloride liquid junction, the activity of plasma potassium appeared to be very similar to that for similar concentrations of potassium in physiological saline. An apparent difference in potassium activities between the plasma of whole blood and separated plasma was found to be comparable to the discrepancy between the pH of whole blood and plasma, and was attributed to an effect of red blood cells on the liquid junction of the measuring cell.The magnitude of the liquid junction potential between a bridge solution and blood is a major uncertainty both in the determination of activities of ions by electrode measurements (e.g., the biological pH scale) and in the measurement of potential differences across biological membranes. The advent of the highly selective plastic membrane electrodes enables this problem to be re-examined. Plasma protein determination ; potassium activity ; ion-selective Keywords electrodes ; valinomycin ; liquid junction potential The use of electrochemical cells for the measurement of inorganic ions, other than hydrogen, in biological fluids offers a number of advantages over the customary emission flame photo- meter.Direct measurement without pre-treatment is possible and the technique can be adapted for continuous irt vivo recording. The voltage of such cells, when giving negligible current, approximates to the Nernst equation: E = Eo + RT/zF In a'/ao where a' and ao are activities of the ion. This response to changes in activity rather than total concentration can be a fundamental advantage in investigating the distribution of ions across biological membranes. The development of ion-selective glasses that respond to cations other than hydrogen led to great interest in this method of analysis,l but for potassium a glass with a selectivity of much better than 10: 1 over sodium was not achieved.Attempts were made to correct for the sodium error by using sodium and potassium glasses in combination and so correcting the potassium measurement~.~t~ Liquid ion exchangers based on the lipophilic tetraphenyl- borate salt of potassium achieved a selectivity of 80: 1 over odium,^ but were still not really adequate for use with plasma, where sodium is in 30-fold excess. The macrotetrolide antibiotics, particularly valinomycin, provided the first adequately selective ionophore. These molecules have the property of rendering oil films selectively permeable to potassium ions. Valinomycin extracted from cultures of Streptomyces fulvissimusb was used initially to study permeability in mitochondria6 and artificial bi-layer membranes.' Its behaviour in rendering biological membranes selectively permeable to potassium was found to apply to macroscopic liquid membranes supported, for instance, in sintered glass,* and this property was tried as an analytical method in aqueous solutions9 and later in serum samples.lO Thin liquid membranes have serious limitations in biological use, however, owing to lack of structural stability and rapid deterioration. Valinomycin is an electrically neutral compound, acquiring a charge of +1 when carrying a potassium ion, which may be the reason why thick films of oil that contain valinomycin do not conduct * Presented at the Anglo - Czech Symposium on Polarography, London, September 13th to 15th, 1977.BAND, KRATOCHVIL, POOLE WILSON AND TREASURE 247 sufficiently well to operate as satisfactory electrodes.A combination of valinomycin and potassium tetraphenylborate, with the valinomycin in molar excess, combines high selectivity at the interface with conductivity in the bulk of the oil.ll However, such membranes still rapidly become contaminated on contact with plasma unless the surface is renewed continu- A poly(viny1 chloride) (PVC) membrane for calcium was investigated by Moody et aZ.12 and this flexible polymeric membrane was shown to be a considerable advance on the liquid membranes. Valinomycin is soluble in bis-2-ethylhexyl adipate , which acts as a plasticiser for PVC. The membrane electrode used for the present study has in addition potassium tetraphenylborate dissolved in nitrobenzene, which also acts as a plasticiser.This has the effect of lowering the electrical resistance13 and improving the anion reje~ti0n.l~ This membrane has been incorporated in a thermostated cell designed for use with samples of and has been used for in vivo recording as an intravascular catheter tip membrane.ls Although the selectivity of membranes incorporating valinomycin for inorganic species has been firmly established, the possibility of interference by lipid-soluble or surface-active components of the blood exists. Further, as the activity coefficient for potassium in complex biological fluids is not known and even the possibility of binding of the ion by the plasma proteins has been suggested,17 it follows that before such electrode results can be fully interpreted their relationship to conventional concentration measurements must be deter- mined.In this paper an attempt has been made to assess the magnitude of the problem by