摘要:

11Analytical Journals fromThe Royal Society of ChemistryThe AnalystAn inpernational journal of high repute containingOrlginOl research papers on the theory and pmctice ofoll aspects of anatytiwl chemistry drawn from a widerangeof sources. tt also publishes regular critical reviewsof Important techniques and their applications, shortpapers and urgent communications (which arepublished in 5-8 week) on important new work, andbod< reviews. Speclal issues devoted to majorconferences and particular analytical themes areregulaty published.Over 400 papers are submMed to The Analyst each year,70% of which are from outside the UK; papers aremceived from about 40 different countries annually.Wm, dfculdon to nearly 100 countries and 70% ofsutwcribers being outside the UK The Anulyst is a trulyintenrotlonal journal.Rapid publication is a feature ofthe/ournal, about 20% of all the papers being publishedIn 25 weeks or less from receipt and the median time topublicotron from occeptance being about 20 weeks.Subsuibers are therefore assured of reading the lutest,upbdate wok in The Analyst.12 htrrwr per annumSbsdpWm 1 986 €1 47.00 ($285.00) o/seas €1 62.00Special Issues of The AnalystSpedd issues of The Analyst are available seporatetyCOnDdnlng papers from The First and Second Biennial NuliarmiAkmk spectsoopV Symposia and SAC '83.mkkvvkrg ksues am Stln awllabk:The kralyst. February 1983 containing papers from the FirstBiem# National Atomic Speclroscopy Symposium.fhehlyst, March 1984 contahing papers from the SAC '83contererrce.fhel\nolYst.Moy 1985contuinjng papers from The Second Biennial Natlonol Atomic v Sympoaium.Rk. tl2.00 ($22.00)S9.50 ($1 9.50)Rk.S9.50~17.00)--m:Orders should be sent to The Royal Society of ChemishyDislrlbution Centre, Blackhorse Road, Letchworth, Herts.S G 6 1 ~ U KW-RSC member prices quoted. RSC members are entitled toodkcwnton mostpublMions Detalls wailable from:AsslsfantMembmh/p Omcer. The Royul socletyof C h e m m 30 Ruse// W m Londorr H B 5 D l : UKossmces quoted.Anal ytica I ProceedingsAnu/yfku/ Proc&hgs Is the news and informationjournal of the Analytical Divislon of the Royal Society ofChemi- It contains special articles, reports of meetings,extended summaries of original papers presented atmeetings orgonlsed by the Analytical Division, safetyarticles, reports of recent legislation, details of newequipment, and many other Items of general interest toanalytical chemists both in Main and overseas.This journal provides authors wlth an excellentopportunity to publtsh details of ongoing projects andinterim results for methods which are not yet sufficienllyadvanced to be submttted to o primary journal, tocontribute non-comprehenslve reviews of small areas ofa particular discipline and, via the Analytical Viewpointscolumn, to submlt philosophical contributions discussingthe future of o particular technique or controversial issuesaffecting anatytical chemists.Letters to the Editor arealso published, making the p a g e s of Anu~cdProceedings the most important forum for the discussionof issues of analytical importonce.12 bsws p e r annumSubrcrfptkn 1986 €69.00 ($1 34.00) o/seu~~ €76.00Analytical AbstractsAnu/flku/Abstrucfs endeavours to cover the whole fieldof anatytical chemistry, providing around 12,000abstracts per annum of papers and books consideredto be of importance and interest to analytical chemists.The journal comprises nlne main section& generalanalytical chemistry; lnorganlc chemistry; organicchemistry; biochemm phormaceUnwl chemistry;food; agriculture environmental chemistry; andapparatus and techniques.A subject index Is providedin each issue and a volume index is produced annually.12 issws per annumSubscription 1986 C219.50 ($425.00) o/sw~ €241 .OOAnalytical Abstracts OnlinePlans are in hand to make Analylkol Abstracts accessibleonline.Further detulls may be obtained from Sales & PromotionDept. The Royal Soclely of Chemistry, The UnlversM NofflnghamNG7 2RD England.Package Prices 1986The Analyst Analytical Abstracts and AnalyticulProceedings€375.00 ($726.00) o/seas S412.00The Analyst and Analytical AbstractsE329.00 ($636.00) O/MS €361 .OOThe Analyst and Anulyticof Proceedings€1 84.00 ($356.00) o/teas €202.0IIIIIIIIIIIIIIIITHE ANALYST READER ENQUIRY SERVICE FEB'86For further information about any of the products featured in the advertisements in this issue, please writethe appropriate number in one of the boxes below.Postage paid if posted in the British Isles but overseas readers must affix a stamD.PLEASE USE BLOCK CAPITALS LEAVING A SPACE BETWEEN WORDS Valid 12 monthsI n l l l i I i LI 1 - l_Ii-lM I 1 1 - Ill I 1 L 1 - 1 I11-1 I I I I I I I I E I : D n - m : E mI I I I I I I I illlllllllllllllill" IrilI I I L I I I I T l _ l l r T - T I T 1 l r I I I I I 1 I 1 1 I I rmiii? POSTCODE I 1 - 1 1 1 I I-!I I I I I I I Imi I r i I I I I I I I -TI1 IT-- I I I i r - r mL I 1 I I I I I I I l I T n 1 - r m : T m m - mI I I I I I 1 I I I 1 I I r-1 I I 1 I I i 1 I I I I I1 NAME-~ 2 COMPANYPLEASE GIVE YOUR BUSINESS ADDRESS IF POSSIBLE.IF NOT, PLEASE TICK HERE3 STREET4 TOWN5 COUNTY '6 COUNTRY7 DEPARTMENT/DIVISION- 8 YOUR JOB TITLE/POSITION m m OFFICE USE ONLY REC D L 1 L-Lj PROC D L u A 9 TELEPHONE NOFOLD HEREiII IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII1Postagewill bepaid byLicenseeDo not affix Postage Stamps if posted in Gt. Britain,Channel Islands, N. Ireland or the Isle of ManBUSINESS REPLY SERVICELicence No. WD 106Reader Enquiry ServiceThe AnalystThe Royal Society of ChemistryBurlington House, PiccadillyLONDONWIE 6WFEngland250v,InUI m50 mPIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII1IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII!IIIIIIIIIIIIIIIIIIIIIIIIIIIII

ISSN:0003-2654    

DOI:10.1039/AN98611BP005

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALAO 1 1 l(2) 129-264 (1 986)The AnalystFebruary 198612913313914515115716316717117517918318919319720120520921 321 7221225227231The Analytical Journal of The Royal Society of ChemistryCONTENTSDetermination of Hydroquinone in Skin-toning Creams Using High-performance Liquid Chromatography-Jane Firth,Ian RixStability-indicating Assay for Oxyphenbutazone. Part II. High-performance Liquid Chromatographic Determination ofOxyphenbutazone and Its Degradation Products-Huguette Fabre, Andrianandrasana Ramiaramana, Marie-Do m i n iq ue B I a nc h in, Bernadette Ma nd ro uPrevention of Artifactual Formation of Nitrosamines During the Analysis of Baby Bottle Rubber Nipples-Nrisinha P.Sen, Stephen W. Seaman, Santosh C.KushwahaGas Chromatographic Determination of Triclopyr in Environmental Waters-Tadashi Tsu kioka, Ryuzo Takeshita,Tetsu ro M ura kamiPreparation of a Chloride-selective Electrode Based on Mercury(1) Chloride - Mercury(l1) Sulphide on an ElectricallyConductive Epoxy Support-J. L. F. C. Lima, A. A. S. C. MachadoReduction in Size by Electrochemical Pre-treatment at High Negative PoYentials of the Background Currents Obtainedat Negative Potentials at Glassy Carbon Electrodes and Its Application in the Reductive Flow InjectionAmperometric Determination of Nitrofurantoin-Ahmad 8. Ghawji, Arnold G. FoggDetermination of Vitamin C by Flow Injection Analysis-Fernando Lazaro, Angel Rios, M. D. Luque de Castro, MiguelVa lcarcelDetermination of Vitamin C in Urine by Flow Injection Analysis-Fernando Lazaro, Angel Rios, M.D. Luque de Castro,Miguel ValcarcelFlow Injection - Hydride Generation System for the Determination of Arsenic by Molecular Emission CavityAnalysis-M. Burguera, J. L. BurgueraDetermination of Sulphur, Phosphorus, Magnesium, Silicon and Aluminium in Washing Powders by X-RayFluorescence Spectrometry-Joh n Wil hamsDetermination of a-Impurities ir the fi-Polymorph of lnosine Using Infrared Spectroscopy and X-ray PowderDiffraction-David H. Doff, Frank L. Brownen, Owen I. CorriganContinuous Spectrophotometric Monitoring of Chlorine in Air-Aviva Shina, J. GabbayMorpholine as an Absorbing Reagent for the Determination of Sulphur Dioxide-V. Raman, J. Rai, M. Singh, D. C.Kinetic - Catalytic Determination of Cobalt by Oxidation of Pyrogallol Red by Hydrogen PeroxideM.Llobat-Estelles,Effect of Anion-exchange Resin on the Formation of Iron(ll1) - Tiron Complexes-Mohamed M. A. Shriadah, KunioMulti-component Quantitative Analysis of Fluorescent Mixtures Not Obeying Beer’s Law-A. T. Rhys Williams, R. A.Minimisation of Bilirubin Interference in the Determination of Fluorescein Using First-derivative SynchronousNovel Aryl Oxalate Esters for Peroxyoxalate Chemiluminescence Reactions-Kazuhiro Imai, Hiroyoshi Nawa, MotoakiAccurate Determination of Platinum, Palladium, Gold and Silver in Ores and Concentrates by Wet Chemical Analysis ofGravimetric Determination of Iron by Precipitation as the [(C,H,),N],[Fe(SCN),] Ion Pair-J. Hernandez Mendez,ParasharA.Sevillano-Cabeza, J. Medina-Escriche0 hze kiSpraggExcitation Fluorescence Spectroscopy-Frank V. Bright, Linda 8. McGownTana ka, H iroshi Og atathe Lead Assay Button-Andreas DiamantatosA. Alonso Mateos, E. J. Martin MateosSHORT PAPERSSize Distribution of Particulate Matter in Exhaust Gases in Inductively Coupled Ptasma Atomic EmissionSpectrometry-J. Alwyn Davies, Adrian C. JefferiesSequential Determination of Arsenic, Antimony and Bismuth in Low-alloy Steels by Hydride Generation InductivelyCoupled Plasma Atomic Emission Spectrometry-Stephen J. WaltonDetermination of Rock-forming Elements in the Presence of Large Amounts of Uranium in Zirconium Using InductivelyCoupled Plasma Atomic Emission Spectrometry-Irene Gal, Ludwik HaliczComparison Between lsobutyl Methyl Ketone and Diisobutyl Ketone for the Solvent Extraction of Gold and ItsDetermination in Geological Materials Using Atomic Absorption Spectrometry-Charles H.Branch, DawnHutchisoncontinued inside backcoverElectronically typeset and printed by Heffers Printers Ltd, Cambridge, Englan235 Study of the Fluorescence of the Lead - Morin System in the Presence of Non-ionic Surfactants-J. Medina,F. Hernandez, R. Marin, F. J. Lopez239 Thin-layer Chromatographic Separation and Determination of Dibutylphosphoric Acid in a Mixture of Monobutyl-phosphoric Acid and Tributyl Phosphates. C. Tripathi, A. Ramanujam, M. N. Nadkarni, C. Bandyopadhyay241 Extraction and Gas Chromatographic Determination of Residual Formaldehyde in Micro-surgical Materials-AntonellaProfumo, Maria Pesavento243 Diazotised Sulphanilic Acid as a Spectrophotometric Reagent for the Determination of Trace Amounts of lndole inAqueous Solution-Ahmad K.Ahmad, Younis I. Hassan, W. A. Bashir245 Selective Spectrophotometric Kinetic Determination of Cobalt with o-Hydroxyphenylthioure-s. J. Rao, G. S. Reddy,J. K. Kumari, Y. K. Reddy247 Spectrophotometric Determination of Selenium(lV) with Potassium Butyl Xanthate-Nepal Singh, Arvind K. Garg249 Oxidative Amperometric Flow Injection Determination of Oxalate at an Electrochemically Pre-treated Glassy Carbon253 Z’-Mercapto-4-propylacetanilide: an Alternative t o Thionalide for Precipitating Lead from Weak Acid Solution-John255 Titrimetric Determination of Catecholamines and Related Compounds via Bromine Oxidation and Substitution-Electrode-Arnold G.Fogg, Rosa M. Alonso, Miguel A. Fernandez-ArciniegaC. Burridge, Irene J. Hewitt, Hamish A. AndersonDarwish AminCOMM U NlCATlON259 Supported Chemoreceptive Lipid Membrane Transduction by Fluorescence Modulation: the Basis of an IntrinsicFibre-optic Biosensor-Ulrich J. Krull, Chrisula Bioore, Gareth Gumbs263 BOOK REVIEWSERRATUM264 Automatic Two-stage Thermal Desorption Gas Chromatography for Low-volatility Organic Vapour Determination-J. F. Alder, E. A. Hildebrand, J. A. W. SykesThe Periodic Tableof the ElementsThe Royal Society of Chemistry has produced acolourful wall chart measuring 125cm x 75cmcovering the first 105 elements as they exist today.Each group is pictured against the same tintedbackground and each element, where possiblephotographed in colour aria discussed with regardto its position in the hierarchy of matter. Additionalinformation for each element includes chemicalsymbol, atomic number, atomic weight and orbitsof electrons.The chart is particularly useful for both teachersand students and would make a worthwhileaddition to any establishment.Price: Non-RSC Members f3.00 including VATRSC Members f2.00 including VATTeacher Members f 12.00 for 10 including VATRSC members should send their orders to: TheRoyal Society of Chemistry, The MembershipOfficer, 30 Russell Square, London WClB 5DT.Non-RSC members should send their orders to:The Royal Society of Chemistry, DistributionCentre, Blackhorse Road, Letchworth, HertsSG6 IHN.Please contactBUREAU OF ANALYSED SAMPLES LTDfor a copy of their list ofOVERSEASREFERENCE MATERIALSproduced byAECAN (Canada)BAM (W.Germany)CANMET (Canada)CKD (Czechoslovakia)CTIF (France)IRSID (France)NBS (USA)SABS (S. Africa)JERNKONTORET (Sweden)Please write, telephone or telex to:BAS Ltd., Newham Hall, Newby,Middlesbrough, Cleveland, TS8 9EATelephone: Middlesbrough (0642) 317216Telex: 587765 BASRIDA201 for further information. See page iv