simply comparing duplicate electrode and emission flame photometer measurements using samples of human blood. The ion-selective cell used, together with some preliminary results, has been demonstrated to the Physiological Society.15 ously. Methods Composition and Preparation of Potassium-selective Membranes Potassium tetraphenylborate was prepared by the method of Vogel18 by precipitation from sodium tetraphenylborate. The valinomycin, potassium tetraphenylborate and PVC were weighed into the same glass specimen tube ; the nitrobenzene, dioctyl adipate and finally the tetrahydrofuran were added. The mixture was stirred magnetically for 2 h before pouring it into a PTFE (Fluon; ICI Ltd.) casting pot.lg It was then covered loosely and left to evaporate slowly over 24 h.The resulting membrane measured 2 cm in diameter and was about 1 mm thick. These mem- branes were stored dry at room temperature. The membrane constituents are shown in Table I. TABLE I COMPOSITION OF THE PVC POTASSIUM-SELECTIVE MEMBRANE Component Amount Potassium tetraphenylborate . . .. .. . . . . 0.000025g Valinomycin (Sigma) . . .. .. .. .. .. 0.0015g Bis-2-ethylhexyl adipate (dioctyl adipate) (Merck) . . . . 0.15 g Nitrobenzene (BDH Chemicals) .. .. .. . . 0.05 g High-molecular-weight PVC (BDH Chemicals) .. . . 0.075 g Tetrahydrofuran (BDH Chemicals) . . . . .. .. 3ml Construction of the Cell Portions of the membrane were cut and glued to the ends of 3-cm lengths of 4 mm 0.d.PVC tubing, using a solution of PVC in tetrahydrofuran. A PVC washer was glued around the edges of each membrane [Fig. 1 (a)] so that an electrically tight junction resulted. Silver - silver chloride internal reference electrodes were made in batches from Diamel- coated silver wires 0.013 mm in diameter (Johnson Matthey Metals Ltd.). The varnish was stripped off the ends to a distance of about 1.5cmJ and after cathodic cleaning the silver was chloridised using 1.5 V and a current density of 10 mA cm-2. Each electrode was completed by filling the PVC tube with 4 mmol l-1 potassium chloride internal reference solution and securing the internal reference electrode.248 BAND et aZ. : RELATIONSHIP BETWEEN ACTIVITY AND Analyst, VoZ.103 Potassium- selective membrane Ion-selective half cell as in (a) Fig. 1. (a) The potassium-selective half-cell. (b) Thermostated Perspex jacket containing the potassium-selective half-cell and cuvette. The electrode was held in a machined Perspex (ICI Ltd.) holder [Fig. 1 ( b ) ] so that the membrane and a small portion of the surrounding washer formed one side of a 50-pl cuvette. Samples were introduced from a syringe, filling the cuvette from below and also filling the polythene capillary tube that forms the outlet. These outlet tubes were made from the replaceable capillaries used in conjunction with the Radiometer K497 reference electrode system for blood pH measurements and served the same purpose. The cell was completed by dipping this capillary into the saturated potassium chloride bridge of the calomel reference electrode (Radiometer K497).The electrode, liquid junction and calomel reference electrode were maintained at 37 "C by circulating water. Calibration For evaluation of the membranes, cell voltages were measured with a Vibron 33B electro- meter (EIL Ltd.) in conjunction with a precision voltage source (Time Electronics Ltd.) used potentiometrically. Both pure potassium chloride and potassium chloride - sodium chloride standards were used. All of the electrodes showed a linear relationship between the logarithm of the calculated activity and voltage for the pure potassium standards and a linear relationship between the logarithm of the concentrations and voltage for the potassium chloride - sodium chloride standards (showing a constant activity coefficient due to the excess of sodium chloride).The slopes of the lines for all electrodes were between 97 and 99% of the theoretical Nernst response. The compositions are given in Table 11. TABLE I1 POTASSIUM STANDARDS USED TO STUDY THE ELECTRODE AND CALIBRATE IT FOR PLASMA MEASUREMENTS Mixed potassium chloride - sodium chloride standards Pure potassium chloride r A \ standards : K+/mmol 1-1 K+/mmol 1-1 Na+/mmol 1-1 1 1 140 1.5 140 2 2 140 2.5 140 3 3 140 3.5 140 4 4 140 6 5 140 7 140 10 10 140 For convenience in the comparison of the blood measurements with the flame photometric results, an electrometer with a logarithmic scale was used (the p,,, scale of a Radiometer PHM 27), calibrated in concentration terms using the potassium chloride - sodium chloride standards.The results are given directly as the readings on this scale and are thereforeMarch, 197'8 CONCENTRATION MEASUREMENTS OF PLASMA POTASSIUM 249 strictly activity relative to the activity of potassium in the standards (activity coefficient 0.72), rather than concentrations. This is a practical convenience as it avoids two systems of labelling for the standard solutions for the two instruments. Blood Samples Heparinised venous blood samples were obtained from patients in