ISSN:0003-2654    

DOI:10.1039/AN98611BX007

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 129 Determination of Hydroquinone in Skin-toning Creams Using High-performance Liquid Chromatography Jane Firth and Ian Rix Consumer Hazards Group, Laboratory of the Government Chemist, Corn wall House, Waterloo Road, London SE18XY, UK A simple, rapid reversed-phase HPLC method for the determination of hydroquinone in skin-toning creams is described that is suitable for routine use. The sample is dissolved in methanol or methanol - light petroleum and directly injected without further purification. The method successfully passed a ruggedness test and was applied to a range of 35 creams. Keywords: Hydroquinone determination; high-performance liquid chromatography; skin-toning creams; cosmetic product The major user of skin-toning creams in the UK is the West Indian population.The creams are applied to even out the skin colour on the facial areas. They are believed to function by decolourising the melanin in the skin and preventing new melanin being formed. The most favoured material for use in skin-toning preparations is hydroquinone, which was reported by Spencer1 to be effective at 1.5-2% in a vanishing cream producing a temporary lightening of skin colour. Spencer found that a concentration of 5% was liable to cause redness and burning. Cases of patchy de-pigmentation have arisen following the use of some skin-toning creams available on the retail market.2 Council Directive 76/768/EEC of the European Communi- ties makes the general point that cosmetic products must not be harmful under normal or foreseeable conditions of use and specifically allows hydroquinone to be used in cosmetic products at a level of 2% m/m subject to certain conditions of use and warnings that must be printed on the label.The field of application is not specified in the basic Directive 76/768/ EEC.3 The second amendment , Council Directive 82/368/ EEC,4 specified the field of application as oxidising colouring agents for hair dyeing and excluded hydroquinone for use as a skin lightener from the scope of the Directive. The fifth Commission Directive 84/415/EEC5 permitted the use of hydroquinone as a localised skin-lightening agent subject to a maximum concentration of 2% m/m in the finished cosmetic product and a warning on the label containing the information "contains hydroquinone, avoid contact with the eyes, apply to small areas, if irritation develops discontinue use, do not use on children under the age of 12." Member States were asked to bring into force the laws, regulations and administrative provisions necessary to comply with the Directive by not later than 31st December 1985.The method described in this paper will be submitted to the EEC for consideration as the adopted method of analysis. Hydro- quinone is permitted at a concentration of up to 2% by mass under the Cosmetic Products (Safety) Regulations; Statutory Instrument 1984: No. 1260, which had to be complied with by 1st January 1986. However, recent studies have shown that some available products contain more than the permitted level of hydroquinone .6 In the light of these facts it was decided to carry out an extensive survey of skin-toning creams to determine their hydroquinone content.There is a dearth of literature on the determination of hydroquinone in skin-toning creams. The most recent pub- lished work of Popov and Yanishlieva7 involved extraction of Crown Copyright. hydroquinone with acetic acid and conversion of the hydro- quinone into p-benzoquinone, with subsequent spectrophoto- metric determination. The determination of hydroquinone in various sample matrices has been described although none included skin-toning creams.Gl1 The method described here is of wide application and is rapid, allowing ten samples and associated standards to be analysed in less than 3 h. Experiment a1 Apparatus Reversed-phase HPLC was performed at ambient tempera- tures using a Spectrophysics SP800 solvent delivery system and a Shimadzu SPD-MIA diode-array UV - visible spectro- photometric detector.Reagents All reagents were of analytical-reagent grade. The methanol was of solvent for liquid chromatography grade. Chromatographic Conditions The analytical column used was of stainless steel (250 mm X 4.6 mm i.d.) packed with Spherisorb S O D S of 5 pm. The sample injection volume was 10 p1. The mobile phase was methanol - water (10 + 90 V/V) pumped at a flow-rate of 1.5 ml min-1. The detector was operated at 226 nm with a sensitivity of 0.50 A full scale and a chart speed of 5 mm min-1. Skin-toning Cream Samples Samples of cream were bought from a number of retail outlets in the South London area.They were stored at room temperature throughout the investigation. Procedure Extraction of hydroquinone Transfer 0.05 g of cream containing 0.2-4.0% of hydro- quinone into a 10-ml calibrated flask and add 8.0 ml of methanol. Heat to 40 "C in a water-bath and shake occasion- ally until dissolved. Allow to cool and make up to the mark with methanol. If the cream fails to dissolve under these conditions repeat the procedure with 4.0 ml of light petroleum (60-80 "C boiling range) and make up to the mark with methanol.130 ANALYST, FEBRUARY 1986, VOL. 111 Table 1. Multi-factorial experiments for the ruggedness test Experiment Factor Ala . . . . Blb . . . . CIC . . . . Dld . . . . E/e . . . . F l f . . . . . . Glg . . . . Results . . . . Hydroquinone, YO 1 .. . . A . . . . B . . . . c . . . . D . . . . E . . , . F . . . . G . . . . 2.062 . . * . s Wavelength of measurementlnm . . Mass of cream taken for analysislg . . Detector type . . . . . . . . Columnlengthlmm . . . . . . Internaldiameter/mm . . . . . . Packing material . . . . . . . . Constructionalmaterial . . . . Mobile phase flow-ratelm1 min-1 . . Final volume of extracting solution/ml 2 A B D e g t 2.045 C f . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mobile phase composition (methanol - water) 3 4 5 6 A A a a b b B B C C C C d d d d E e e E F F f G 8 G g f U V W X 2.127 2.020 2.086 2.070 Factor 7 a b C D e G Y 2.100 f 8 a b D E F g 2.056 C 2 ~~~~~~~ ~~ A B C SPD-MIA D . . 250 . . 4.6 . . Spherisorb SS-ODS . . Stainless steel E .. 1.5 F . . 10 G . . 10 + 90 . . 226 . . 0.05 . . Shimadzu ~ ~~ a 230 b 0.08 Pye Unicam PU4020 d 100 3 Cp-t,-Spher C18 Glass e 2.0 25 g 12.5 + 87.5 C f t v) 0 5? U 0 5 10 0 5 10 Ti meimin Fig. 1. Typical chromatograms for ( a ) a standard solution containing 0.152 g 1-1 of hydroquinone and (b) a skin-toning cream containing 1.6494% m/m of hydroquinone Preparation of the stock standard solution Prepare a stock solution of standard hydroquinone by dissolving the solid hydroquinone in methanol at a concentra- tion of 10 g 1-1. Preparation of the calibration graph Prepare a range of solutions by diluting aliquots of the stock hydroquinone standard with methanol to 100 ml in calibrated flasks. Inject the standard solutions and measure the peak- height absorbance. A straight-line calibration graph of absor- bance versus concentration was obtained which passed through the origin.Table 2. Effect of each factor in the ruggedness test D, = 1/4 (S + t + u + v - w - X - y - Z) = 0.0145 D,= 1 / 4 ( ~ - t+ u - v + w - x + y - Z) =0.0460 D, = 1 / 4 ( ~ - t + u - v - w + X - y + Z) = 0.0160 Db = 114(~ + t - u - v + w + x - y - 2) = 0.0100 Dd= 1/4(s + t - u - v - w - x + y + 2) = 0.0100 Df= 114(~ - t - u + v + w - x - y + 2) = 0.0295 Dg = 1/4(s - t - u + v - w + x + y - 2 ) = 0.0155 Results and Discussion Under the experimental conditions used, hydroquinone had a retention time of 5.0 min. Fig. l ( a ) and ( b ) depict typical chromatograms of a standard solution and skin-toning cream. It can be seen that there is no overlap from other compounds present in the cream as these are not eluted by the relatively weak mobile phase composition employed in the method.Performance Characteristics The limit of detection, which was based on a solution containing 0.001 g 1-1 of hydroquinone and defined as 5 times the standard deviation of this standard, was found to be 0.13% mlm with 4 degrees of freedom. A skin-toning cream containing 2% mlm of hydroquinone produces a solution that gives an absorbance of approximately 0.17. A proprietary skin-toning cream containing 1.694% rnlm of hydroquinone and spiked with 2.000% rnlm of hydroquinone gave a recovery of 3.658 k 0.232% rnlm (95% confidence limits, 4 degrees of freedom). A proprietary moisturising cream base spiked with 2.000% mlm of hydroquinone gave a recovery of 2.045 k 0.0186% mlm (95% confidence limits, 4 degrees of freedom). The full analytical procedure was followed through for fiveANALYST, FEBRUARY 1986, VOL.111 131 Table 3. Results for a range of skin-toning creams Sample No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Test No. 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 Hydroquinone, Yo 5.40 5.69 8.30 8.07 2.25 2.20 1.94 1.86 2.09 2.08 5.99 6.28 5.55 5.41 1.90 2.01 1.75 1.60 1.74 1.61 1.88 1.87 1.96 1.90 5.20 5.08 2.06 1.93 2.37 2.24 2.11 2.20 5.10 4.86 1.52 1.56 Sample No. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Test No. 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 Hydroquinone, % 2.07 1.89 1.62 1.76 2.08 1.99 1.76 1.74 1.94 1.95 1.62 1.64 2.37 2.49 6.44 6.63 1.72 1.75 1.64 1.71 1.66 1.64 1.98 2.07 1.79 1.76 1.80 1.76 2.18 2.18 2.03 2.11 5.36 5.19 samples of the same tube of skin-toning cream; a mean of 1.694% mlm and standard deviation of 0.017% mlm of hydroquinone were found.Replicate injections of standard solutions of hydroquinone of concentrations of 0.010 and 0.200 g 1-1 gave standard deviations of 0.000299 and 0.00217, respectively (4 degrees of freedom), which is the same order of magnitude to that obtained with skin-toning creams, indi- cating that the extraction procedure does not contribute sig- nificantly to the over-all error of the procedure. Ruggedness Test The results presented so far indicate that the method is precise and without significant bias. However, these results were obtained by a single analyst using a rigidly defined set of operating conditions in one laboratory.It was thought desirable to simulate use of the method in other laboratories by altering slightly the various analytical parameters and determining the effect on the result. The Youden and Steinerl2 model was employed for this test. The ruggedness test is carried out by deliberately varying the factors that are likely to have an effect on the result, from value A to value a. Youden and Steiner give a set of multi-factorial experiments for varying up to seven factors simultaneously (Table 1). From these eight results the effect of each factor can be calculated for a proprietary skin-toning cream (Table 2). The standard deviation (a) of the eight individual results found on this occasion was 0.0334.Any valueof Dlcan be considered significant (P<0.05) if 1 D 1 > v 2 a As no value of D exceeds d- 20 the method can be considered to be rugged for the parameters chosen. v 2 a = 0.0472). Application to Samples The method was applied to a range of skin-toning creams. Four products were waxy in nature and required preliminary dissolution in light petroleum before addition of methanol. Each sample was analysed in duplicate and the results are given in Table 3. Eight products contained hydroquinone well in excess of the 2% mlm permitted by the Cosmetic Products (Safety) Regulations, 1984. In order to obtain some indication of the relative errors of the procedure the difference between the duplicate results was plotted against the mean value for each sample.This plot revealed two distinct clusters at about 2% mlm and 5.5% rnlm of hydroquinone. The ratio of the mean of the differences to the mean of the sample means was taken for each cluster and was found to be very similar at the two hydroquinone levels (0.0368 at 2% hydroquinone, 0.0371 at 5.5% hydroquinone) indicating that the relative errors of the procedure are not dependent on hydroquinone concentration in the sample. The method presented here allows the determination of hydroquinone in a complex sample matrix with relative ease, accuracy and precision and has been shown to be insensitive to changes in many instrumental parameters, permitting its application in a wide range of laboratories. References 1. 2. 3. Spencer, M. C., Arch. Dermatol., 1961, 84, 131. Ridley, C. M., Br. Med. J., 1984, 287, 1537. Council Directive 76/768/EEC, Off. J . Eur. Commun., 1976, L262.132 ANALYST, FEBRUARY 1986, VOL. 111 4. Second Amendment Council Directive 82/368/EEC, Of$ J . Eur. Commun., 1982, L167. 5. Fifth Amendment Commission Directive 84/415/EEC, Off. J. Eur. Commun., 1984, L228. 6. Bush, S. J . , personal communication, 9 November 1984. 7. Popov, A., and Yanishlieva, N., Fresenius 2. Anal. Chem., 1970, 249, 191. 8. Miller, R. L., Chromatogr. Newsl., 1981, 9, 10. 9. Greenlee, W. F., Chisn, J. P., and Richert, D. E., Anal. Biochern., 1981, 112, 367. 10. 11. 12. Sticher, O., Soldati, F., and Lehmann, D., Plant. Med., 1979, 35, 253. Raghavan, N. V., J. Chrornatogr., 1979, 168, 523. Youden, W. J., and Steiner, E. H., “Statistical Manual of the Association of Official Analytical Chemists,” AOAC, Wash- ington, DC, 1975. Paper A51283 Received August 2nd, 1985 Accepted August 22nd, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100129

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 133 Stability-indicating Assay for Oxyphenbutazone Part It.* High-performance Liquid Chromatographic Determination of Oxyphenbutazone and Its Degradation Products Huguette Fabre,t Andrianandrasana Ramiaramana, Marie-Dominique Blancslin and Bernadette Mandrou Laboratoire de Chimie Analytique, Faculte de Pharmacie, 34060 Montpellier Cedex, France A high-performance liquid chromatographic method is proposed for the simultaneous determination of oxyphenbutazone and six potential decomposition products, using a reversed-phase column and ultraviolet detection. The method is more sensitive than thin-layer chromatography and allows the determination of 0.1% of each degradation product (with respect to oxyphenbutazone). It has been applied to the analysis of commercial tablets, capsules and ointments.Keywords: Oxyphenbutazone determination; degradation products determination; high-performance liquid chromatography; stability-indicating assay; reversed-p hase chromatography In Part I,1 we outlined the difficulties relating to the establishment of a thin-layer chromatographic (TLC) method to be used as a stability-indicating assay of oxyphenbutazone. We proposed a quantitative thin-layer chromatographic pro- cedure that prevents air oxidation on the plate of oxyphen- butazone. As reversed-phase high-performance liquid chromato- graphy (HPLC) is particularly suitable for the determination of easily oxidised compounds, we propose here a reversed- phase HPLC procedure for separating and determining oxyphenbutazone and six potential decomposition products.Method Development Oxyphenbutazone and its potential decomposition products were chromatographed under different conditions in order to optimise the separation. The decomposition products, the formulae of which were given in Part I,1 were 4-hydroxy-4- butyl-l-phenyl-2-(4-hydroxy)phenylpyrazolidine-3,5-dione (I), 2-butyl-N-(4-hydroxyphenyl)-N’-phenylpropanediamide (11), 2-[1-phenyl-2-(4-hydroxy)phenylhydrazino]-3-0~0-2- butylpropionic acid (111), 4-hydroperoxy-4-butyl-l-phenyl-2- (4-hydroxyphenyl)pyrazolidine-3,5-dione (IV) , 2-butyl-N- { 3- [4-butyl-3,5-dioxo- 1-(4-hydroxyphenyl)-2-phenylpyrazolidin- 4-yl]-4-hydroxyphenyl}-N’-phenylpropanediamide (V) and 2-oxo-3-butyl-3-phenylcarbamoyl-5-hydroxyindole (VI). The starting solvent system was the mobile phase we previously used in the stability-indicating assay of phenylbutazone, viz., 0.1 M tromethamine (THAM) citrate buffer (pH 5.25) - acetonitrile (60 + 40).2 As this mobile phase could not achieve the separation of the compounds, the influence of pH and acetonitrile content of the mobile phase on the capacity factor was investigated.The method was developed using a mixed standard solution (5 pg ml-1) of each compound (Fig. 1). The capacity factor, k’, was calculated using the equation k’ = (u/L) (fR + I), where tR is the retention time of the compound, u the linear flow-rate of the mobile phase given by the supplier and L the length of the column. As a stability-indicating assay of oxyphenbutazone in drugs involves the determination of less than 1% of each decompo- sition product (relative to the drug), the resolution factor (R.F.) between oxyphenbutazone and VI (pH 4.1) and oxyphenbutazone and I (pH 5.25 and 6.4) was calculated using * For Part I of this series, see reference 1.t To whom correspondence should be addressed. the equation R.F. = 2(tR2 - tR1)/(W2 + w1), where t~~ and fR2 are the retention times and w1 and w2 the peak widths. The influence of the acetonitrile content of the mobile phase on the resolution factor is shown in Fig. 2. The pH values 5.25 and 6.4 gave a satisfactory resolution (R.F. > 2 with equal concentrations of each compound) but pH 6.4 was discarded because of the elution of compound I11 [Fig. l(a)] in the solvent peak (the mobile phase cannot be used as the solvent because of the instability of the compounds in the solvent system).Therefore, the mobile phase 0.1 M THAM citrate buffer (pH 5.25) - acetonitrile (65 + 35) was selected in further development of the method. In order to increase the resolution, the effect of the addition of tetrahy- drofuran (THF) (2-6%) on this mobile phase was investi- gated. The addition of 6% of THF increased the selectivity and gave a resolution factor R.F. = 1.32 between oxyphenbu- tazone and I using a concentration of 500 pg ml-1 of oxyphenbutazone and 5 pg ml-1 of I. The mobile phase A finally selected was 0.1 M THAM citrate buffer (pH5.25) - acetonitrile - THF (65 + 29 + 6). For the determination of V a more strongly eluting mobile phase B, 0.1 M THAM citrate buffer (pH 5.25) - acetonitrile (45 + 59, was selected from Fig.l(b). Experimental Apparatus A high-performance liquid chromatograph (Merck LMC System), equipped with a variable-wavelength ultraviolet detector (LC 313), a 10-p1 loop injector and fitted with a 25 x 0.4 cm i.d. stainless-steel cartridge, packed with 7 pm LiChrosorb RP-18 (Merck), was used. The column was equilibrated with the mobile phase for 30 min before use. Reagents and Materials Tromethamine and citric acid were of analytical-reagent grade. Oxyphenbutazone and its degradation products (I-VI) were the same as used previously.’ Tanderil ointment (5% oxyphenbutazone), Tanderil tablets (100 mg of oxyphenbutazone per tablet) were commercial formulations (Ciba-Geigy Laboratories). Kymalzone capsules (75 mg of oxyphenbutazone per tablet) were commercial formulations (Biocodex Laboratories).ANALYST, FEBRUARY 1986, VOL. 111 'bj 35 40 45 50 55 I C) \ 35 40 45 50 - 55 Ec------p-- Acetonitri le,% Fig. 1.Influence of the acetonitrile content of the mobile phase (0.1 M THAM citrate buffer - acetonitrile) on the capacity factors (k') of (0) oxyphenbutazone and its degradation products: (0) I, (A) 11, (D) 111, (A), IV, (+) V and (0) VI. (a) Buffer pH 4.10; ( b ) buffer pH 5.25; and ( c ) buffer pH 6.40 12 10 8 L w 0 m C + .s 6 - :: a 4 1 fn C 0 Q al U - (a L 0 Fig. 3. I 1 8 12 16 20 24 Time/m in Separation of oxyphenbutazone and potential degradation products under the optimised conditions. (a) Mobile phase A, 0.1 M THAM citrate buffer (pH 5.25) - acetonitrile - THF (65 + 29 + 6); flow-rate, 1.3 ml min-1; pressure, 98 bar; detector sensitivity, 0.02 a.u.f.s.; chart recorder speed, 0.5 cm min-'; wavelength, 239 nm; oxyphenbutazone, 1.15 pg ml-1; I, 1.30 pg ml-l; 11, 1.75 pg ml-I; 111, 1.37 pg ml-1; IV, 1.47 pg ml-1 and VI, 1.30 pg ml-1.( b ) Mobile phase - _ _ Fig. 2. Resolution factor versm the acetonitrile content of the mobile phase (0.1 M THAM citrate buffer - acetonitrile): (A) between oxvDhenbutazone and VI, buffer pH 4.10; (H) between B, 0.1 M THAM &rate buffer (PH 5.25) - acetonitrile (45 + 55); flow-rate, 1.8 ml min-l; pressure, 105 bar; detector sensitivity, 0.02 a.u.f.s.; chart recorder speed, 0.5 cm min-1; wavelength, 239 nm; oxyphenbutazone and 1, butter pH 5.25; and (0) between oxyphen- butazone and I, buffer pH 6.40 Table 1.Chromatographic parameters for oxyphenbutazone and its decomposition products oxypnenoutazone 1.17 pg ml-1; IV, 1.22 pg ml-1; V, 1.37 pg ml-I; VI, 1.67 pg ml-1 pg mIPL; 1, 1.w pg mlr L; 11, 1.J.L pg 1111 *; 111, Mobile phase A Mobile phase B Compound k' Hlmm As k' Hlmm As 1.16 0.35 0.067 - I . . . . . . 4.11 0.077 1 .oo 0.43 0.062 1 .oo I1 . . . . . . 11.32 0.040 1.16 1.08 0.042 1.20 I11 . . . . . . 0.74 0.166 1 .oo 0.00 IV . . . . . . 6.79 0.052 1.25 0.64 0.046 1 .oo v * . . . . . 3.26 0.043 1 .oo VI . . . . . . 8.79 0.047 1.40 0.88 0.043 1.30 Ox.* . . . . 2.63 0.460 - * OxyphenbutazoneANALYST, FEBRUARY 1986, VOL. 111 135 Table 2. Repeatability, sensitivity and detectability data for oxyphenbutazone and its decomposition products Repeatability, Sensitivity/ Detectability*/ Compound a,% mm pg-1 Ox.? .. _. . 0.72 2 435 0.0040 I . . . . . . 0.75 6 615 0.0015 I1 . . . . . . 1.26 3 314 0.0030 I11 . . . . . . 0.68 11 387 0.0010 IV 1.18 2 041 0.0050 V . . . . . . 1.59 7 299 0.0010 VI . . . . . . 0.89 3 231 0.0030 * Determined in the presence of oxyphenbutazone (5 pg). t Oxyphenbutazone. . . . . . . Table 3. Recovery of oxyphenbutazone and its degradation products from pharmaceutical formulations Formula- Initial Amount Amount tion Compound contentjmg added/mg found/mg Ointment. . . .Ox.* I I1 I11 IV V VI I I1 I11 IV V VI I I1 111 IV V VI * Oxyphenbutazone Tablet . . Ox.* Capsule . . Ox.* 50.200 0.034 0.009 Traces - - - 94.000 0.394 0.014 Traces Traces 78.750 0.065 - - - - Traces 0.051 - - 0.131 0.176 0.138 0.148 0.138 0.131 0.259 0.385 0.274 0.294 0.326 0.259 0.195 0.263 0.206 0.221 0.210 0.195 - - - 0.161 0.172 0.142 0.003 0.141 0.132 0.678 0.375 0.296 0.363 0.325 0.268 0.264 0.263 0.212 0.216 0.218 0.244 - - t s O n 0) a I X U 1 I I 4 8 12 16 20 24 Methanol, acetonitrile and THF were of HPLC solvent grade.Distilled water was filtered through a 0.45-pm filter (Millipore). Standard Solutions A mixed stock standard solution containing 250 pg ml-1 each of oxyphenbutazone and I-VI was prepared in methanol. This solution was suitably diluted with methanol to give a concen- tration range from 1 to 10 pg ml-1. Test Solutions Ointment An accurately weighed amount of about 250 mg of ointment was sonicated for 5 min with 25 ml of methanol in a 50-ml centrifuge tube. The emulsion was rotated at 4000 rev min-1 for 10 min.Tablets The coatings of ten tablets were carefully removed with a cutter and the average mass of one core was determined. The ten cores were combined and powdered in a mortar. A core mass of about 20 mg was accurately weighed into a 50-mi centrifuge tube, then sonicated for 5 min with 25 ml of methanol. The suspension was rotated at 4000 rev min-1 for 10 min. Capsules An accurately weighed amount of about 34 mg of capsule powder was sonicated for 5 min with 25 ml of methanol in a 50-ml centrifuge tube. The suspension was rotated at 4000 rev min-1 for 10 min. After centrifugation, the supernatant from the ointment, tablets or capsules was injected on to the chromatograph without dilution for the determination of the degradation products, then diluted (1 + 399) in methanol for the determination of oxyphenbutazone.Chromatography Duplicate injections (10 pl) of each standard solution and test solution were injected under the following isocratic condi- tions: mobile phase A, 0.1 M THAM citrate buffer (pH 5.25) - acetonitrile - THF (65 + 29 + 6); flow-rate 1.3 ml min-1; Time/min Fig. 4. Chromatograms of a test solution from a tablet formulation. ( a ) Without addition of degradation products; mobile phase A. ( b ) Spiked with I-IV and VI; mobile phase A. ( c ) Without addition of degradation products; mobile phase B. ( d ) Spiked with I-VI; mobile phase B. U, unidentified decomposition product136 ANALYST, FEBRUARY 1986, VOL. 111 t m c 0 Q Q, U 0 w 4 8 12 16 20 24 I I , I I I 4 8 12 16 20 24 Tirnehnin Fig.5. Chromatograms of a test solution from capsule formulation. Chromatograms (a)-(d) as in Fig. 4 )X U I I I 1 I I 4 8 12 16 20 24 ox I V I 4 8 0 4 8 Ti rn e/rn in Fig. 6. Chromatograms of a test solution from an ointment. Chromatograms (a)-(d) as Fig. 4 pressure, 95 5 0.1 bar; chart recorder speed, 0.5 cm min-l; detector sensitivity, 0.04 or 0.02 a.u.f.s.; detection wavelength, 239 nm; mobile phase B, 0.1 M THAM citrate buffer (pH 5.25) - acetonitrile (45 + 55); conditions as above except flow-rate, 1.8 ml min-1; pressure, 105 k 0.1 bar. Results and Discussion Specimen chromatograms of a standard solution recorded at 239 nm (a suitable wavelength for the simultaneous determi- nation of all the compounds), using mobile phases A and B, are given in Fig.3(a) and (b), respectively. Table 1 gives the chromatographic data for a mixed standard solution of oxyphenbutazone and its decomposition products (10 pg ml-1) expressed as the capacity factor, k ’ , the theoretical plate height, H , and the asymmetry factor, As, under the conditions used in this study. As, was calculated using the equation As = b/a, where b , is the distance after the peak maximum and a the distance before the peak maximum, both being measured at 10% of the total peak height. The stability of a solution of oxyphenbutazone and its decomposition products was tested by separately injecting on to the chromatograph, at different time intervals, a solution of oxyphenbutazone (500 pg ml-1) and I-VI (10 pg ml-l) in methanol.After 4 h at ambient temperature under diffused light, no decomposition was observed for oxyphenbutazoneANALYST, FEBRUARY 1986, VOL. 111 137 and I-VI within the limit of sensitivity of the method (at 0.02 a.u.f.s.). The stability of test solutions of the ointment, tablet core and capsule formulation (equivalent to 500 yg ml-1) was also investigated. These solutions can be kept for 6 h without any detectable decomposition. Validation of the HPLC Procedure The linearity of the response was examined by plotting the peak-height measurement for each solute against solute concentration in the range &lo0 yg ml-1 for each compound. The calibration graph was rectilinear and passed through the origin in all instances. The correlation coefficient of the linear regression analysis was higher than 0.999 for each compound.The repeatability, assessed by five replicate analyses of a mixed standard solution (10 pg ml-1) and expressed as the coefficient of variation, is shown in Table 2. The sensitivity, defined as the change in the peak height (mm) measured at the maximum detector sensitivity resulting from a concentration change of one unit (pg), is also given in Table 2, together with the detectability, defined as the amount of compound that yields a signal to noise ratio of 2. The limit of determination can be evaluated as about three times the detectability. Commercial formulations were spiked with known amounts of I-VI (0.25% of each with respect to the theoretical oxyphenbutazone content) and these spiked formulations were treated as indicated in the method. The chromatograms obtained are shown in Figs.4-6 and average results of duplicate injections are given in Table 3. Satisfactory results were obtained except for IV in the ointment, probably because a physical and/or a chemical interaction took place. Attempts to solve this problem using different extraction solvents in the procedure (acetonitrile, the mobile phase solvent system, aqueous alkaline solutions) were unsuccessful. In addition , oxyphenbutazone was very unstable in these solvents. Inert interference corresponding to a check on the placebo effect was carried out by treating a placebo of tablets, capsule formulation and the ointment as required in the method. No interference was observed. Conclusions The proposed procedure allows the detection and determi- nation of the potential degradation products of oxyphen- butazone at trace levels. The method is more sensitive than TLC,1 as less than 0.1% of the decomposition products (with respect to oxyphenbutazone) can be quantified. In addition, the sample preparation is simple and rapid and allows the method to be used easily for routine control purposes. References 1. Fabre, H., Ramiaramanana, A., Blanchin, M. D., and Mandrou, B., Analyst, 1985, 110, 1289. 2. Fabre, H., Mandrou, B., and Eddine, H.,J. Pharm. Sci., 1982, 71, 120. Note-Reference 1 is to Part I of this series. Paper A51267 Received July 22nd, 1985 Accepted September loth, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100133