St. Thomas's Hospital Intensive Care Unit, at the same time as routine tests were being performed. Part of each sample was read directly in the electrode, the other was centrifuged and the plasma potassium was measured both in the electrode and with an emission flame photometer (IL 343) cali- brated with the same standards.When a whole blood sample is in the electrode cuvette, the potassium in the plasma determines the voltage and intracellular potassium does not contribute. All measurements were made in duplicate. Results The results from 50 blood samples are summarised in Fig. 2, where the electrode results for the whole blood and plasma are compared with the emission flame photometric readings on plasma. The span and stability of the electrode cell were not affected by exposure to blood or plasma. Most of the readings with the electrode were reproducible to better than 0.01 mmol l-l; the maximum difference was 0.04 mmol l-1. Potassium in plasma by ion-selective electrode/mmol I-' 1 2 3 4 5 6 " 1 2 3 4 5 6 Potassium in plasma by emission flame photometer/mmol I-' Fig.2. (a) A comparison of potassium determinations in whole blood and separated plasma of the same sample, both measurements made with the ion-selective electrode. The line of identity and the regression line (Y = - 0.073 + 1.045X) are drawn. (b) A comparison of potassium determinations in whole blood (ion- selective electrode) and separated plasma (emission flame photometer). The line of identity and the regression line (Y = 0.042 + 1.043X) are drawn. (G) A comparison of ion- selective electrode and emission flame photo- meter determinations in plasma. The line of identity and the regression line (Y = 0.111 + 0.998X) are drawn. Potassium in plasma by emission flame photometer/mmol I-'250 Analyst, VoZ.103 Fig. 2 (a) shows the electrode results for the whole blood plotted against the results for plasma. The linear regression line has a slope slightly greater than unity, i.e., the electrode tends to read higher on the whole blood. This is reflected in the graph of the electrode results for whole blood against the molar concentrations determined by the flame photometer [Fig. 2 ( b ) ] , whereas the slope of the line for the electrode results on the plasma against the molar concentration approaches unity [Fig. 2 ( c ) ] . A convention of linear scales has been adopted in these plots for clarity in the comparison of the two methods. The discrepancy in the electrode results on whole blood and plasma was further analysed by subtracting -log,, of the readings for each pair.The mean difference was -0.011 (standard deviation 0.006) “pIon” unit; n = 50, $J < 0.001. BAND et aZ. : RELATIONSHIP BETWEEN ACTIVITY AND Discussion The performance of the electrode was not impaired by exposure to stagnant blood even for long periods. This differs from our experience with earlier liquid membrane electrodes and shows that if the PVC membrane is indeed poisoned by blood, the effect is a t least immediately reversible. The electrode consistently read higher on whole blood than on its separated plasma. Leakage of potassium from cells does not cause a significant increase in the plasma potassium in heparinised samples for at least 1 h. In this work the electrode measurements in the plasma of whole blood were made before centrifugation of the samples and therefore leakage could not have contributed to the discrepancy between the plasma and whole blood deter- minations.An explanation for this effect can be found in the work of Sigaard Andersen,20 who investigated the properties of liquid junctions and reference systems in conjunction with blood pH measurements. He reported an apparent difference of 0.010 & 0.005 pH between blood and plasma, the blood appearing to be the more acidic, using a saturated potassium chloride junction at 38 “C of similar configuration to that selected for the present study. The difference was attributed to red blood cells affecting the liquid junction potential of the cell (I?$). In order to compare our results, the differences in the logarithm of the readings for each blood - plasma pair were calculated.The mean difference was found to be -0.011 (standard deviation 0.006) "pion" unit. The very close agreement for two completely different ions strongly supports Sigaard Andersen’s conclusion. The actual magnitude of the difference in E, for blood - saturated potassium chloride and plasma - saturated potassium chloride would therefore be approximately 0.6 mV. A constant voltage error of this kind would have the effect of altering the slope of line shown in Fig. 2 (a). The comparison of the electrode readings on plasma and the concentration measured using the emission flame photometer can now be considered. A discrepancy between the activity coefficients for potassium in plasma and the potassium chloride - sodium chloride standards, any binding of potassium to plasma proteins or the effect of a liquid junction potential formed between plasma and saturated potassium chloride could affect the slope of