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST FEBRUARY 1986 VOL. 111 139 Prevention of Artifactual Formation of Nitrosamines During the Analysis of Baby Bottle Rubber Nipples Nrisinha P. Sen and Stephen W. Seaman Food Research Division Food Directorate Health Protection Branch Health and Welfare Canada Ottawa, Canada KIA OL2 and Santosh C. Kushwaha Scientific and Laboratory Services Division Product Safety Branch Department of Consumer and Corporate Affairs Ottawa Canada KIA OC9 It has been found that considerable amounts of nitrosamines may be formed as artifacts during the analysis of rubber nipples by a method that involves Soxhlet extraction of the samples with dichloromethane. The extent of such formation was monitored by incorporating morpholine as a marker amine and studying the formation of nitrosomorpholine which varied between 9 and 80 ng per analysis depending on the type of sample analysed and the brand of dichloromethane used.The problem could be minimised by pre-testing dichloromethane for its N-nitrosation potential and by incorporating propyl gallate an N-nitrosation inhibitor, in the method. Keywords Volatile nitrosamine determination; artifactual formation; gas - liquid chromatography - thermal energy analysis; rubber nipple analysis; propyl gallate During the past few years considerable effort has been spent by various researchers in developing sensitive and specific methods for the determination of trace levels of N-nitros-amines (mainly volatile nitrosamines) in baby bottle rubber nipples and pacifiers.l-4 The reason for this intense interest in this field stems from the fact that most of the nitrosamines detected in these products are potent carcinogens in labora-tory animals5 and that trace amounts of these chemicals can migrate to liquid infant foods or infant saliva thus posing a potential health hazard to infants using these products.For a detailed background on the subject the reader is advised to consult earlier publications on this topic. 1,276-8 Like all other trace analyses the analysis of rubber nipples for volatile nitrosamines is a complex and difficult task. The analyst has to pay attention to the usual problems such as developing efficient extraction and clean-up techniques, avoiding contamination from reagents and glassware and developing sensitive and specific detection methods.The situation is further complicated by the fact that some of the nitrosamines for which the analysis is carried out may be formed as artifacts during work-up of the samples or during final analysis (e.g. in the hot injector of the gas chromato-graph) of the extract by gas - liquid chromatography (GLC). This can happen because the rubber products themselves may contain the necessary precursors (amines and nitrosating agents) for nitrosamine formation or nitrogen oxide gases (NO,) present in the air can enter into the system and nitrosate the amines in the nipple extracts. For this reason, additional precautions have to be taken and safeguards have to be incorporated in the methods to minimise or eliminate such formation. Krull et al.9 discussed various ways of minimising artifactual formation during analysis and emphas-ised the importance of correctly determining the concentra-tion of carcinogenic nitrosamines in consumer products.The first evidence that artifactual formation of nitrosamines might pose a problem during analysis of rubber nipples was observed by Sen et a1.3 while comparing two different methods of analysis. It was noted that one of the methods that incorporated a nitrosation inhibitor and employed a simpler extraction technique (avoiding Soxhlet extraction) gave much lower values for nitrosamines in some samples than those obtained by another method2 that did not include any nitrosation inhibitor and employed Soxhlet extraction with dichloromethane. Although the latter method had been tested2 for artifactual formation by adding (after Soxhlet extraction) nitrite and morpholine to nipple extracts and analysing for possible formation of N-nitrosomorpholine it was felt that the test was inadequate because the nitrosamine precursors were not added at the start of the analysis.Therefore the possibility of artifactual formation in the latter method could not be completely ruled out; the need for further work was suggested. In this paper we present further evidence to indicate that artifactual formation of nitrosamines can indeed take place during the extraction of rubber nipples with dichloromethane. An improved method is suggested that greatly alleviates the problem. Experimental Apparatus A Model 502 thermal energy analyser (TEA) (Thermo Electron Corp.Waltham MA USA) coupled to a Varian VISTA 6000 gas chromatograph was used for the determina-tion of volatile nitrosamines. The gas chromatographic col-umns and operating conditions were similar to those reported previously.7 The TEA was operated in the GLC mode with a stainless-steel cold trap immersed in liquid nitrogen and a furnace temperature of 475 "C. The vacuum chamber pressure was ca. 1.8 Torr. Reagents and Standards Glass-distilled dichloromethane (DCM) was purchased from three suppliers. Of these two were sold as nitrosamine-free reagents and for convenience these will be referred to as brands A and C. The third (brand B) was of the regular glass-distilled variety and contained cyclohexene as a preser-vative. It is understood that the DCM of brands A and C did not contain any known preservative but they may have been processed by special techniques.As commercial DCM had previously been shown10 to be contaminated with N-nitroso-morpholine each bottle of DCM was redistilled from an all-glass apparatus and tested for nitrosamine contamination before use. All other reagents used were of analytical-reagent grade. To ensure the absence of nitrosamine contamination a reagent blank was used with each batch of new reagents and analysed by GLC - TEA. Highly activated basic alumina (IC 140 ANALYST FEBRUARY 1986 VOL. 111 Nutritional Biochemicals Cleveland OH USA) was pre-pared by heating overnight at 500 "C cooling in a desiccator and then storing in a stoppered flask. It was used without any addition of water for the purification of certain batches of DCM as will be discussed later.Ascorbyl palmitate and propyl gallate were obtained from ICN K & K Laboratories, Plainview NY USA and Eastman Chemical Products, Kingsport TN USA respectively. Morpholine was obtained from BDH Chemicals Poole UK. Dilute standards (ca. 10 pg ml-1 of each in ethanol) of a mixture of seven volatile nitrosamines and a separate standard of N-nitrosodipropylamine (100 pg ml-1) were purchased from Thermo Electron Corp. The mixture consisted of the following N-nitrosodimethylamine (NDMA) N-nitroso-diethylamine (NDEA) N-nitrosodipropylamine (NDPA) , N-nitrosodibutylamine (NDBA) N-nitrosopiperidine (NPIP) N-nitrosopyrrolidine (NPYR) and N-nitrosomor-pholine (NMOR) .The solutions were appropriately diluted with DCM before use. NoteReference to a brand or company name does not constitute endorsement by the Consumer and Corporate Affairs of Canada or by Health and Welfare Canada over others of a similar nature not mentioned. Samples All of the samples except one were purchased locally during October - December 1984 or procured by the Inspector of the Department of the Consumer Corporate Affairs. These are identified by capital letters (D H F etc.). Only a few samples of brand H from a recent (February 1985) production lot were obtained directly from the manufacturer. Those identified by lower-case letters (a e etc.) were left-over samples from an AOAC collaborative study in which we participated during June - July 1984.Caution-As all of the nitrosamine standards mentioned above are potent carcinogens extreme precaution should be taken when working with or handling the chemicals. Contact with skin or inhalation of vapour must be avoided. All solutions containing nitrosamines should be destroyed by a well established procedure11 before disposal. Analysis of Rubber Nipples Method 1 The method was essentially the same as that described by Havery and Fazio2 but later modified to include the addition of 2 g of barium hydroxide to prevent foaming during distillation. Basically it consisted of (a) overnight soaking of the cut nipple pieces with DCM followed by Soxhlet extrac-tion (b) distillation of the DCM extract from an aqueous alkaline solution (c) re-extraction of the aqueous distillate (containing the volatile nitrosamines) with DCM (d) concen-tration of the DCM extract to a small volume (ca.1 ml) using a Kuderna - Danish concentrator and a micro-Snyder column (for the final concentration step to 1 ml) and (e) GLC - TEA analysis of the final extract. In this study 1.0 ml of NDPA (10 ng ml-1) was added as an internal standard to each sample at the beginning of the analysis. This was carried out only to check the performance of the method; the final results were not corrected for percentage recoveries of NDPA. Also in some instances, 1 mg of morpholine (in 1 ml of DCM brand B) was added at the start to the sample as a marker amine to monitor the extent of artifactual formation of nitrosamines. Method 2 The method has been described in detail elsewhere.3 In summary it consisted in overnight extraction of the sample with DCM in the presence of 100 mg of ascorbyl palmitate as an N-nitrosation inhibitor rinsing of the nipple pieces using a special extraction technique that did not involve Soxhlet extraction concentration of the extract to 1 ml as in method 1 and final analysis by GLC - TEA.Testing the Effect of Different Brands of DCM on Nitrosamine Levels in Rubber Nipples as Determined by Method 1 As nitrosamine levels may vary from nipple to nipple such tests were always carried out in pairs using half of a nipple or using aliquots from a composite mixture of cut nipple pieces. Also everything else such as size of the distillation and Soxhlet flasks size of the Soxhlet extractor and other reagents was kept constant.The size of the apparatus was kept as small as possible so as to avoid an excessive reduction in volume of DCM from the Soxhlet extraction flask (at the bottom) that, otherwise could have caused overheating. The paired experi-ments were carried out simultaneously in the same fume hood. This ensured identical atmospheric conditions such as NO, levels that may have an influence on the formation of nitrosamines during Soxhlet extraction. A 1-ml volume each of NDPA internal standard and morpholine marker amine were also routinely added at the start to check the perfor-mance of the method and to monitor the artifactual formation of NMOR respectively. Prior to this the samples were analysed without any addition of morpholine to ensure the absence of pre-formed NMOR in them.Morpholine blanks (1 mg) were also run in the same manner with different lots and brands of DCM the only difference being that no nipples were present in these tests. The Soxhlet extracts in these instances were concentrated directly (omit-ting the aqueous distillation) and analysed by GLC - TEA. Testing the Effect of Propyl Gallate (PG) While carrying out this test all the factors including the DCM solvent were kept constant. Half of a nipple was analysed by the above method2 and the other half by the same method but with added PG (100 mg). The stoppered flask containing the nipple pieces PG and DCM was gently shaken (to aid in dissolving PG) overnight in the dark while the other (without PG) was allowed to sit in the dark without shaking as specified in the original method.2 The other steps were unchanged.Results and Discussion The data (Table 1) indicate that the levels of nitrosamines detected in a sample of rubber nipple by method 1 can vary widely depending on the type of DCM used for the analysis. The variation in results due to the use of different DCM could not all have been due to random errors because duplicate results (using brand B DCM) run under such controlled conditions usually were within 10%. Therefore the difference (considered to be significant if 3 _+30%) in the two sets of results in Table 1 was probably caused by artifactual forma-tion the extent of which varied from brand to brand of DCM and to a smaller extent from bottle to bottle of DCM within the same brand.This theory was also supported by the fact that there was a concomitant rise in the artifactual formation of NMOR from added morpholine for a particular DCM used for the analysis. As neither the morpholine nor the nipples (when analysed alone) contained any detectable amount of NMOR the NMOR detected in these experiments must have formed as an artifact. Similar conclusions could also be drawn from the results for the morpholine blanks (Table 2). Some typical chromatograms obtained from these experiments are shown in Fig. 1. Although the exact nature of the nitrosating agent(s) responsible for the artifactual formation of nitrosamines in the above method is not known it is well established that DCM is an excellent medium for N-nitrosation.12313 Both inorgani ANALYST FEBRUARY 1986 VOL. 111 141 Table 1. Effect of different brands of DCM on the levels of nitrosamines in various nipples and that of artifactually formed NMOR in the presence of added morpholine Brand of nipple e . . . . e . . . . b . . . . b . . . . e . . . . e . . . . H . . . . H . . . . H . . . . H . . . . H . . . . H . . . . H . . . . Blanks . . * . . . * . . . . . . . . . . . . . . . . . . . . . . . Brand of DCMused for analysis B C B A B A B A A passed through basic A purified by sulphamic alumina acid wash and alkali wash, then dried over anhydrous sodium sulphate B C C passed through basic A B or C (without added alumina morpholine) Nitrosamines detected in nipples/ I % kg-l NDEA,5.5 NDEA 11.7 Negative NDMA,55.3;NDEA,4.3; NDBA,4.8;NPYR,6.7; NMOR,3.5 NDEA,3.9;NMOR,2.3 NDEA 10.0; NMOR 2.2 NDMA 9 NDMA 83 NDMA 38.9 NDMA,8.2 NDMA,8.3 NDMA 19 NDMA,7.9 All negative * When analysis was carried out with 1 mg of added morpholine as a marker amine.t Not included. 10.9 74.3 41.8 11.5 11.8 23.5 11.2 NMOR Recovery of formed as NDPA internal an artifact*/ng standard % 9.2 -$ 67.8 -- 97.5 100 75 80 95.2 96.3 103 80.2 83.3 93.8 75.7 Table 2. Artifactual formation of NMOR from added morpholine in blank runs by method 1 NMOR formed from 1 mg of added morpholine/ng DCM A 30.6 47 > 60 11.5 (after special processing) * * See text.DCM B 16.9 22.9 41.8 14.5 14.5 12.4 DCM C 13.7 15.7 19.6 80.3 38.6 31.3 24.1 14.2 (after special processing) * 59.6 nitrite and NO gases can efficiently nitrosate secondary amines in DCM.12J3 The nitrosating agents could have originated in several ways. Firstly they could have been present in the nipples1.14 or formed as a result of thermal degradation during Soxhlet extraction. Transnitrosation by organic nitro or nitroso compounds present in the nipples might have been partly responsible for the artifactual formation of nitrosamines observed in this study. This was demonstrated by analysing (by method 1) a 5-g aliquot of cut nipple pieces (brand H) with 1 mg of added N-nitrosodiphenylamine a well known rubber curing ingre-dient,l4 and showing increased formation of NDMA during the analysis (10000 pg kg-1 compared with 13 pg kg-1 in the absence of added N-nitrosodiphenylamine) .This suggested that there was enough amine precursors present in the sample that reacted with N-nitrosodiphenylamine which is an excel-lent transnitrosating agent to produce the excess of NDMA. From the excessively high result obtained in the above experiment one would expect the formation of substantial amounts of NDMA even in the presence of much smaller t fn C 0 a 2? L c 0 al U c z I NDMA - 4 8 ’ NDMA 83 p.p.b. 96.3% ecovery II I I 4 8 1 Ti me/m in 98.4% recovery 0 I 0 4 8 1 2 Fig. 1. Typical chromatograms showing the effect of special process-ing of DCM on the results obtained by method 1.1 Nitrosamine standards; 2 a nipple (brand H) analysed using DCM of brand A (before special processing or clean-up); a total of 74.3 ng of NMOR was formed from the added morpholine marker amine; 3 direct injection of morpholine solution showing the absence of NMOR; 4, DCM of brand A (before special processing) reagent blank taken through all the steps of method 1 (without morpholine marker amine); and 5 above nipple analysed using specially processed (sulphamic acid and alkaline washes etc.) DCM from same bottle as above; only 11.5 ng of NMOR were formed. Percentage recoveries refer to those with added NDPA internal standard. For details see text and Table 142 ANALYST FEBRUARY 1986 VOL. 111 amounts (e.g.W100 pg) of N-nitrosodiphenylamine. Alter-natively NO and O2 in the air could have entered the system during Soxhlet extraction. According to Mirvish,13 nitrosation of certain amides by NO in DCM is about 30000 times faster than that by nitrous acid in aqueous solution. Therefore even trace amounts of NO gases in the air could form significant amounts of nitrosamines because the reaction is almost quantitative. The fact that NMOR could form in the absence of rubber nipples (e.g. morpholine blanks) led us to speculate that NO, gases in the air might be partly responsible for such artifactual formation. As the extent of such formation varied with different brands of DCM (run side by side) the existence of some other factors such as catalysts (e.g.trace amounts of HCl) or of other nitrosating agents in the DCM was a possibility. Other possible mechanisms include (a) participa-tion of certain transition metals (e.g. zinc dithiocarbamates are used as additives in rubber manufacture) as catalysts of N-nitrosation14.15 and (b) formation of amine - NO Drago complexes followed by oxidation by O2 (in the air) to the nitrosamines. 16 As reagent blank determinations (Fig. 1) carried out according to the protocol of method 1 cannot distinguish a good from a bad DCM (both give negative blanks) we have developed a procedure that allows one to pre-test the DCM. This is done by carrying out a morpholine blank determination as described under Experimental (Table 2). The greater the extent of NMOR formation in such a test the greater will be the chance of artifactual formation of nitrosamines during analysis of rubber nipples using the particular DCM.The best commercially available glass-distilled DCM tested gave a value of ca. 10 ng of NMOR in such tests. Therefore it is recommended that any DCM that gives a significantly higher value for NMOR formation (say >15 ng) should not be used for such analyses otherwise the results in extreme instances (e.g. sample H Table 1) could be inflated by as much as 950%. To our knowledge the possibility of this happening has not been investigated or reported previously. It should be emphasised that neither of the DCMs of brand A or C was of sub-standard quality. Both were of glass-distilled varieties and both were redistilled and tested to be nitrosamine free before use.Also both gave negative reagent blanks (Fig. 1D) when taken through all the steps of method 1. Therefore without a knowledge of the data presented here, an analyst would have no justification for rejecting them for use in the analysis of rubber nipples for nitrosamines. As all the DCMs were redistilled before use the responsible nitrosating agent or catalyst in the DCM must be volatile and should not be removed by distillation. Therefore alternative methods were developed to purify DCM. A 1-1 volume of brand C DCM (initially forming 25-30 ng of NMOR in the morpholine test) was purified by passage through a column containing 50 g of highly activated (activity I) basic alumina. The purified DCM on re-testing formed only 14 ng of NMOR.Similar treatment through alumina however was only partially effective for a sample of brand A DCM that initially formed >60 ng of NMOR in the morpholine test (Table 2). This DCM was further purified by shaking vigorously for 5 min in a separating funnel with 1% sulphamic acid (prepared in 0.5 M H2S04) back-washing with 1 M KOH solution and drying over anhydrous sodium sulphate . The treatment greatly improved the DCM which on re-testing gave a very low value of 11.5 ng of NMOR in the morpholine test (Table 2). These two specially purified DCMs were further tested for the analysis of rubber nipples and the results were compared with those obtained with unpurified DCM from the respective bottles (Table 1). The results (Table 1 nipple H DCM of brands A and C) clearly indicate a noticeable reduction in the artifactual formation of NMOR and also lower results for the nipples compared with those obtained with the corresponding unpurified DCM.The respective new (with purified DCM) results were also comparable to those obtained with the brand B DCM the best commercially available. The fact that it was possible to purify two poor lots of DCM and obtain results comparable to those obtained with a third brand (B) of good DCM lends further support to the conclusion that the results obtained by method 1 are subject to extreme variations depending on the quality of DCM used for the analysis. Therefore the importance of pre-testing DCM before starting the analysis cannot be overemphasised.Next the possibility of similar artifactual formation of nitrosamines that could be occurring even with the best DCM, i.e. brand B was investigated. In previous studies with cured meats,I7 fried bacon,17 beer18 and rubber nipples,3 research-ers have recommended incorporation of N-nitrosation inhibi-tors in the analytical protocols in order to minimise such formation. After some preliminary trials with various inhibi-tors (ascorbic acid ascorbyl palmitate a-tocopherol pyrrole and propyl gallate) propyl gallate (PG) was selected for this purpose because it gave the most consistent results. The results of several analyses in which duplicate halves of rubber nipples were analysed with or without PG are presented in Table 3. As can be seen from the data the inclusion of 100 mg of PG in method 1 gave much lower values for nitrosamines in most instances suggesting artifactual formation of nitros-amines in its absence.In a few instances this was also confirmed by adding morpholine as a marker amine (Fig. 2). PG also reduced the formation of NMOR. It should be emphasised that this phenomenon was not observed with all nipples tested (e.g. samples D F H6 and i). This was particularly noticeable with the most recent nipple of brand H (H6). Probably the nitrosamines in these instances were already present in the nipples or the necessary precursors were absent (as a result of the introduction of improved rubber curing practices). The extent of inhibition of nitrosamine formation in the presence of PG varied depending on the brand of DCM used and also with the type of nipple.Even with the best commercially available DCM (brand B) one could obtain a result that was 375% higher (Table 3 sample H4) if PG was omitted. In an extreme instance (e.g. sample H5 with brand A DCM) the difference was 700%. The addition of PG did not in any way affect the performance of the method nor did it affect the recoveries of added nitrosamines. In the presence of PG, the recoveries of all seven volatile nitrosamines (see Reagents and Standards) added to rubber nipples at ca. 20 pg kg-1 levels and also that of NDPA internal standard added to all samples were excellent (8&100./,). This ruled out the possibility of any loss of nitrosamines due to breakdown or interaction with PG. A few analyses were carried out (using brand H nipple) in which PG or N-nitrosodiphenylamine was added at various stages of method 1.The results suggest that in the absence of PG artifactual formation of nitrosamines can take place during both DCM extraction and alkaline distillation. As information regarding the detailed composition of various nipples was not easily available it was difficult to investigate the problem more thoroughly. Further work might be desirable to understand fully the mechanism of artifactual formation observed in this study. In a few limited instances the results obtained by the improved method 1 (with PG and pre-testing DCM) were compared with those obtained by method 2 (Table 4). They were in excellent agreement. In a previous study3 with rubber nipples it had been observed that method 1 (without PG) could give results higher than or comparable to (never lower than) those obtained by method 2 which included an N-nitrosation inhibitor.The findings presented in this paper have been helpful in explaining the reasons behind these discrepancies. The occasional higher results previously ob-served with method 1 were probably due to artifactual formation. This was also substantiated by the data in Fig. 3 ANALYST FEBRUARY 1986 VOL. 111 143 Table 3. Inhibition of artifactual formation of nitrosamines by PG during analysis of rubber nipples for volatile nitrosamines Brand of nipple * HI . . . . Hz . . . . H3 . . . . H4 . . . . H5 . . . . H6$ . . . . a . . . . . . a . . . . . . a . . . .. . F . . . . . . g . . . . * . i . . . . . . e . . . . . . d . . . . . . D . . . . . . Volatile nitrosamines detected DCM brand used for analysis Nitrosamine CLgkg-lt PG/pg kg- 1 t Amount by method 1/ Amount by method 1 with . . C NDMA 60 (81.3%) 13.6 (93.6%) . . c NDMA 78.1 (90.3%) 15.4 (96.9%) . . B NDMA 9.8 (86.5%) 4.0 (98.0%) . . B NDMA 13.5 (76.7%) 3.6 (76.7%) 107 (85%) 14.0 (93.8%) . . A NDMA . . B NDMA 2.0 (goo/,) 2.0 (85%) NDEA 1.5 1.5 NMOR 4.0 4.0 . . c NDMA 79.1 (85.4%) 10.3 (94.4%) NDMA 14.5 (92.2%) 5.9 (87.5%) . . B . . B NDMA 8.5 (100%) 3.6 (100%) . . B NDBA 181 (102.7%) 167 (89.3%) NPIP 39.4 36.3 . . B - Negative (86.7%) Negative (92%) . . B NDBA 81.9 99.4 . . B NDEA . . B NDEA 15.0 (81%) 5.6 (77%) 3.2 (75.6%) 2.6 (85%) .. B NDBA 87.7 (94%) 79.3 (94.5%) NPYR 13.7 12.3 NMOR 10.9 12.3 * Different subscripts indicate samples from different lots or different nipples from the same lot. t Figures in parentheses represent recoveries of NDPA internal standard. $ Obtained directly from the manufacturer in February 1985. Table 4. Comparison of results obtained by improved* method 1 with those obtained by method 2 NDMA detected/pg kg-17 Brand of Improved nipple Method 1 method 1 Method 2 a 8.5 (100%) 3.6 (100%) 3.3 (97%) a 14.5 (92.2%) 5.9 (87.5%) H 18.9 (89.2%) 5.1 (82.5%) 5.9 (97.5%) H$ 71.4 (82.1%) - 9.7 (98.5%) * Using brand B DCM and including 100 mg of PG. t Figures in parentheses represent percentage recoveries of NDPA $ Using brand C DCM. internal standard.( a ) NDMA 14. e 1.p.b. 92.2% ecovery NDMA 5.9 p.p.b. 4 8 12 0 4 8 Timelmin Fig. 2. GLC - TEA chromatograms (a) a nipple of brand a analysed by method 1 using DCM of brand B; (b) the same nipple analysed in the presence of PG. About 24.3 and 12.6 ng of NMOR were formed, respectively from the morpholine added in each instance (a) Method 2 I I 9.7 p.p.b. recovery u-0 4 8 12 1 6 0 4 8 12 Ti me/m in Fig. 3. GLC - TEA chromatograms of a ni ple (brand H see Table 4) analysed by (a) method 2 and (b) metgod 1 (with mor holine marker amine added in each instance) using DCM of brand 8. Both the NDMA levels detected in the nip le and the amounts of NMOR formed were higher with method 1. d e GLC column and conditions were slightly different from those in Figs.1 and 2 hence the slight difference in retention times which show increased formation of NMOR in method 1 compared with method 2. It has been previously shown3J8 that the use of a micro-Snyder column (instead of blowing down in a stream of nitrogen) for the final concentration from 4 to 1 ml and the use of a Graham condenser which is more efficient than a straight-jacket condenser during the distillation step can give consistently good recoveries of the volatile nitrosamines. It is therefore recommended that both of these modifications also be incorporated in the improved version of method 1. Finally it should be emphasised that this work should not be viewed as an undue criticism of method 1. The main objective was to determine the cause of variations in results produced by method 1 and to improve it.Basically it is a good method that needed refinement. With the modifications suggested above and those mentioned earlier (i. e. pre-testing DCM and including PG) the method would improve greatly 144 ANALYST FEBRUARY 1986 VOL. 111 and should give more precise and accurate results. Also the effect of PG on nitrosamine levels detected by method 1 should be tested for any new product not analysed before. This will offer a safeguard against any possible catalytic influence it might have on nitrosamine formation in the presence of new rubber curing agents. 1. 2. 3. 4. 5. 6. 7. 8. References Spiegelhalder B. and Preussmann R. ZARCSci. Publ. 1982, No. 41 231. Havery D.C. and Fazio T. Food Chem. Toxicol. 1982,20, 939. Sen N. P. Seaman S. W. Clarkson S. G. Garrod F. and Lalonde P. IARC Sci. Publ. 1985 No. 57 51. Billedeau S. M. Thompson H. C. Jr. Miller B. J. and Wind M. L. 98th Annual AOAC Meeting October 29-November 2 1984 Washington DC Abstract No. 252. Preussmann R. and Stewart B. W. in Searle C. E. Editor, “Chemical Carcinogens,” Second Edition ACS Monograph No. 182 American Chemical Society Washington DC 1984, p. 643. Babish J. G. Hotchkiss J. H. Wachs T. Vecchio A. J., Guttenmann W. H. and Lisk D. J. J . Toxicol. Environ. Health 1983 11 167. Sen N. P. Kushwaha S. C. Seaman S. W. and Clarkson, S. G. J. Agric. Food Chem. 1985,33,428. Osterdahl B.-G. Food Chem. Toxicol. 1983 21 755. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Krull I. S. Fan T. Y. and Fine D. H. Anal. Chem. 1978, 50 698. Eisenbrand G. Spiegelhalder B. Janzowski C. Kann J., and Preussmann R. ZARCSci. Publ. 1978 No. 19 311. Castegnaro M. Eisenbrand. G Ellen G Keefer L. Klein, D. Sansone E. B. Spincer D. Telling G. and Webb K. S . , in “Laboratory Decontamination and Destruction of Nitros-amines in Laboratory Wastes,” ZARC Sci. Publ. 1982 No. 43. Angeles R. M. Keefer L. K. Roller P. P. and Uhm S. J., ZARCSci. Publ. 1978 No. 19 109. Mirvish S . in Magee P N. Editor “Nitrosamines and Human Cancer,” Banbury Report No. 12 Cold Spring Harbor Laboratory New York 1982 p. 227. International Agency for Research on Cancer “IARC Mono-graphs on the Evaluation of the Carcinogenic Risk of Chem-icals to Humans The Rubber Industry,” IARC Lyon 1982, Chapter IV p. 89. Croisy A. F. Fanning J. C. Keefer L. K. Slavin B. W. and Uhm S.-J. IARC Sci. Publ. 1980 No. 31 83. Hansen T. J. Croisy A. F. and Keefer L. K. ZARC Sci. Publ. 1982 No. 41 21. Hotchkiss J. H. Libbey L. M. Barbour J. F. and Scanlan, R. A. IARCSci. Publ. 1980 No. 31 361. Sen N. P. Seaman S . and Bickis M. J. Assoc. Off. Anal. Chem. 1982 65 720. Paper A51236 Received July lst 1985 Accepted September 16th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100139

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 145 Gas Chromatographic Determination of Triclopyr in Environmental Waters Tadashi Tsukioka Nagano Research Institute for Health and Pollution, 1978, Komemura, Amori, Nagano-shi, Nagano, Japan Ryuzo Takeshita School of Pharmaceutical Science, Toho University, 22 I , Miyama, Funabashi-shi, Chiba, Japan and Tetsuro Murakami Department of Chemical Engineering, Kogakuin University, 1-24-2, Nish ish inju ku, Shinjuku-ku, Tokyo, Japan The reaction of BF3-trifluoroethanol with an extract of triclopyr from an acidified sample solution to form the trifluoroethyl ester has been applied to the determination of triclopyr in environmental waters. The product is cleaned up by silica-gel column chromatography and determined by gas chromatography with electron- capture detection.The detection and determination limits were 0.005 ng and 0.00025 pg ml-1, respectively. The recovery and coefficient of variation were found to be 90-93% and less than 4%, respectively ( n = 71, for recovery experiments on river waters. Keywords: Triclopyr determination; gas chromatography; herbicide residue; river water; halogenated alk yla tion Triclopyr (3,5,6-trichloro-2-pyridyloxyacetic acid) is a hor- mone-type herbicide that is effective for the destruction of arrowroots (Pueraria thunbergianal) and deciduous shrubs. It is on the market as the triethylammonium salt (triclopyr- TEA) or butoxyethyl ester (triclopyr-BE). This herbicide, singly or mixed with Frenock (sodium 2,2,3,3- tetrafluoropropionate) , is used extensively in woods and forests.2 It is therefore necessary to determine the pollution of natural waters by this herbicide, for which purpose a simple and highly precise microanalytical method is required.A microanalytical method for the determination of triclopyr has not been reported, although there are many reports of the determination of phenoxy herbicides,sl6 which have similar properties to triclopyr. Phenoxy herbicides have a high polarity and low volatility, preventing the use of a direct GC method for their determination. Thus, they are subjected to GC or GC - MS after conversion into more volatile com- pounds, i.e. , alkyl esters,sg halogenated alkyl esters10Jl or halogenated aromatic esters. 12-16 In the environment triclopyr- BE is hydrolysed gradually into triclopyr-BE is more easily total amount of triclopyr and triclopyr-BE is more easily determined than are the two compounds separately.A method of determination has been developed in which triclopyr and triclopyr-BE in acidic, aqueous solution are extracted with diethyl ether, converted into the trifluoroethyl (TFE) ester, cleaned up by silica-gel column chromatography and determined by gas chromatography with electron-capture detection. (ECD - GC). This method has sufficient sensitivity, precision and manageability to be applicable to environmental waters. Experimental Reagents 3,5,6-Trichloro-2-pyridyloxyacetic acid (triclopyr) , triethyl- ammonium 3,5,6-trichloro-2-pyridyloxyacetate (triclopyr- TEA) and butoxyethyl 3,5,6-trichloro-2-pyridyloxyacetate (triclopyr-BE) were obtained from Dow Chemical Japan.Standard solutions of each of these compounds were prepared from a concentrated solution of 1000 pg ml-1 in acetone by diluting with acetone to 10, 1 or 0.1 yg ml-1. Polychlorinated biphenyls (PCB) , dibutyl phthalate, 2,4-dichloro- phenoxyacetic acid (2,4-D) , 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) and 2-methyl-4-chlorophenoxyacetic acid (MCP) were obtained from Wako Pure Chemical Co., Japan. Silica gel, Wako-gel S-1 from Wako Pure Chemical Co. , Japan, was activated by heating for 12 h at 130 "C before use. A solution of 15% mlV boron trifluoride in 2,2,2-trifluoroethanol (BF3- TFE) was obtained from Tokyo Kasei Co., Japan. Hexane, benzene, acetone and diethyl ether were of the grade suitable for detection of pesticide residues.All the other reagents were of guaranteed grade. Apparatus The gas chromatograph was a Shimadzu Model GC-3BE equipped with an electron-capture detector (63Ni) , the Reacti- Therm was from Pierce Chemical Co., USA, the gas chromatograph - mass spectrometer was a Model JMS-D300 from Japan Electron Optics Laboratory Co., Japan, and the chromatographic column was 10 mm in diameter and 300 mm in length. GC Conditions The stationary phase was 5% XE-60 on Chromosorb W (60-80 mesh) , packed into a glass column (3 mm in diameter and 200 cm in length). The carrier gas was nitrogen with a flow-rate of 28 ml min-1. The temperature was 155 "C for both the column and the detector and 200 "C for the injection port.146 > .- c ANALYST, FEBRUARY 1986, VOL. 111 I I t a c 0 a a a L I 1 5 70 Tim e/m i n Fig.1. Gas chromatogram of triclopyr-TFE. Column, 2 m, 5% XE-60; column and detector temperature, 155 "C; injection-port temperature, 200 "C; and carrier gas, N2 at a flow-rate of 28 ml min-* looo t 210 11 - 1000 c t M' 337 0 250 300 350 400 m/z Fig. 2. EI mass spectrum of triclopyr-WE. Column, 2 m, 5% XE-60; column t e y erature, 160 "C; injection-port and enricher temperatures, 200 8; ion-source temperature, 250 "C; ionisation voltage, 70 eV; and carrier gas, He at a flow-rate of 40 ml min-1 Table 1. Effects of reaction temperature and reaction time on the esterification of triclopyr Esterification at reaction temperature, % Reaction time/ min 50°C 60°C 70°C 80°C 90°C ' 10 44 63 90 100 100 20 70 82 100 100 100 40 94 100 100 100 100 60 100 100 100 100 100 80 100 100 100 100 100 Standard Procedure The standard procedure consists of four steps: extraction, esterification, clean-up and determination.Table 2. Effects of reaction temperature and reaction time on the ester-group exchange reaction Exchange at reaction temperature, "/o Reaction time/ min 50°C 60°C 70°C 80°C 90°C 10 22 46 61 76 93 20 51 68 87 95 99 100 100 40 73 94 99 100 100 60 85 100 100 100 100 80 91 100 100 q , , g 75 .- 0.5 1 .o .+I - a Volume of reagent/ml LK Fig. 3. Effect of amount of reagent on the esterification 8 2 100 2{ s,' a, 75 Fig. 4. Effect of amount of reagent on the exchange reaction of ester groups Extraction A 200-ml portion of the sample water is placed in a 300-ml separating funnel, to which 6 g of NaCl and 1 ml of 9 M H2SO4 are added and shaken to 10 min with each of two 50-ml portions of diethyl ether.The combined diethyl ether extracts are washed with 20 ml of 10% mlV NaCl solution, dried with anhydrous Na2S04 and concentrated to less than 5 ml in a Kuderna - Danish (KD) concentrator. Esterification The concentrate is transferred into a 5-ml vial and the solvent is removed with a gentle stream of Nz. A 0.25-ml portion of BF3-TFE is added to this vial, which is covered by a Teflon cap and placed on a Reacti-Therm at 80 "C and allowed to react for 1 h. After cooling, the contents are transferred into a 100-ml separating funnel with 30 ml of hexane, washed twice with 20 ml of 10% mlV NaCl solution, dried with anhydrous Na2S04 and concentrated to less than 5 ml in the KD concentrator.Clean-up The concentrate is transferred on to a 3-g silica-gel column (10 mm i.d.) that has been slurry-packed in hexane. The column is washed with 100 ml of benzene - hexane (10 + 90), and the adsorbed compound is eluted with 100 ml of benzene - hexane (35 + 65). Determination The eluate is concentrated in the KD concentrator to less than 5 ml, diluted to the appropriate volume and subjected to ECD - GC.ANALYST, FEBRUARY 1986, VOL. 111 50 23 d c 3 -Z 25 r 3 E" a 147 - - H2S04 concentrationh Fig. 5. Effect of acid concentration on the recovery I I 1 I I I 0 50 Volume of solvent/mI t a, t 0 Q a n Fig. 6. Elution pattern of triclopyr-TFE 0 10 Timeim in 20 Fig. 7. Gas chromatogram of river water before clean-up. Condi- tions as in Fig.1 A blank test is conducted by the same procedure with 200 ml of distilled water. Results and Discussion Formation and Identification of Triclopyr-TFE The followingexperiment was conducted in order to prevent tailing during gas chromatography and to produce ECD - GC analysis of high sensitivity. A 1-mg mass of triclopyr and 0.5 ml of BF3-TFE were placed in a vial, heated at 80 "C for 1 h, and extracted with hexane. To select the column conditions for the GC of the reaction product, a test was conducted in the range 150-200 "C on 2% OV-17, 2% OV-101, 5% SE-52, 5% DEGS and 5% XE-60. Fig. 1 shows the chromatogram obtained with XE-60, which was the best with respect to the peak shape and separation from coexisting substances. A mass spectral measurement gave a peak at mlz = 337 corresponding to the molecular ion (M+) of triclopyr-TFE (Fig.2). Investigation of the Conditions for TFE Esterification Esterification To ascertain the optimum reaction temperature and time for the esterification, 2.5 ml (containing 2.5 pg) of triclopyr standard solution were placed in a 5-ml vial and the solvent was removed in a gentle stream of N2. A 0.25-ml portion of BF3-TFE was added to this vial, which was covered by a Teflon cap. Esterification was conducted at 50-90 "C for 10-80 min. As Table 1 shows, the ester was obtained quantitatively under the following reaction conditions: 60 min at 50 "C, 40 rnin at 60 "C, 20 min at 70 "C and 10 rnin at 80 or 90 "C. The amount of triclopyr, the temperature and the reaction time were kept constant at 20 pg, 60 "C and 60 min, respectively, and the amount of the reagent was changed in the range 0.05-1 ml.Fig. 3 shows that the esterification is constant with a minimum amount of 0.1 ml of BF3-TFE, but 0.25 ml of BF3-TFE was used for actual samples in the event that they contain other substances that consume BF3-TFE. Ester-group exchange reaction The effects of reaction temperature, reaction time and exchange ratio on the ester-group exchange reaction were investigated with 2.5 ml (containing 2.5 pg) of triclopyr-BE standard solution in a vial, under the same conditions as the esterification. Table 2 shows that the ester-group exchange reaction does not occur as easily as the esterification; even an 80-min reaction time at 50 "C did not yield a 100% exchange.Quantitative exchange requires 60 rnin at 60 "C, ca. 40 rnin at 70 "C, 40 min at 80 "C and 20 min at 90 "C. The amount of reagent required for a quantitative reaction was found by using 20 pg of triclopyr-BE under the same conditions as the esterification. Although 0.1 ml of reagent was sufficient, 0.25 ml was used for the same reason as in the esterification (Fig. 4). Hence the optimum conditions for esterification are as follows: addition of 0.25 ml of BF3-TFE and reaction for 1 h at 80 "C. Investigation of the Extraction Conditions Extraction of triclopyr Diethyl ether was selected as the extraction solvent because it is easily removed after extraction. The optimum acid concentration for extraction was deter- mined in the following manner. A 2.5-ml volume of triclopyr- TEA standard solution (containing 2.5 pg) and different amounts of 9 M H2S04 were added to 100 ml of distilled water to produce the final solution with H2S04 concentrations of 0-0.5 M.Each was extracted with 50 ml of diethyl ether and then submitted to the standard analysis procedure. The relationship between the acid concentration and the recovery148 ANALYST, FEBRUARY 1986, VOL. 111 L 0 5 10 Timeimin Fig. 8. Gas chromatogram of river water after clean-up. Conditions as in Fig. 1 0 4 1 I 10 Time/m in 20 Fig. 9. Gas chromatogram of TFE esters of acid herbicides. Conditions as in Fig. 1. Peaks: 1, MCP, 2 ng; 2, triclopyr, 0.2 ng; 3, W D , 0.4 ng; 4, 2,4,5-T, 0.3 ng t a rn C 0 P a a 0 5 10 Time/min Fig. 10. Gas chromatogram of a river water. Conditions as in Fig.1 was found and the results are shown in Fig. 5. The recovery of triclopyr-TEA was ca. 55% with no acid added and ca. 100% for an acid concentration above 0.025 M. To determine the optimum salt concentration for ex- traction, the relationship between the salt concentration and the extraction ratio was investigated by changing the NaCl concentration in the range &30% mlV, keeping the acid concentration at 0.05 M. The recovery did not change. As environmental samples sometimes contain suspended sub- stances that form emulsions which prolong the time required for separation, the addition of up to 3% NaCl solution was adopted to prevent such an emulsion formation. Extraction of triclopyr-BE Triclopyr-BE can be extracted by ordinary hydrophobic solvents and hence was extracted with diethyl ether, as was triclopyr-TEA.The dependence on the recovery with acid and salt concentrations was similarly investigated and was found to be independent of both acid and salt concentration over the range examined. Hence 6 g of NaCl and 1 ml of 9 M H2S04 were added to 200 ml of sample before extraction. Clean-up by Silica-gel Column Chromatography A variety of substances in natural waters can be simul- taneously extracted and esterified by TFE and therefore can cause interference in the determination of triclopyr-TFE. Hence the esterification products were purified by silica-gel column chromatography. Triclopyr-TFE (10 pg) was loaded on to a column (10 mm i.d.) containing 3 g of silica gel and eluted with 100 ml of benzene - hexane (10 + 90); no ester species were found in the eluate.The ester was completely eluted with a mixture (100 ml) of benzene - hexane (35 + 65) (Fig. 6).ANALYST, FEBRUARY 1986, VOL. 111 149 Table 3. Results of the addition and recovery experiments for triclopyr-TEA and triclopyr-BE Coefficient of Amount variation Sample Compound added/yg Recovery, YO (n = 7), Yo Distilled water . . . . Triclopyr-TEA 0.5 92 3.1 Triclop yr-BE 0.5 93 2.7 Riverwater . . . . . . Triclopyr-TEA 0.5 90 3.5 Triclopyr-BE 0.5 92 3.4 The clean-up procedure was as follows: the column was washed with 100 ml of benzene - hexane (10 + 90) and the ester was eluted with 100 ml of benzene - hexane (35 + 65). The usefulness of this procedure was demonstrated with 0.5 ml (containing 0.5 pg) of triclopyr standard solution in 200 ml of an actual river water (Figs.7 and 8). Effect of Interfering Substances The influence of phthalate esters, PCB, HCB, MCP, 2,4-D and 2,4,5-T, which can be extracted with diethyl ether, was examined by taking 5 pg of each substance in distilled water. The phthalate esters were extracted and esterified with TFE but eluted slightly from the silica-gel column. Their GC retention time (ca. 2.5 min) was considerably different from that of triclopyr-TFE. PCB and HCB were extracted but eluted by the washing solvent from the silica-gel column. MCP, 2,4-D and 2,4,5-T were extracted, esterified with TFE and eluted from the silica-gel column. However, they caused no interference as they have different GC retention times (Fig. 9).Calibration Graphs and Recovery Experiments A calibration graph was prepared as follows: standard solutions containing 0.2, 0.4, 0.6, 0.8 and 1.0 pg of triclopyr were placed in 5-ml vials, the solvent was removed in a gentle stream of N2, the residue was esterified with TFE and the amount of the ester measured by ECD - GC. The calibration graph was linear over the range examined. The detection limit was 0.005 ng with a 5-pl sample injection and the determina- tion limit was 0.00025 pg ml-1 using 200 ml of sample solution. The recovery was examined by the standard procedure after adding a given amount of triclopyr to 200 ml of distilled water or river water without triclopyr. Table 3 shows that the proposed method is satisfactory, with recoveries of more than 90% and a coefficient of variation of less than 4%.Application to Real Samples The effectiveness of the proposed procedure was tested by analysing actual samples of river water (n = 12). It was found that 0.25 ml of BF3-TFE was sufficient, its use in excess causing no interference, and that the coexistence of pollutants or natural organics did not interfere with the determination. A replicate experiment on a river water containing 0.001 pg ml-1 of triclopyr gave a standard deviation of less than 4% (n = 5 ) . Fig. 10 shows the detection of 0.0015 pg ml-1 of triclopyr in a river water sampled near a site where triclopyr had been released 1 week before. Conclusion Triclopyr can be successfully extracted from acidified aqueous samples with diethyl ether and esterified with BF3-TFE.Interfering substances can be eliminated by a clean-up method with a silica-gel column and using benzene - hexane as an eluting solvent. The recovery from actual river waters is 9&93%, with coefficients of variation of less than 4%. The detection and determination limits are 0.005 ng and 0.00025 pg ml-1, respectively. The proposed procedure is also useful for the simultaneous determination of MCP, 2,4-D and 2,4,5-T if the clean-up method is modified. The authors thank Dow Chemical Japan for the gifts of triclopyr, triclopyr-BE and triclopyr-TEA and Professor I. Matsuzaki and Messrs. S. Shimizu and H. Ozawa for many helpful suggestions. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Makino, T., “Makino’s Illustrated Flora in Color,” Hokuyokan, Tokyo, 1982, p. 263. Noyaku Yoran Henshu Iinkai, “Noyaku Yoran,” Nippon Shokubutsu Boeki Kyokai, Tokyo, 1983, p. 159. Howard, S. F., and Yip, G., J. Assoc. Off. Anal. Chem., 1971, 51, 970. Khan, S . U., J. Assoc. Off. Anal. Chem., 1975, 58, 1027. Thio, A. P., Kornet, M. J., Tan, H. S. I., and Tompkins, D. H.,Anal. Lett., 1979, 12, 1009. Cotterill, E. G., Analyst, 1982, 107, 76. Goetz, R., Fresenius 2. Anal. Chem., 1983, 314, 131. Sekita, H., Takeda, M., Saito, Y., and Uchiyama, M., Eisei Kagaku, 1982, 28, 219. Chmil, V. D., Zh. Anal. Khim., 1981,36, 1121. Gutermann, W. H., and Lisk, D. J., J. Assoc. Off. Anal. Chem., 1977, 60, 1070. Woodham, D. W., Mitchell, W. G., Loftis, C. D., and Collier, C. W., J. Agric. Food Chem., 1971, 19, 186. Agemian, H., and Chau, A. S . Y., Analyst, 1976, 101, 732. Johnson, L. G., J. Assoc. Off. Anal. Chem., 1973, 56, 1503. Cotterill, E. G., J. Chromatogr., 1979, 171, 478. Draper, W. M., J. Agric. Food Chem., 1982, 30,227. Lee, H. B., and Chau, A. S . Y., J. Assoc. Off. Anal. Chem., 1983, 66, 1023. Paper A51259 Received July 16th, 1985 Accepted August 2nd, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100145