the line.Although the results shown in Fig. 2 (c) are close to a line with a slope of unity, this may be slightly misleading as the flame photometric values are molar concentrations. A correction to concentrations relative to plasma water would be about 5%, giving the line a slope of greater than unity. It seems likely, therefore, that the results contain some self- cancelling errors. One possibility is to assume that potassium ions exist free in solution in the plasma water and have an activity coefficient very close to that for equivalent concentra- tions of potassium in physiological saline (0.72), but that the plasma proteins contribute approximately 1 mV to the E, value above the effect of the red blood cells.Whatever the explanation, the observation is of interest in the calculation of the electrochemical gradient for potassium across biological membranes, when concentration measurements only are available. That the choice of reference liquid junction is extremely important in interpreting the results of ion-selective electrodes in biological work is shown by further findings of Sigaard Andersen20 on the composition of the salt bridge. He showed that either physiological potassium chloride or sodium chloride, in comparison with saturated potassium chloride, would change the biological pH scale by -0.105 or -0.136, respectively.This would not be of fundamental importance for pH measurements where a conventional activity scale is all that is available, but would drastically affect the relationship between electrode andMarch, 1978 CONCENTRATION MEASUREMENTS OF PLASMA POTASSIUM 251 concentration measurements for other ions. With the sodium chloride bridge, the potassium electrode would read 8.4 mV higher on each blood sample, which would give the appearance of 37% higher activity in blood than in a potassium chloride - sodium chloride standard of the same concentration-the activity coefficient for potassium in plasma would appear to be greater than unity! The finding that the saturated potassium chloride liquid junction at 37 “C gives an apparent activity coefficient for potassium in plasma within a few per cent.of that for physiological saline gives some justification for the belief that this bridge does reduce the residual liquid junction potential. This is further supported by similar findings for the sodium ion using glass electrodes.21 As both the activity coefficient and the liquid junction potential affect the slope of the line relating activity to concentration, it is not possible to separate the two effects. Any scale of activity is therefore operational in that it depends upon the bridge used. The saturated potassium chloride junction was chosen for the present work because its repro- ducibility and characteristics have been extensiveIy investigated in relation to the biological pH scale. The results are, however, sufficiently consistent to suggest that the PVC-based membrane electrodes offer a useful means of studying the state of ions in the blood. Further, more accurate work may include the use of theoretically more equitransferent bridges and accurate determination of plasma water concentration. 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. References Eisenman, G., Rudin, D. 0.. and Casby. J. U., Science, N.Y., 1957, 126, 831. Dahms, H., Clin. Chem., 1967, 13, 437. Nefi, G. W., Radke, W. A., Sambucetti, C. J., and Widdowson, G. M., Clin. Chem., 1970, 16, 566. Wise, W. M., Kurey, M. J., and Baum, G., Clin. Chem., 1970, 16, 103. Brockman, H., and Schmidt-Kastner, G., Chem. Ber., 1955, 8, 57. Moore, C., and Pressman, B. C., Biochem. Biofihys. Res. Commun., 1964, 15, 562. Mueller, E. E., and Rudin, D. O., Biochem. Biophys. Res. Commun., 1967, 26, 398. Stefanac, Z., and Simon, W., Microchem. J., 1967, 12, 126. Pioda, L. A. R., Stankova, V., and Simon, W., Analyt. Lett., 1969, 2, 665. Pioda, L. A. R., Simon, W., Bosshard, H.-R., and Curtius, H. Ch., Clinica Chim. Acta, 1970, 29, 289. Ebden, P., PhD Thesis, “Studies of the Applicability of Sodium and Potassium Ion-selective Electrodes to Measurement in Plasma,” University of London, 1975. Moody, G. J., Oke, R. B., and Thomas, J . D. R., Analyst, 1970, 95, 910. Band, D. M., and Kratochvil, J., J . Physiol., 1974, 239, 1OP. Morf, W. E., Kahr, G., and Simon, W., Analyt. Lett., 1974, 7, 9. Band, D. M., Kratochvil, J., and Treasure, T., J . Physiol., 1977, 265, 5P. Treasure, T., and Band, D. M., J . Med. Engng Technol., 1977, 1, 271. Gerbrandy, J., Hellendoorn, H. B. A., deVries, L. A., and Van Leeuwen, A. M., Scand. J . Clin. Lab. Vogel,. A. I., “Quantitative Inorganic Analysis,” Third Edition, Longmans, London, 1961. Kratochvil, J., Band, D. M., and Treasure, T., Lab. Pract., 1977, 26, 102. Sigaard Andersen, O., Scand. J . Clin. Lab. Invest., 1961, 12, 125. Moore, E. W., and Wilson, D. W., J . Clin. Invest., 1963, 42, 293. Invest., 1957, 10, 272. Received September 21st, 1977 Accepted October 24th, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300246

出版商:RSC    

年代:1978

数据来源: RSC