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST FEBRUARY 1986 VOL. 111 151 Preparation of a Chloride-selective Electrode Based on Mercury(1) Chloride - Mercury(l1) Sulphide on an Electrically Conductive Epoxy support J. L. F. C. Lima and A. A. S. C. Machado* Chemistry Department Faculty of Science University of Oporto 4000 Oporto Portugal A chloride-selective electrode based on mercury salts on an electrically conductive epoxy support was prepared and tested. A 1 + 1 molar mixture of mercury(1) chloride and mercury(l1) sulphide was used as the sensor. Response characteristics (linear response range limit of detection slope stability and time of response pH range and potentiometric selectivity coefficients with respect to bromide iodide and thiocyanate) were determined. The standards of performance of the electrode were found t o be better than those of a commercial electrode with the same type of sensor.Reactions taking place in the membrane are discussed. Keywords Chloride-selective electrode; mercury salts; conductive epoxy support We have previously reported a simple and easy to implement procedure for the construction of inexpensive “all-solid-state” ion-selective electrodes in which a layer of very finely powdered sensor is applied to a base of electrically conductive (silver loaded) epoxy r e ~ i n . l - ~ Using this procedure elec-trodes for silver(1) and sulphide,l32 halides,’J copper(II)1.4 and other divalent cations,l with silver salts as sensors were obtained. Their response performances were found to be similar to those of the respective commercial electrodes.1-4 This work has been extended to include electrodes based on mercury salts as sensors. This type of sensor was introduced by Sekerka and Lechner,5-7 who found that an electrode containing a mixture of mercury(I1) sulphide and mercury(1) chloride as a sensor showed improved response characteristics to chloride compared with electrodes with a sensor based on a mixture of silver sulphide and silver chloride .578 The lower detection limit of this type of chloride-selective electrode has made its use very convenient for the determination of low levels of chloride in water.8-13 The performance of a commer-cial version of the electrode (Graphic Controls PHI 91100) has also been recently evaluated.14 The use of mixtures of mercury(I1) sulphide and mercury(1) chloride as sensors for the self-construction of selective electrodes has presented some difficulties,lOJ5 probably owing to the high pressures required to obtain membranes with good mechanical proper-ties.10 However Marshall and Midgley 11 successfully applied this type of sensor to the graphite surface of a RfiiiCka Selectrode. The purpose of this study was to investigate the response characteristics of a chloride-selective electrode with this type of sensor obtained by our procedure for the construction of “all-solid-state” selective electrodes which does not require high pressures for the preparation of the membrane. Experimental Apparatus Potentials were measured with an Orion 811 digital pH meter (reading to kO.1 mV) and an Orion 605 manual electode switch.Graphs for the determination of response times were obtained with a Radiometer PHM 64 pH meter and a Servograph Rec 61 plotter. A Philips GAH 110 glass electrode was used for the measurement of pH. Orion 90-02-00 double-junction elec-~~ * To whom correspondence should be addressed. trodes (of silver - silver chloride type) were used as reference electrodes (inner filling solution Orion 90-00-02; outer filling solution 10% potassium nitrate). Reagents The water used in the preparation of all the standard reagent solutions was de-ionised (Elgastat B114 mixed-bed column unit) and distilled in a quartz still (Heraeus B1 18 double distillation unit). All chemicals were of analytical-reagent grade and were used without further purification.When necessary stock solutions were standardised by potentio-metric titration. Further details are given elsewhere.14 Preparation of the Electrodes Mercury(I1) sulphide (black) was prepared by precipitation, initiated by slowly mixing equal volumes of 0.1 M sodium sulphide and 0.1 M mercury(I1) nitrate solutions. The sensor was prepared by thorough grinding of a 1 + 1 molar mixture of mercury( 11) sulphide and mercury(1) chloride (Merck Ref. No. 4425). The procedure used previously1 for the preparation of electrodes with silver salt sensors was followed. A piece of silver-loaded commercial epoxy (EPO-TEK 410) was applied to an end of Perspex tube (0.d. 1 cm length ca. 15 cm) to constitute a layer ‘of about 0.7 cm thickness; a shielded cable was fixed to the epoxy inside the tube and after hardening (at 100 “C for 1 h) a conical cavity was drilled in the epoxy layer.A new piece of epoxy was applied to this cavity and the very thinly powdered sensor was blown against it while still fresh from a Pasteur pipette this operation being repeated several times to obtain a continuous coat of sensor on the epoxy. After hardening (at 80 “C for 4 h) the other end of the tube was closed with Perspex glue. Finally the sensor layer was polished over glass (Wilks 004-10001) and then with polishing paper (Orion 94-82-01). Procedure for the Evaluation of the Electrodes Standard techniques were used for evaluating the response characteristics of the prepared electrodes. All the measure-ments were made with the electrodes immersed in solutions kept at 25.0k0.2 “C.Except where otherwise stated the ionic strength of the solutions was adjusted to 0.1 M and the pH to 3 with potassium nitrate and nitric acid 152 310 290 270 > E 250 230 210 190 ANALYST FEBRUARY 1986 VOL. 111 -------The slope S and standard potential EO were obtained from the experimental points in the linear range of calibration by a least-squares adjustment performed by standard pro-grams of pocket calculators. R is the correlation coefficient of linear regression given by the programs which measures the goodness of fit. Results Characteristics of Electrode Response to Chloride Reproducibility of preparation In order to assess the reproducibility of the preparation procedure and the stability of response (see below) five units were simultaneously calibrated daily with standard solutions of sodium chloride in the range 4 X 10-4-10-2 M for a period of more than 2 weeks.2-4 Between calibrations the electrodes were left in de-ionised water as this was found to be a suitable conditioning medium.(When not in use the electrodes were stored dry and in the dark and before their re-use were polished and conditioned.) Table 1 presents typical results of the calibrations obtained with one of the units (A). Table 2 gives the average values of the calibration parameters and their standard deviations for the five units (A-E). The results presented in Table 2 show that the procedure used for electrode construction yields units with reproducible response characteristics.Stability of response With respect to calibration graphs the electrodes retain their characteristics over several weeks (Table 1) without any need Table 1. Stability of the calibration graph* of an electrode unit?,$ E (10-3)/mv S/mV Time/d decade-' E'ImV R Calculated Read 1 1 2 2 3 3 8 8 9 9 10 10 18 18 -56.7 -57.0 -56.3 -56.3 -55.8 -54.9 -57.4 -57.5 -57.4 -57.5 -57.9 -57.3 -57.5 -57.6 40.0 39.6 40.2 39.8 40.2 40.0 36.8 36.8 36.6 36.9 35.4 36.0 35.2 34.5 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 210.0 209.5 209.2 208.7 207.4 204.6 209.2 209.2 208.8 209.5 209.0 208.0 207.7 207.4 210.0 209.3 208.9 208.3 207.3 204.3 209.0 209.2 208.9 209.5 208.9 208.1 207.4 207.5 * In the range 4 x 10-4-10-2 M chloride.t Unit A in Table 2. $ Symbols S = slope; E" = standard potential (vs. S.C.E.); = response M chloride; Calculated from calibration graph; Read direct R = correlation coefficient of linear regression; E to reading. Table 2. Average calibration parameters of several electrode units* ,? S/ Unit mV decade-' EOImV A 56.9 (0.8) 38 (2) B 57.8 (0.8) 33 (2) C 57.0 (1.0) 36 (2) D 56.0 (2.0) 37 (4) E 57.0 (2.0) 37 (4) * Averages of 14 calibrations over 18 d with standard deviations in t For symbols see Table 1. parentheses. for restoring the membrane by polishing provided that they are kept in water between measurements.The reproducibility of repeated calibrations obtained during the same day was generally found to better than 0.5 mV decade-' for the slope and 1 mV for the standard potential (Table 1). A decrease in slope was normally found after a month or more of use but the response characteristics could be restored by polishing the membrane with Orion paper followed by conditioning in water. This treatment was found to be very effective even when the membrane had been subjected to strong interferences. Under normal use the electrodes do not require frequent polishing and consequently although the membrane thick-ness is small (less than 1 mm) they have high durability. Several units have been used for the determination of chloride in high-purity waters for almost two years and still show good response characteristics.Such durability is a definite advan-tage over electrodes with the same type of sensors in pressed membranes which as mentioned by Tacussel and Fomboml5 and confirmed by ourselves in a previous evaluation of two units of the Graphic Controls electrode,14 show frequent periods of irregular response and require repeated polishing, which wears out the membrane. Lower limit of linear response and limit of detection Typical calibrations by the standard additions technique for the determination of these parameters are shown in Fig. 1. Values of ca. 10-5 and ca. 5 x 10-6 M for the lower limit of linear response and the limit of detection ,I6 respectively were obtained both being similar to the corresponding values found for a commercial electrode with the same type of sensor.14 Although less intense than for the commercial electrode a "memory effect" was found; for example if the electrode had been immersed in a 10-1 M chloride solution before calibration values of ca.10-4 M were found for the lower limit of linear response. The re-establishment of the normal value above for this parameter requires polishing followed by immersion in water for at least 30 min. The value for the lower limit of linear response falls within the range of values for the parameter reported in the literature (2 X 1W6-2 X 10-5 ~8,11,14,15) and extension of the linear range is characteristic of the mercury salt-based chloride electrodes compared with those based on silver salts.A 100-fold increase in linear response range has been reported5 for pressed membrane electrodes but previously with a commercial electrode with this type of membrane an approxi-mately 20-fold increase was found14 as in this work. An interesting feature shown by the calibration (Fig. 1) is the small variation of potential below the limit of linear 1.00 x 10-6 1.00 x 10-5 1.00 x 10-4 1.00 x 10-3 acl-Fig. 1. Typical calibration graph for the electrod ANALYST FEBRUARY 1986 VOL. 111 190 > E Li 220 153 . 2.32 x ' I 0 - 3 ~ 1.03 x 10-4M 4.54 x 10-4 M t-l 12.19 x 1 0 - 4 ~ I min 250 I 1.00 x 10-4M 280 ' Time -Fig. 2. Typical recorder output of dynamic response time determina-tions for varying concentrations of chloride response only about 20 mV in the decade from 10-6 to 10-5 M.A similar situation was found by us for the Graphic Controls electrode.14 In both situations the potential varia-tion is smaller than that shown by the electrode prepared by Marshall and Midgley11J2 by coating a Rfiiitka Selectrode with the same type of sensor. As a consequence in this instance the practical limit of detection is close to the lower limit of linear detection and the usefulness of the electrode below the latter limit is limited unless perhaps rigorously controlled conditions12 are applied. Slope The slope of ca. 57 mV decade-1 for the units studied (Table 2) was slightly lower than the theoretical value. Although Sekerka and Lechner5.8 reported a Nernstian response Tseng and Gutknechtl7 were unable to obtain the same slope for the electrodes prepared by a similar procedure and obtained slopes of 50-51 mV decade-1.However other worker~1~J1J5 have reported slopes in the range 57-59 mV decade-l for electrodes with the same type of sensor constructed by different procedures as was found by us in this and in previous work. 14 Standard potential Correcting the average value of this parameter for the units included in Table 2 +36 mV by +242 mV to express S.C.E. against N.H.E. and -7 mV ( S x logfcl-,fcl- = +0.755 at Z = 0.1 MIS) to compensate for the ionic strength a value of +271 mV vs. N.H.E. was obtained as the standard potential. This value is close to +268 mV the standard potential of the Hg2C12 - Hg system,19 suggesting that there is metallic mer-cury in the membrane to fix aHg at 120921 (see Discussion).Response time As shown in Fig. 2 the dynamic response time of the electrode for increases in chloride concentrations in the range 10-4-10-2 M is less than 1 min. Sekerka and Lechner5.8 quoted shorter response times for their electrode than for chloride-selective electrodes with silver salts as sensors but the present electrode and the corresponding one of the latter type3 showed similar response times. The electrode prepared by Marshall and Midgleyll by application of the sensor on a Rfiiitka Selectrode was also found to be slower than the pressed membrane electrode of Sekerka and Lechner.5~8 Our previous evaluation14 of the Graphic Controls commercial electrode did not show any improvement of response times over electrodes with silver salts as sensors.When the electrode is exposed to a decrease in chloride concentration response times are much longer especially if the initial value of concentration is high. This "memory effect" 340 300 260 220 > 5 180 UI 140 100 60 20 0 2.00 4.00 6.00 8.00 10.00 12.00 PH Fig. 3. Variation of response potential with pH for various concen-trations of chloride A 1.00 x 10-5 M; B 1.00 X 10-4 M; C 1.00 X M; D 1.00 x 10-2 M; and E 1.00 x lo-' M (cf. discussion under Lower limit of linear response and limit of detection) is much more pronounced for the electrode reported here than for the electrode based on silver salts. This has also been found for the Graphic Controls electrode,l4 based on a pressed membrane sensor and also for the electrode of Marshall and Midgley.11 As the degree of compactness of the sensor in the membrane is different for these three electrodes the effect cannot be ascribed only to surface irregularities.A possible explanation is a tendency for adsorption of chloride at the membrane surface which may be related to the high values of the stability constants of [HgCl,](n-2)- complexes. Indeed such values are much higher than those of the corresponding [AgCl]("-l)- species.22 It was observed that when this electrode was immersed in water for conditioning its potential suffered a quick change until a stable value was reached which may be due to the washing out of chloride from the surface of the membrane.Effect of pH on the Response The influence of the pH on the response of the electrode at fixed chloride concentrations between 10-1 and loe5 M (in solutions with an ionic strength adjusted to 0.2 M with potassium nitrate) is presented in Fig. 3. The potential is independent of pH from a lower limit of pH between 1.5 and 2 and an upper limit that depends much more markedly on the chloride concentration than for the hydroxide interference in other solid membrane electrodes e.g. for chloride-selective electrodes with silver salts as sensors.3 The marked variation of the upper limit of the operative response plateau can be understood when the value of KCI,OH = aC1-/aOH- is calculated from the solubility products of Hg2C12 and Hg20,14 the large value obtained (ca.3 x 102) explaining the observed variation. This feature of the chloride-selective electrode with an Hg2C12 - HgS sensor has not been discussed in the literature where a value of pH 6 is invariably indicated as the upper limit of operation.s.15.23 These results show that this value of pH is too high for the measurement of chloride at low concentrations (less than ca. 10-4 M) and support the pro-cedure established by Marshall and Midgleyl' where the hydrogen ion concentration is fixed at a constant value of M 154 ANALYST FEBRUARY 1986 VOL. 111 Table 3. Potentiometric selectivity coefficients gg:x This work* at chloride concentration Reference X 10-3 M 10-4 M Calculatedt 5 15 14$ 23 - SCN- 70 k 12 7 k 1 40 2 - 25 -Br- 70 k 11 43 k 6 2 x 104 6.3 x 102 - 102 -25 6 x 102 1- 30 k 9 5 + 1 3 x 10'0 3.2 x 103 -10 - 25 3 x 103 * Values are averages of six results (duplicate determinations with three units) obtained at the chloride concentrations given.t Calculated14 by K E F Kso(Hg2C12) - Kso(Hg2X2) using values of K, given in reference 27. $. Obtained at a M chloride concentration. A comparison of Fig. 3 with similar results for the Graphic Controls electrode14 shows that this electrode is more sensitive to pH changes below 2 which may be explained by the sensor being less compact when on the epoxy support than in a pressed membrane where solubilisation is more difficult. Interferences For this type of electrode there are discrepancies between literature values5J4J5J3 for the potentiometric selectivity coefficients relative to interferences of anions whose mer-cury(1) salts are more insoluble than mercury(1) chloride (namely bromide iodide and thiocyanate) as well as anomal-ies in their relative values found by some workers.14J5 In this study the coefficients were determined by the mixed solution method (with the chloride concentration fixed at 10-3 and 10-4 M without adjustment of the ionic strength but with the pH fixed at ca.4 M with nitric acid). The results of replicate determinations showed a certain degree of variability owing to the lack of reproducibility of the straight segments corre-sponding to response to interferences even when the experimental conditions (rate of interferent addition criterion of readings etc.) were kept constant.Therefore the experimental values presented in Table 3 which are averages of the results of six determinations should be considered only as orders of magnitude of the parameter. In Table 3 literature values are also included for comparison as well as values obtained from the solubility products which were calculated using standard procedures. The experimental values of the selectivity coefficients obtained for the electrode on a conductive epoxy support were much lower than those predicted by calculations. This result is similar to that found with electrodes with silver salts,lJ but the differences between the experimental and calculated values appear to be more accentuated in this instance. Moreover as found by Tacussel and Fombomls for the bromide and iodide interferences the relative strength of interferences found follows the order SCN- 3 Br- 3 I- which is opposite to the order predicted from calculations.In our previous study of the Graphic Controls electrode which for interferences was less detailed than this one about the same value was obtained for the selectivity coefficients with respect to the three interfer-ents.14 Tacussel and Fombomls have suggested that such anomalies have kinetic causes. Another interesting point shown by these results is that the values of the selectivity coefficients at the 10-4 M chloride level are smaller than the corresponding values at 10-3 M. The electrode seems to feel the effect of these interferences less extensively at lower concentrations and this may also be a consequence of slower response to lower concentrations of interferents.However it should be pointed out that in these experiments in contrast to observations in similar experi-ments with chloride-selective elctrodes with silver salts as sensors,3 visual inspection of the membrane did not show any strange precipitates on its surface. As the chemical behaviour of an Hg2C12 - HgS membrane seems to be extremely complex (see Discussion) an interference mechanism more complex than that accepted for electrodes with silver salts as sensors cannot be ruled out. Response to the Mercury(1) Ion Calibrations for the response to mercury(1) ion were obtained by the titration technique using lo-210-4 M mercury(1) nitrate solutions. These showed a slope slightly higher than the theoretical one 32.0 (0.5) mV decade-1 and a standard potential of 819 (2) mV vs.N.H.E. These are average values obtained from four determinations with three units the standard deviations being given in parentheses. The standard potential includes a correction of +13 mV for ionic strength compensation calculated as before with fHg2+ = 0.355 at Z = 0.1 ~ . 1 8 The value of the standard potential is similar to that found previously14 for the Graphic Controls electrode (813 mV) both being greater than the standard potential of the Hg2+ - Hg system (792 mV).19 The difference between the electrode standard potential and the standard potential of the relevant couple is larger in this instance than when the response to chloride is considered. Discussion Our previous work on ion-selective electrodes based on mixtures of silver salts of divalent metal sulphides and silver(1) sulphide on a conductive epoxy support has shown that this construction procedure yields electrodes with response characteristics similar to those of the corresponding commer-cial electrodes.1-4 There is indirect evidence that in these electrodes the metallic silver in the epoxy support does not contact the solution.374J4 This is also suggested by observation of the surface of the membranes by scanning electron microscopy which shows that in the small resin areas (ca.10-3-10-4 mm2) exposed between sensor microcrystals the silver signal is very weak (less than 1% of the signal of microcrystals and probably having this origin) .25 In this situation the operation of the constructed electrode was found to be less troublesome than for the commercial electrode evaluated previously,l4 even though the numerical values of the characteristic parameters of electrode response are very similar.Less frequent polishing for maintaining electrode performance was required and sudden outbreaks of bad behaviour of the electrode were not observed as for the Graphic Controls electrode. 14 Other problems of the chloride and other selective electrode with pressed membranes based on mercury salts e.g. irreproducibility of response15 or troublesome operation,26 have been discussed in the litera-ture. 1 0 ~ 5 ~ 7 ~ 2 6 The proposed method of construction minimises such problems. The values of the standard potentials of this electrode in response to chloride (+271 mV) and mercury(1) ions (+819 mV) are respectively close to the values for the Hg2C12 - Hg and Hg2+ - Hg couples and yield a value of 2.8 X 10-19 for the solubility product of Hg2C12 in agreement with literature values,27 e.g.1.3 x 10-18 ~ 3 . These data show that the electrode responds to chloride as a second kind electrode ANALYST FEBRUARY 1986 VOL. 111 According to Koebel20 and Buck and Shepard,21 this requires the occurrence of free mercury in the membrane to fix the value of the activity of the metal equal to unity. The presence of the free metal in the sensor is understood if the value of the equilibrium constant for the disproportiona-tion of Hg2C12 (s = solid sol = solution) is considered.It can be calculated from the solubility products of Hg2C12 (1.3 X 10-18) and HgS (black) (1.6 x lO-52)27 and the equilibrium constant of the disproportionation (Kdisp) by the expression &isp can be calculated from the standard potentials of the couples Hg2+ - Hg22+ (+907 mV) and Hg22+ - Hgo (+792 mV),19 to be &iSp = 10-1.95. A value of ca. 1032 is obtained for K. This value is so large that the reaction can occur even for large although reasonable values of chloride concentration in solution. Sulphide provided by the intrinsic solubility of the mercury(I1) sulphide is used up i.e. the reaction (1) occurs instead of the dissolution of this salt. Alternatively if the intrinsic solubility of Hg2CI2 is con-sidered [or when the electrode is immersed in a solution of mercury(1) ion] precipitation of free mercury at the mem-brane surface is explained by the reaction Hg2C12(~) + S’-(Sol) = Hg(1) +HgS(s) + 2C1- (sol) (1) Hg22+ (sol) = Hg(1) + Hg2+ (sol) .. (2) K = [CI-l2/[S2-] = &is, X Kso(Hg~C1~)IKso(HgS) (3) Hg22+(~01) + S2-(sol) = Hg(1) + HgS(s) . . (4) of which the equilibrium constant K’ = l/([Hg22+] X [S2-]) = Kdisp/Ks,(HgS) . . ( 5 ) is also very large (K‘ = 1050). These calculations show that the disproportionation of mercury(1) ion in the membrane is spontaneous in the thermodynamic sense and is expected to occur indefinitely with conversion of Hg2C12 into HgS (and free mercury) with the release of chloride into the solution. It is interesting to observe that Hulanicki et al.,26 in a paper discussing the construction of a bromide-selective electrode based on mer-cury salts where reactions similar to those above are involved,28 reported the appearance of small droplets of metallic mercury on the walls of the die where a membrane consisting of Hg2C12 HgS and Ag2S had been pressed with thermal treatment.This thermodynamic instability may explain the problems connected with the troublesome operation observed for the chloride and other electrodes with pressed membranes based on mercury salts.1O714J5J7926 A practically stable response of such electrodes requires that a metastable equilibrium state be reached in the membrane. The results of this work and their comparison with previous14 and literature results suggest that the use of an unpressed mixture makes the attainment of such a metastable state easier than when the mixture is disturbed by pressure and heating.Conditioning of the mercury salt electrodes in water after polishing the final stage of the construction or when the membrane is regenerated has been recommended5J3J4J3 and was found to be necessary in this work in order to obtain acceptable response characteristics. The effectiveness of such conditioning may result from the removal of chloride and other soluble species formed by the disproportionation reactions (discussed above) from the membrane surface. Financial support from INIC Lisbon (C.I.Q.U.P. Line 4A) and JNICT Lisbon (Research Contract 50.78.127) is grate-155 fully acknowledged as well as helpful discussions with Dr. J. D.R. Thomas (Cardiff Wales) made possible by a Travel Grant from the Scientific Affairs Division of the North Atlantic Treaty Organisation (Grant No. 069/84). We thank Mr. A. J. T. Sousa and Mrs. M. Isabel R. G. F. Sampaio for carrying out routine work. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. References Lima J. L. F. C. and Machado A. A. S . C. in Albaiges J., Editor “Analytical Techniques in Environmental Chemistry,” Volume 2 Pergamon Press Oxford 1982 p. 419. Lima J . L. F. C. and Machado A. A. S . C. Rev. Port. Quim., 1979 21 15. Lima J. L. F. C. and Machado A. A. S . C. Rev. Port. Quim., 1979 21 153. Lima J. L. F. C. and Machado A. A. S .C. Rev. Port. Quim., 1982 24 156. Lechner J. F. and Sekerka I. J. Electroanal. Chem. 1974, 57 317. Sekerka I. and Lechner J. F. J . Electroanal. Chem. 1976, 69 339. Sekerka I. and Lechner J. F. Anal. Lett. 1976 9 1099. Sekerka I. Lechner J. F. and Wales R. Water Res. 1975,9, 663. Sekerka I. Lechner J. F. and Harrison L. J. Assoc. Off. Anal. Chem. 1977,60,625 Bailey P. Wilson J. Karpel S. and Riley M. in Pungor E., Editor “Conference on Ion Selective Electrodes Budapest, 1977,” Elsevier Amsterdam 1978 p. 201. Marshall G. B. and Midgley D. Analyst 1978 103 438. Marshall G. B. and Midgley D. Analyst 1979 104 55. Ryan,.T. E. Peterson A. J. and Subsara W. P. in Moody, G. J. Editor “International Symposium on Electroanalysis in Clinical Environmental and Pharmaceutical Chemistry,” UWIST Cardiff 1981 paper 5.Lima J. L. F. C. andMachado A. A. S . C. Rev. Port. Quim., 1982 24 61. Tacussel J. and Fombom J . J. in Pungor E . Editor, “Conference on Ion Selective Electrodes Budapest 1977,” Elsevier Amsterdam 1978 p. 567. Guilbault G. G. Editor “Recommendations for Publishing Manuscripts on Ion-Selective Electrodes,” Pure Appl. Chem., 1981 53 1907. Tseng P. K. C. and Gutknecht W. Anal. Chem. 1976 48, 1996. Kielland J. J. Am. Chem. SOC. 1937 59 1675. Lurie J . “Handbook of Analytical Chemistry,” Mir Moscow, 1975 p. 305. Koebel M. Anal. Chem. 1974,46 1559. Buck R. P. and Shepard V. R. Anal. Chem. 1974,46,2097. Sillen L. G. and Martell A. G. Editors “Stability Constants of Metal Ion Complexes,” Special Publication No. 17 Chem-ical Society London 1964 pp. 286-288 and 292-293. “Ultra-Sensitive Solid State Chloride Electrode PHI 91 100 (Instruction Manual) ,” Graphic Controls Buffalo NY USA. da Silva M. G. P. Lima J . L. F. C. and Machado, A.,A. S . C Port. Electrochim. Acta 1984 2 29. Lima J. L. F. C. Machado A. A. S. C. and SB C. M., “Abstracts of the Sixth Meeting of the Sociedade Portuguesa de Quimica,” Sociedada Portuguesa de Quimica Aveiro 1983, paper PC36. Hulanicki A. Lewandowski R. and Lewenstam A. Anal. Chim. Acta 1979 110 197. Lurie J. “Handbook of Analytical Chemistry,” Mir Moscow, 1975 p. 110 Lima J. L. F. C. and Machado A. A. S . C. Port. Electrochim. Acta submitted for publication. Paper A51269 Received July 22nd 1985 Accepted August 8th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100151

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 157 Reduction in Size by Electrochemical Pre-treatment at High Negative Potentials of the Background Currents Obtained at Negative Potentials at Glassy Carbon Electrodes and its Application in the Reductive Flow Injection Amperometric Determination of Nitrofurantoin Ahmad B. Ghawji and Arnold G. Fogg Chemistry Department, Loughborough University of Technology, Loughborough, Leicestershire LEI I 3TU, UK The reduction of dissolved molecular oxygen a t a glassy carbon electrode was shown to be made more difficult on electrochemically pre-treating a newly polished glassy carbon disc electrode (3 mm in diameter) in 0.1 M sulphuric acid solution at -2.7 Vfor 1 min. By this means a background current of only 1 pA was obtained when this electrode was held at -1.05 V in a flow injection system incorporating extensive PTFE transmission tubing and using a deoxygenated pH 7 Britton - Robinson buffer as eluent (flow-rate 6.5 mi min-1).The size of the signal obtained when 100 1-11 of eluent that had not been deoxygenated were injected as the sample blank was only 0.02 FA at -0.7 V. When nitrofurantoin was determined using these latter conditions, this blank signal was equivalent to about 2 x 10-7 M nitrofurantoin. Keywords: Electrochemical pre-treatment; amperometric detection; reduction; flow injection analysis; nitro furantoin determination Increasing attention is being paid to the advantages of electrochemically pre-treating glassy carbon electrodes used for amperometric detection in HPLC and in flow injection analysis.1-8 The studies reported to date have been made to improve the performance of glassy carbon electrodes used for monitoring oxidation processes at positive potentials.In many irreversible oxidation processes, electrochemical pre- treatment first at a high positive potential and then at about -1.0 V reduces the overpotential for oxidation of the determinand such that an improved hydrodynamic voltammo- gram is obtained. Oxidation occurs more completely and at a less positive potential such that a higher and more reproduc- ible signal is obtained. At any particular potential the background signal is also increased, but this does not detract significantly from the technique. An important HPLC method that involves reductive amperometric detection at a glassy carbon electrode held at negative potentials is the determination of vitamin K and its analogues.9JO Hart et ~1.10 determined vitamin K1 in a 95% methanol eluent that was 0.05 M in a pH 3 sodium acetate - acetic acid electrolyte holding the potential of the glassy carbon electrode at - 1.0 V; the eluent was deoxygenated with nitrogen and an all-metal solvent delivery system was used to prevent the re-entry of oxygen.Calibration graphs were obtained by injecting 1-10 ng of vitamin K1. Adsorbed product on the electrode was removed periodically by holding the electrode at +0.7 V, which re-oxidised the product as the reduction process is quasi-reversible. Hanging mercury drop electrodes have been used by other workers, notably by Lloydll-13 for determining explosives residues, as detectors of reductive processes.The rigorous exclusion of oxygen has been an important feature of all of these methods. In the work described in this paper a study was made of the possibility of improving signals for reductive processes at glassy carbon electrodes held at negative potentials by applying electrochemical pre-treatment. Experimental Flow injection analysis was carried out in a single-channel system that has been described previously.14 Eluent flow was produced by means of an Ismatec Mini-S peristaltic pump. Sample (approximately 100 pl) was injected with a Rheodyne 5020 low-pressure injection valve connected to a laboratory- built detector cell by means of 50 cm of 0.58 mm bore PTFE tubing. The detector cell holds the glassy carbon electrode only, eluent being presented to it in a wall-jet configuration.The cell is used partially immersed in an electrolyte having the same composition as the eluent. A counter platinum and a conventional potentiometric calomel reference electrode are placed in the electrolyte to obtain electrical contact with the working electrode. The glassy carbon disc electrode (3 mm diameter) was constructed from Le Carbonne glassy carbon and was mounted in PTFE. An eluent flow-rate of 6.5 ml min-1 was used. The potential of the glassy carbon electrode was controlled by means of a PAR 174 polaro- graphic analyser and current signals were monitored on a Linseis L650 y - t recorder. Linear sweep voltammetry was carried out at a sweep rate of 10 mV s-1 using the same working, counter and reference electrodes immersed in the appropriate measuring solution.Preliminary Linear Sweep Experiments in a Static System During studies of the effect of positive- and negative-potential electrochemical pre-treatments of glassy carbon electrodes on oxidation processes at low positive potentials, it was noticed that electrochemical pre-treatment at high negative potentials was effective in making smaller the background currents obtained at negative potentials. This is clearly illustrated in Fig. 1, in which base-line linear sweep voltammograms obtained with a static electrode system in 0.01 M sulphuric acid before and after electrochemical pre-treatment are shown; the electrode was electrochemically pre-treated in 0.1 M sulphuric acid. Pre-treatment at -3 V in the static mode is seen to remove the oxygen reduction wave most effectively.Further, electrochemical pre-treatment was shown to be effective only when carried out in dilute sulphuric acid; attempts to effect pre-treatment in Britton - Robinson buffer solution of pH between 2 and 8 were unsuccessful. The pre-treatment that had been effected in dilute sulphuric acid, however, was also as effective when the electrode was used in these buffer solutions. Linear sweep voltammograms obtained for the reduction of158 11 8 f . c If 2 3 0 4 0 -0.4 -0.8 PotentialiV Fig. 1. Blank linear sweep voltammograms in 0.01 M sulphuric acid without deoxygenating the solution. A and A’, first and second scans at a newly polished glassy carbon electrode; B, scan after pre- treatment at -2.5 V for 1 min in 0.1 M sulphuric acid; and C, scan after pre-treatment at -3 V for 1 min in 0.1 M sulphuric acid I 4 f .5 0 c L 3 0 8 4 0 I / -0.4 -0.8 Potent i a I N Fig. 2. Linear sweep voltammograms of nitrofurantoin (2 x M) in undeoxygenated pH 7 Britton - Robinson buffer. (a) At a newly polished glassy carbon electrode; and (b) at an electrode pre-treated at -3 V for 1 min. The blank linear sweep voltammograms are given as broken lines in both instances 8 f . c C ? 3 0 4 0 ANALYST, FEBRUARY 1986, VOL. 111 -0.4 -0.8 Potent ia I/V 1.2 Fig. 3. Linear sweep voltammograms of cephalonium (100 pg ml-1) in 0.1 M sulphuric acid. A, Without deoxygenating the solution at a newly polished glassy carbon electrode; B, without deoxygenating the solution at an electrode pre-treated at -3 V for 1 min; and C, after deoxygenating the solution at a newly polished glassy carbon electrode nitrofurantoin, which occurs at about -0.58 V in pH 7 Britton - Robinson buffer, are shown in Fig.2. When the electrode is newly polished a hump due to the reduction of dissolved molecular oxygen is apparent as a post-peak. After pre- treating the electrode at -3 V this hump is no longer apparent. An illustration of the oxygen reduction process occurring before that of a determinand is shown in Fig. 3, in which linear sweep voltammograms for the reduction of the cephalosporin cephalonium are shown. Here also electrochemical pre- treatment at -3 V removes visible signs of the oxygen reduction process. Polarographic methods are available for the determination of nitrofurantoin15 and cephalonium.16 Effect of Electrochemical Pre-treatment at High Negative Potentials in Flow Injection Analysis In using a glassy carbon electrode for amperometric detection in HPLC or flow injection analysis, two characteristics of the system should be considered before the quality of the signal obtained with the determinand is studied. These are the background current associated with the eluent and the blank signal obtained when a control blank is injected. When the eluent is used as the control blank, clearly eluent and sample are the same and no signal should be observed when the control blank is injected, except at high sensitivities owing to disturbance to the flow of eluent caused by the process of injecting the eluent.In determinations made at potentials where oxygen reduction occurs, however, a finite blank signal will be observed if the oxygen contents of the eluent and the blank sample solution differ. A negative signal will be observed if the eluent contains more oxygen than the blank sample. Clearly analytical determinations become very unreli- able when the level of interferent in the solvent system and sample solution have to be balanced, and this is particularly so with dissolved molecular oxygen. In general, with increasing background current the detection limit attainable is increased; with the system used in this work it has generally been observed that if the background current reaches 1 yA then determinations can only be made down to about 5 x 10-6 M and that this also applies at positive potentials where the background current is not associated with the reduction of oxygen.The results reported here for electrochemically pre-treated electrodes were obtained with electrodes pre-treated either atANALYST, FEBRUARY 1986, VOL. 111 3 - f . 5 2 - 4- L J u 1 - 0 - -2.7 V for 1 min in a static system before being inserted into the detector cell or at -3 V for 1 min on-line in 0.1 M sulphuric acid at a flow-rate of 2 ml min-1. These were found to be the optimum off-line and on-line electrochemical pre-treatment conditions. This latter process was readily effected by switch- ing eluents before the pump. The use of higher pre-treatment potentials than those recommended led to higher background noise. The background current levels obtained at various potentials with the flow injection system in which pH 7 Britton - Robinson buffer was used as the eluent are shown in Fig.4. These were obtained for a newly polished electrode and for pre-treated electrodes in all instances with and without deoxygenation of the eluent with nitrogen (it should be borne in mind that the term “deoxygenation,” which is used extensively by polarographers, is misleading in that the oxygen concentration is reduced only to a particular level that is determined by the effectiveness of the “deoxygenation” process and also by the effectiveness of preventing oxygen from re-entering the eluent before the measurement is made). It is clear from Fig. 4 that electrochemical pre-treatment extends the useful range of the electrode to more negative potentials both when the eluent is deoxygenated and when it is not and that the static electrochemical pre-treatment process is more effective than the on-line pre-treatment.In effect, on pre-treatment the overpotential for the reduction of oxygen at the glassy carbon electrode is being increased, i.e., the reduction of oxygen is being made more difficult by the pre-treatment process. Perhaps not surprisingly, electrochemical pre-treatment has a more significant effect on the useful range of the electrode when the oxygen content of the eluent has been reduced to a lower level by deoxygenation. Nevertheless, deoxygenation of eluent and sample solutions is a time-consuming task and there is a distinct advantage to be gained in avoiding the necessity of having to carry it out.Compounds such as vitamin Ks, which can be determined at potentials less negative than -0.5 V, can be determined at low levels even with an unpre-treated electrode without having to deoxygenate the eluent and sample solutions. Nevertheless, even in these instances the background current is reduced and the detection limit should be lowered by using a pre-treated electrode. Electrochemical pre-treatment produces a slight extension of the useful range of the electrode in an eluent that has not been deoxygenated and this should allow other compounds to be determined without the need to deoxygenate the eluent or sample solutions, particularly if determinations are to be made at high concentrations. The extension of the useful range on pre-treating the glassy carbon electrode, however, is much greater for the deoxygen- ated eluent.From Fig. 4 it can be seen that the potential at which a background current of 1 pA is obtained is moved from -0.72 to -1.05 V on pre-treating the electrode at -2.7 V in the static mode. Hence electrochemical pre-treatment should make amperometric detection possible for compounds that are reduced (or oxidised) at these more negative potentials. Again, the added advantage that lower background currents are obtained in determining compounds at lower negative potentials should not be overlooked. Deoxygenation of eluent in a flow injection system by means of nitrogen is readily carried out, and nitrogen can be bubbled continuously through the eluent in the eluent reservoir during determinations with no great inconvenience or loss of time once the initial deoxygenation has been effected.Deoxygenation of every sample solution is extremely time consuming, however, and the need to do this should be avoided if at all possible. The size of signals obtained at various potentials on injecting pH 7 Britton - Robinson buffer that had not been deoxygenated into deoxygenated eluent of the same composition is illustrated in Fig. 5. These results were obtained with both newly polished and electrochemically pre-treated electrodes. The marked effect of electrochemical 15 f 10 2- 4- E 3 0 5 1 0 -0.8 PotentiaW 159 Fig. 4. Background currents obtained with flow injection ampero- metry using pH 7 Britton - Robinson buffer as eluent.A, Without deoxygenating the solution at a newly polished glassy carbon electrode; B, without deoxygenating the solution at an electrode pre-treated at -3 V for 1 rnin on-line; C, after deoxygenating the solution at a newly polished glassy carbon electrode; D, after deoxygenating the solution at an electrode pre-treated at -3 V for 1 rnin on-line; and E, after deoxygenating the solution at an electrode pre-treated at -2.7 V for 1 rnin (in a static system) I 4 t r I -0.5 -0.7 -0.9 PotentiaVV Fig. 5. Blank hydrodynamic voltammograms representing the size of signals obtained on injecting undeoxygenated eluent into deoxy- genated eluent ( H 7 Britton - Robinson buffer). A, At a newly polished electrog; B, at an electrode pre-treated on-line in 0.1 M sulphuric acid at -3 V for 1 min; and C, at an electrode pre-treated in the static mode at -2.7 V for 1 min pre-treatment on the size of the signal obtained can be clearly seen.The potential at which the blank signal due to oxygen in the blank sample reaches 1 pA is moved from -0.77 to -0.95 V on pre-treating the electrode at -2.7 V in the static mode. In Figs. 6 and 7 are shown hydrodynamic voltammograms of nitrofurantoin in pH 7 Britton - Robinson buffer using newly polished and pre-treated electrodes, respectively. In both instances the hydrodynamic voltammograms that are shown were obtained using deoxygenated eluent. The effect of deoxygenating the sample on the hydrodynamic voltammo- grams obtained is also clear. Blank hydrodynamic voltammo- grams in which undeoxygenated eluent was injected intoANALYST, FEBRUARY 1986, VOL.111 D - - T I 0.1 pA & -0.4 -0.8 Potent i a l/V Fig. 6. Hydrodynamic voltammograms obtained at a newly polished electrode for injection of nitrofurantoin (2 x 10-4 M) into deoxygen- ated pH 7 Britton - Robinson buffer. A, Sam le solution undeoxygen- ated; B, sample solution deoxygenated; and 8, undeoxygenated blank injection I 0.1 pA D -Time -0.4 -0.8 PotentialN Fig. 8. Signals obtained near the determination limit for nitrofuran- toin at (a) a newly polished electrode and (b) at an electrode pre-treated at -2.7 V in the static mode for 1 min. Measurement potential = -0.65 V. Nitrofurantoin concentration: A, 0; B, 2 X 10-7 M; C, 5 x M; and D and D’, 10 x 10-7 M. Eluent and sample solution D’, deoxygenated; sample solutions, A-D, undeoxygenated Fig. 7.Hydrodynamic voltammograms obtained at an electrode pre-treated in the static mode at -2.7 V for 1 min for injection of nitrofurantoin (2 x 10-4 M) into deoxy enated pH 7 Britton - Robinson buffer. A, Sample solution uncfeoxygenated; B, sample solution deoxygenated; and C, undeoxygenated blank injection deoxygenated eluent are also shown. It should be noted that the signals due to the reduction of nitrofurantoin are made smaller by the pre-treatment process. Clearly, the reduction of nitrofurantoin is also being inhibited, although not to the same extent as the reduction of oxygen. The precision of the signals, however, remained excellent. The beneficial effect of the electrochemical pre-treatment in making smaller the size of the oxygen signal can be clearly seen in Fig.7. At the current sensitivity used in obtaining the hydrodynamic volt- ammograms shown in Fig. 7 there is no difference in the signal at -0.7 V on deoxygenating the sample solution. Hence it is clear that at these levels of determinand there is no need to deoxygenate the sample solutions. Fig. 8 shows signals obtained near the determination limit both with a newly polished electrode and a pre-treated electrode. The measurement potential used here was -0.65 V to reduce the oxygen blank to an acceptable level for this concentration of determinand. The large-scale removal of background noise on electrochemically pre-treating the elec- trode can be seen clearly. The extensive reduction in the signal from the oxygen dissolved in the sample solution on electro- chemical pre-treatment is also apparent; this blank is equiv- alent to 6 x 10-8 M nitrofurantoin. At significantly higher concentrations determinations would normally be made at -0.7 V where the blank signal is equivalent to 2 X M nitrofurantoin. At levels of nitrofurantoin above 5 X M coefficients of variation for five injections at the same concentration were typically less than 1%.Conclusions Electrochemical pre-treatment at high negative potentials in dilute sulphuric acid is effective in inhibiting the reduction of dissolved molecular oxygen at glassy carbon electrodes and therefore lowers the background currents caused by reduction of dissolved oxygen when such electrodes are used at negative potentials in flow injection analysis and, by extrapolation, in HPLC applications. With nitrofurantoin a slight loss of determinand signal also occurs, but without loss of precision.For most reversible systems it is expected that little or no loss of signal would be experienced, i.e., that the systems would remain reversible. It is further expected that detection limits even for compounds that are determined at low negative potentials, where oxygen is not a major interferent at high determinand concentrations, will be lowered. Further studies are being made on such systems. In flow injection applications using PTFE transmission tubing it is possible to deoxygenate eluents to a sufficiently low level to enable compounds that are reduced at potentials up to about -0.7 V to be determined at a pre-treated electrode without the need to deoxygenate the sample solutions. The authors thank Dr. J. P. Hart for his interest in this work.ANALYST, FEBRUARY 1986, VOL. 111 161 References 1. Blaedel, W. J., and Jenkins, R. A., Anal. Chem., 1975, 47, 1337. 2. Chan, H. K., and Fogg, A. G., Anal. Chim. Acta, 1979, 111, 281. 3. Van Rooijen, H. W., and Poppe, H., Anal. Chim. Acta, 1981, 130, 9. 4. Engstrom, R. C., Anal. Chem., 1982, 54, 2310. 5. Ravinchandran, K., and Ba1dwin;R. P., Anal. Chem., 1983, 55, 1782. 6. Engstrom, R. C., and Strasser, V. A,, Anal. Chem., 1984,56, 136. 7. Fogg, A. G., Fernfindez-Arciniega, M. A., and Alonso, R. M., Analyst, 1985, 110, 851. 8. Fogg, A. G., Fernindez-Arciniega, M. A., and Alonso, R. M., Analyst, 1985, 110, 1201. 9. 10. 11. 12. 13. 14. 15. 16. Ikenoya, S . , Abe, K., Tsuda, T., Yamano, Y., Hiroshima, O., Ohmae, M., and Kawabe, K., Chem. Pharm. Bull., 1979,27, 1237. Hart, J. P., Shearer, M. J., McCarthy, P. T., and Rahim, S., Analyst, 1984, 109,477. Lloyd, J . B. F., J . Chromatogr., 1983, 256, 323. Lloyd, J. B. F., J . Chromatogr., 1983, 257, 227. Lloyd, J. B. F., J . Chromatogr., 1983, 261, 391. Fogg, A. G., and Summan, A. M., Analyst, 1984, 109, 1029. “United States Pharmacopeia,” Twentieth Revision, United States Pharmacopeial Convention, Rockville, MD, 1980, p. 549. Fogg, A. G., Fayad, N. M., Burgess, C., and McGlynn, A., Anal. Chim. Acta, 1979, 108, 205. Paper A51282 Received August Ist, I985 Accepted August 21st, I985

ISSN:0003-2654    

DOI:10.1039/AN9861100157

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 163 Determination of Vitamin C by Flow Injection Analysis Fernando Lazaro, Angel Rios, M . D. Luque de Castro and Miguel Valcarcel* Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, Cordoba, Spain A photometric procedure for the determination of vitamin C using chloramine T by two flow injection analysis (FIA) methods is described. The first is in the presence of potassium iodide - starch solution using the normal FIA technique (linear determination range: 15-150 pg ml-1, RSD +_0.97O/0 and sampling frequency 90 samples h-1) and the second is in the presence of potassium bromide - methyl red solution using an FIA titration (linear determination range: 0.5-1000 pg mi-’, RSD +0.82°/0 and a sampling frequency of 30 samples h-l).The two methods are compared with each other and with a conventional titration procedure. Keywords: Vitamin C determination; flow injection analysis; potassium iodide - starch; potassium bromide - methyl red Numerous conventional methods have been described for the determination of vitamin C-with electroanalytical detection. most frequently used type of detection. Methods based on the use of Ti(II1) ,1 Ce(1V) ,2 hexacyanoferrate(II1) ,3 bromine cyanide,4 vanadium,5 iodate,6 2,4-dinitrophenylhydrazine7J and a7a’-dipyridyl9-l1 have been described, each with their advantages and disadvantages. Flow injection analysis (FIA) has been applied twice to the determination of vitamin C with electro analytical detection. The first contribution was reported by Strohl and Curran,12 who used a reticulated glassy carbon electrode to carry out the coulometric and amperometric determination of the vitamin, achieving detection limits of a few nanograms.The major advantage of the coulometric method is that the analyte is completely electrolysed and can be determined without calibration graphs. However, the sampling frequency is decreased with respect to the amperometric method. The second contribution was an amperometric method using immobilised enzyme (ascorbic acid oxidase) reactors that allowed the determination of the vitamin in brain tissue, with good results. 13 In order to develop a photometric - FIA method for the determination of vitamin C we used the reagent chloramine T in the presence of starch - potassium iodide solution or methyl red - potassium bromide solution as an indicator,’4 by normal FIA in the first instance and by an FIA titration without a gradient chamber (expansion-scale technique)l5-18 in the second.Both methods are compared with each other and with the conventional titration procedures. Experimental and Results Apparatus An FIA-5020 Tecator Analyzer, an SP6-500 Pye Unicam UV - visible spectrophotometer and a 178.12 QS Hellma flow cell (inner volume 18 1.11) were used. Reagents Chloramine Tsolution, 5 x 10-4 M. Starch - potassium iodide solution. Containing 5.25 g of soluble starch and 0.50 g of potassium iodide in 500 ml of distilled water. Methyl red - potassium bromide solution. Containing 0.5 ml of methyl red, 0.1% in ethanol, made up to 100 ml with 0.5% KBr solution. Ascorbic acid solutions.Prepared by direct weighing. All solutions were stable for at least 24 h. * To whom correspondence should be addressed. Determination of Vitamin C with Chloramine T and Starch - Potassium Iodide Solution Although this determination is commonly carried out by titration14 this is unsuitable for use with FIA when starch is employed as an indicator, due to the fleeting appearance of the blue colour of the starch - iodine complex in the reactor. A plateau is not obtained when a high sample volume is injected because the excess of chloramine T degrades the starch - iodine complex. It is therefore essential that the blue colour should have developed by the time that the sample passes through the detector. Manifold and chemical system The manifold used is shown in Fig.1. The design is very simple. The sample, dissolved in an acidic medium, is injected into the chloramine T stream after merging with a starch - iodide stream. Iodide forms HI in an acidic medium, which subsequently reacts with chloramine T: NaCl + I2 The iodine formed oxidises vitamin C: Once vitamin C has been fully consumed, iodine binds to starch, which consitutes the indicator reaction. Sulphuric acid is added directly to the sample as insertion into the chloramine T stream results in the precipitation of the reagent. When the sample only contains sulphuric acid (blank) no iodine is consumed because of the absence of vitamin C and thus the maximum signal is obtained. The presence of vitamin C causes a weakening of the analytical signal proportional to its concentration.Optimisation of variables The optimum values for the FIA variables (Fig. 1) are as follows: flow-rate (4) = 2.04 ml min-1; injected volume (Vi) = 96.2 p1; reactor length ( L ) = 55 cm; inner diameter of the reactor (@) = 0.5 mm. These values allow the blue colour of the starch - iodine complex to be formed as the sample plug passes through the detector, where it is monitored at 650 nm. The optimisation of the variables ensures the maximum difference between the reference and the sample peaks. The influence of the chloramine T concentration is one of the most important variables in this determination. Its effect can be observed in Fig. 2. For high concentrations of chloramine T the presence of vitamin C has no effect on the164 I Starch - K I Sample (H+) Chlorarnine T i ANALYST, FEBRUARY 1986, VOL.111 650 nrn Photometer Methyl red - KBr Sample (H+) Chlorarnine T I Blank - 1 n q = 1.19 rnl rnin-1 1 W 1 Time - I Sample I Time - Fig. 1. (for details see, text) Manifolds used in the determination of vitamin C and recordings obtained by: (a) normal FIA mode and ( b ) FIA titration analytical signal. The maximum difference between the blank and the sample signal is obtained with 5 x 10-4 M chloramine T solutions. For lower concentrations of the amine the difference in absorbance decreases owing to the drastic weakening of the blank signal. The pH of the sample is adjusted by the addition of H2S04. The reaction medium must be acidic to allow the formation of HI. The difference in absorbance between the blank and the sample is high at low pH, but above a sulphuric acid concentration of 0.75 M the increase in this difference is small and some peak broadening is observed. A 0.90 M sulphuric acid concentration was chosen as the optimum.The most suitable concentrations of KI and starch are indicated under Experimental. It is well known that an increase in temperature is a negative factor for the development of the blue colour of the starch - iodine complex; therefore, the experiments were performed at room temperature. Determination of vitamin C This method allows the determination of vitamin C according to the equation: AA = 0.0035 [vitamin C]-O.O053; r = 0.999 where AA is the difference in the absorbance between the blank and the sample containing vitamin C (whose concentra- tion is expressed in pg ml-1).The equation is satisfied between 15 and 150 pg ml-1, the RSD (P = 0.05) for the determination of 100 pg ml-1 of vitamin C being &0.97%. The sampling frequency is 90 samples h-1. Determination of Vitamin C with Chloramine T and Methyl Red - Potassium Bromide Solution This method, based on the measurement of the peak width at a pre-determined height from the base line, has been termed “high-speed titrations” by Ramsing et al. 15 and “scale- expansion techniques” by Stewart and Rosenfeldl6 following the suggestions from Pardue and Fields.17?1* The chemical system is similar to the one described above but the indicator is methyl red - potassium bromide solution. I, B 1.200 al f (0 2 0.800 v) 2 0.400 0.1 1 .o 5.0 [Chlorarnine TI x 1 0 - 2 / ~ Fig.2. Influence of the chloramine T concentration in the stream. Absorbance values obtained with: A, sample with 5 X M vitamin C; B, blank; and C , difference between sample and blank This system may be used for FIA titrations because the red colour of methyl red in an acidic medium allows a wide peak to be obtained when a large volume of this indicator is injected. Manifold and chemical system The manifold used is similar to the one described above. In this instance the methyl red - potassium bromide stream is substituted for the starch - potassium iodide one. The HBr formed by merging this stream with that of the sample in an acidic medium (H2S04) is oxidised by chloramine T and the Br2 produced oxidises vitamin C to dehydroascorbic acid.An excess of bromine with respect to the stoicheiometric amount of vitamin C present reacts with methyl red, causing its degradation. The presence of the vitamin in the sample (resulting in a decrease in the degradation of the indicator) brings about a widening of the peak obtained at 535 nm. The measurement of the peak width (in seconds) at a pre- determined absorbance value is the basis of the FIA titrations.ANALYST, FEBRUARY 1986, VOL. 111 165 Table 1. Calibration graphs for the determination of vitamin C by FIA titration Absorbance * Range/M Equation 0.020 (0.3-1.7) x At = 8.248 log At = 18.17 log 0.050 (0.3-1.7) x 10-5 At = 4.67108 At = 15.18 log At = 20.60 log 0.100 (0.3-1.7) x 10-5 At= 4.46108 At = 15.88 log (0.1-2.8) x 10-3 (0.2-1.1) x 10-4 (0.2-5.7) x 10-3 (0.02-5.7) X 10-3 *Absorbance at which the peak width (At) is measured vitamin C] + 49.63 vitamin C] + 102.94 vitamin C] + 28.55 vitamin C] + 79.08 vitamin C] + 100.38 vitamin C] + 28.01 vitamin C] + 80.86 r 0.991 0.991 0.993 0.997 0.991 0.990 0.999 Table 2. Determination of vitamin C in synthetic samples [Vitamin C] by normal FIA*/M [Vitamin C] by FIA titration ?/M Added Found Error, YO 9.02 x 10-5 9.66 x 10-5 +6.8 1.70 x 10-4 1.77 x 10-4 +4.1 2.84 x 10-4 2.67 x 10-4 -5.9 5.66 x 10-4 5.56 x 10-4 -1.7 1.14 x 10-3 1.17 x 10-3 +2.6 1.42 x 10-3 1.43 x 10-3 -1.4 * Equation: AA = 615 [vitamin C] - 0.005 t Equation At = 15.882 log [vitamin C] + 80.86 Added Found Error, YO 2.08 x 10-5 2.17 x 10-5 +4.3 5.68 x 10-5 5.82 x +2.5 1.70 x 10-4 1.71 x 10-4 +0.6 4.26 x 10-4 4.23 x 10-4 -0.7 8.52 x 10-4 8.63 x 10-4 +1.3 1.42 x 10-3 1.39 x 10-3 -2.1 "t i 60 v) L Q 40 2o t "\ " Absorbance Fig.3. Variation of At with the value of the absorbance at which the measurement of the peak width is performed. A, Sample with 5 X 10-3 M vitamin C; B, blank; C, difference between sample and blank Optimisation of variables The optimum values for the FIA variables are as follows: q = 1.19 ml min-1; Vi = 923.0 pl; L = 70 cm and @ = 0.5 mm. The large sample volume injected ensures that wide peaks are obtained, the lower the vitamin C concentra- tion, the wider the peaks. This sample volume, which is unusually large for normal FIA, is the key to FIA titrations. The concentrations of some reagents are decisive.Thus, the optimum concentration of chloramine T is 5 X M, which allows maximum differences to be obtained between the sample and the blank. The samples contain 1% of concen- trated H2S04 (0.15 M). Owing to the short reactor length, the temperature does not exert a significant influence. Thus, the experiments were performed at room temperature (18-20 "C). The absorbance at which the peak width is measured influences the sensitivity of the determination. It is advisable to measure the peak width at the absorbance yielding the maximum difference between the sample and the blank. Fig. 3 shows this effect for a sample containing 100 yg ml-1 of vitamin C. The differences increase as the absorbance approaches zero, but owing to instrumental limitations, peak widths over 99.9 s cannot be measured.The increase in peak width observed above 0.100 A (curve C, Fig. 3) is not significant, and as the corresponding peak widths are very small, the errors introduced in these measurements are large. Therefore, the criteria adopted to run the calibration graphs at different absorbances according to the high or low concentra- tion of vitamin C in the samples are as follows: the peak width is measured at 0.100 A for high and at 0.020 or 0.050 for low analyte concentrations. Calibration graphs A plot of the peak width (At) versus logarithm of the unknown concentration, characteristic of these titrations, shows differ- ent linear ranges according to the value of the absorbance at which At is measured (as is shown in Table 1). In each instance the contribution of the blank has been subtracted.This represents a major improvement of the method proposed above; a wider range and lower limit of detection are attained, possibly owing to the instability of the starch - iodine complex in the presence of chloramine T, which was a limiting factor in the previous method. An RSD ( P = 0.05) of +0.82% and a sampling frequency of 30 h-1 were obtained in the determina- tion of 100 pg ml-1 of vitamin C (At measured at 0.050 A). Study of Interferents in Each Method In a study of the interferences in both FIA methods, Fe2+, Ca2+, Mg2+ , NH4+, P043-, C032-, oxalate, glucose, glycine, hystidine, urea, cysteine and uric acid can all be tolerated at a ten-fold excess over vitamin C, except for cysteine and uric acid, which interfere when present at the same concentration as the vitamin in the FIA titration method.It is worth noting the smaller influence of foreign species in the FIA technique compared with conventional methods. 14 This behaviour was observed in earlier studies19 and is attributable to the non-equilibrium state of the chemical system at the point of detection. The higher selectivity of the method employing iodine is a result of the lower oxidising strength of this halogen166 ANALYST, FEBRUARY 1986, VOL. 111 compared with bromine. Likewise, the higher acidity of the sample in the first method facilitates the precipitation and removal of interferents (e.g. , uric acid). Comparison of the Results Obtained by the Two Methods The results obtained by both methods for the determination of vitamin C in several synthetic samples are given in Table 2.A similar degree of accuracy is afforded by both methods. The determination of vitamin C with chloramine T by FIA titration offers the following advantages over the normal FIA technique: a wider determination range; a higher sensitivity; and a slightly higher accuracy. These advantages stem from the high instability of the starch - iodine complex in the presence of chloramine T, which is a limiting factor in the determination of vitamin C by the normal FIA technique. Normal FIA does however have some advantages, such as the need for only one calibration graph, smaller sample volume, higher selectivity, which makes it applicable to the determination of vitamin C in urine, and higher sampling rate.The determination of vitamin C by FIA titration presents the following advantages over the manual procedurel4: it is faster and simpler, as the colour change in the manual method is slow and difficult to detect visually; it uses reagent and sample more sparingly; the manipulation and intervention of the operator is a minimum; and it is more sensitive and accurate . References 1. Gupta, D., Sharma, P. D., and Gupta, Y. K., Talanta, 1975, 22, 913. 2. Rao, G. G., and Sastri, G. S . , Anal. Chim. A m , 1971,56,325, 3. Sastri, G. S., and Rao, G. G. Talanta, 1972, 19, 212. 4. Paul, R. C., Chauhan, R. K., and Prakash, J . , Indian J . Chem., 1971, 9, 879. 5. Eremina, Z. I., and Gurevich, V. G., Zh. Anal. Khim., 1964, 19, 519. 6. Murty, C. N., and Bapat, N. G., Fresenius Z. Anal. Chem., 1963, 199, 367. 7. Roe, J. H., Oesterling, M. B., and Llamron, C. M., J. Biol, Chem., 1948, 174,201. 8. Teruuchi, J., and Kitazato, K . , Arch. Exp. Med., 1951, 1, 55. 9. Zannoni, U., Lynch, M., Goldstein, S., and Sato, P., Biochem. Med., 1974, 11, 41. 10. Okamura, M., Clin. Chim. Acta, 1980, 108, 259. 11. Okamura, M., Bitamin, 1981, 55, 495. 12. Strohl, A. N., and Curran, D. J., Anal. Chem., 1979,51, 1045. 13. Bradberry, C. W., and Adams, R. N., Anal. Chem., 1983,55, 2439. 14. Verma, K. K., and Gulati, A. K.,Anal. Chem., 1980,52,2336. 15. Ramsing, A., RfiiiCka, J., and Hansen, E. H., Anal. Chim. Acfu, 1981, 129, 1. 16. Stewart, K. K., and Rosenfeld, A. G., Anal. Chem., 1982,54, 2368. 17. Pardue, H. L., andFields, B., Anal. Chim. Acta, 1981,124,39. 18. Pardue, H. L., and Fields, B.,Anal. Chim. Acta, 1981,124,65. 19. Linares, P., Luque de Castro, M. D., and ValcArcel, M., Anal. Paper A5134 Received January 29th 1985 Accepted August 28th, 1985 - Left., 1985, 18 B1, 67.

ISSN:0003-2654    

DOI:10.1039/AN9861100163

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1986, VOL. 111 167 Determination of Vitamin C in Urine by Flow Injection Analysis* Fernando Lazaro, Angel Rios, M. D. Luque de Castro and Miguel Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, Cordoba, Spain A spectrophotometric method with flow injection analysis is described for the determination of vitamin C in urine, which presents substantial improvements over the existing conventional manual and automatic methods. Determinations can be carried out at the pg ml-1 level; the method has an average recovery error of +2.5% and a sampling frequency of 90 samples h-1 can be attained. Keywords: Vitamin C determination; flow injection analysis; urine The determination of vitamin C in its three forms (ascorbic acid, AA; dehydroascorbic acid, DHAA; and 2,3-diketogulonic acid, DKGA) has been carried out in numerous ways.l The most frequent method is based on the titration of ascorbic acid with 2,6-dichlorophenolindophenol, 2,6-DCPIP2J; nevertheless, this method is only useful for samples with small amounts of foreign species as it is subject to numerous interferences from reducing species such as sulphur dioxide, tannins, cysteine , sulphydryl compounds, certain metal ions, plant pigments, reductants and similar compounds.For this reason, depending on the sample matrix, this method and others that eliminate possible interferents have been used so far. For the determination of vitamin C in urine and blood the method most commonly used is that of Roe and K ~ e t h e r , ~ which is based on the formation of coloured osazones in a concentrated sulphuric acid medium, via the reaction between 2,4-dinitrophenylhydrazine and 2,3-diketogulonic acid, an oxidised form of ascorbic acid.Nevertheless, in spite of overcoming most interferences this procedure does have several disadvantages: it is complex and time consuming (over 180 min per analysis are required); and it is subject to other interferents such as hexoses, pentoses, glucuronic acid, reductants, hystidine and several other amino acids, although most of them do not interfere at the normal level in which they are normally found in urine and blood. To improve this method several modifications have been suggested: the use of different oxidising agents such as b r ~ m i n e , ~ 2,6-DCPIP6 and benzoquinone7; different reaction conditionss; the use of different acids6; or the use of different separative techniques.5 Automatic methods using reagents such as 2,6-DCPIP,9 o-phenylenediamine (DPD)10 and 2,4-dinitrophenol (2,4-DNP)llJ2 have also been developed with the aid of Technicon AutoAnalyzers.These methods are the most similar to that proposed in this paper, in which the chloramine T - KI - starch reaction and flow injection analysis (FIA) are used jointly. This latter development is an easily automatable technique,l3 which, in addition, is inexpensive. These facets, together with the absence of interferents in the method, make this a good alternative to routine vitamin C analysis. Experimental All reagents and apparatus used in this work were the same as described previously,l4 except for a Metrohm Dosimat E535 automatic burette.The FIA manifold employed has also been described previously.14 Procedure Urine samples were obtained from individuals who had been given a pharmaceutical compound containing 0.5 g of vitamin * To whom correspondence should be addressed. C. Samples were collected in a polyethylene flask to which a final concentration of 500 mg 1-1 of oxalic acid had been added. The urine samples were subsequently diluted 1 + 5 with 0.9 M H2SO4 and pumped to their confluence with a KI - starch stream until they filled the loop of the injection valve, which was inserted into the chloramine T stream. The concentrations of vitamin C in urine shown in Tables 1 and 3 were obtained with diluted samples.Results and Discussion Firstly the two methods proposed in reference 14 were applied to the determination of vitamin C in urine; however, the FIA titration method (in which KBr - methyl red solution is used as the indicator) is subject to major interferences owing to its low selectivity. Urine samples with a concentration of vitamin C of 60 pg ml-1, as determined by the normal FIA method, provided concentrations over 6000 pg ml-1 with the FIA titration technique. This result was predictable as sulphydryl compounds interfere at the same analyte level because of the strong oxidising character of Br2, a product yielded in the indicator reaction (chloramine T - KBr) which acts as an oxidant for vitamin C. Conversely, the normal FIA method, which uses KI - starch solution as the indicator, provided good results and was therefore chosen for the application of the method to real samples.Stability of the Samples The study was performed on urine samples diluted 1 + 5 with 0.9 M &SO4. Oxalic acid, a reducing agent, was used as a preservative.15 Concentrations between 0 and 2000 pg ml-l were used in this study. For concentrations of oxalic acid equal to or higher than 500 pg ml-1, the analytical signal yielded by the sample remained constant for at least 24 h, whilst in the absence of a preservative a decrease in the signal of 10% was observed during the same time interval. A concentration of 500 pg ml-1 was therefore chosen for subsequent experiments. The reproducibility of the method was studied at two analyte concentrations (30.30 and 65.70 pg ml-1) with nine different samples in each instance and at two different times (1.0 and 3.5 h). The values obtained for the relative standard deviation (r.s.d.) were as follows. After 1 h: 30.30 pg ml-l (r.s.d.= +0.93%); 65.70 pg ml-1 (r.s.d. = +0.32%). After 3.5 h: 30.30 vg ml-1 (r.s.d. = +0.900/,); 65.70 pg ml-1 (r.s.d. = +0.30%). The precision was good in both instances but was improved by increasing the analyte concentration. For the study of the recovery of vitamin C in urine, six samples were taken at different times from an individual who had taken 0.5 g of vitamin C . Concentrations of 20,40 and 80 pg ml-l of vitamin C were added to each sample. The results obtained are listed in Table 1; a good recovery was observed (the average recovery = 99.6% with an average error of 2.24%). These data indicate the absence of interferents in this168 ANALYST, FEBRUARY 1986, VOL.111 Precision 1.3 , i Table 1. Results for the recovery of vitamin C in urine 10 Vitamin C addedlpg ml-1 Vitamin C found in urine 20.00 40.00 80.00 Urine diluted 1 + 5/ sample pg ml-1 Found/pg ml-1 Recovery, Yo Found/pg ml-1 Recovery, YO Found/pg ml-1 Recovery, YO 1 20.97 40.06 95.5 60.00 97.6 102.00 102.0 2 29.03 49.68 103.2 69.68 101.6 109.02 100.0 3 36.13 56.13 100.0 74.84 96.8 115 .OO 98.6 4 39.35 59.68 101.6 78.71 98.4 119.50 100.2 5 41.61 61.95 101.6 80.65 97.6 119.50 97.4 6 59.35 78.71 96.8 102.00 106.6 137.50 97.7 Table 2. Comparison between the FIA and conventional titration methods for the determination in vitamin C in synthetic samples FIA method Titration method Error of FIA Amount of method relative vitamin C Amount Relative Amount Relative to titration addedlpg ml-1 found/pg ml-1 error, YO found/yg ml-1 error, '/O method, '/O 20.00 40.00 50.00 60.00 70.00 80.00 100.00 120.00 140.00 Mean error, YO 19.04 43.79 49.56 56.35 68.64 79.78 106.34 120.53 135.98 -4.8 9.5 -0.9 -6.1 -1.9 -0.3 6.3 0.4 -2.9 3.7 21.54 41.57 50.73 62.70 73.97 85.25 106.38 126.11 145.13 5.7 3.9 1.5 4.5 5.7 6.5 6.4 5.1 3.6 4.8 -11.6 5.3 -2.3 - 10.1 -7.2 -6.4 0.0 -4.4 -6.3 4.6 Table 3.Comparison between the FIA and conventional titration methods for the determination of vitamin C in urine samples (diluted 1 + 5 ) and containing 9.26 pg ml-1 of vitamin C Amount of vitamin C addedlyg ml - 1 10.00 30.00 40.00 50.00 60.00 70.00 90.00 110.00 130.00 Mean error, Yo FIA method Titration method Amount found/pg ml- 21.07 39.26 47.85 59.88 71.01 78.60 102.69 119.43 136.29 Relative error, Yo 9.4 0.0 -2.9 1.0 2.5 -0.8 3.5 0.1 -2.2 2.5 Amount found/pg ml-1 21.84 41.57 51.43 61.29 69.57 79.61 95.81 114.34 134.56 Relative error, Yo 13.4 5.9 4.4 3.4 0.4 0.4 -3.5 -3.7 -3.4 4.3 Error of FIA method relative to titration method, YO -3.5 -5.5 -6.7 -2.3 2.1 -1.3 7.2 4.0 1.3 3.8 method, which is attributable to the weak oxidising character of the I2 formed in the reaction.Comparison of Methods The FIA method used in this work was compared with the conventional titration technique16 (Fig. 1) for synthetic samples and urine samples. In addition, both methods were compared with the determination of vitamin C by direct weighing.Analysis of synthetic samples On analysing nine different samples by the two methods average errors of 3.7 and 4.8% were obtained by the FIA and conventional titration methods, respectively, both rela- tive to direct weighing (Table 2). In the conventional titration procedure the errors were always positive, possibly owing to the difficulty in determining the end-point, as it is not based on a colour change, but in the persistence of the colour. Moreover, the kinetics of the reaction are very slow in the vicinity of the equivalence point, so that it is necessary to wait for a certain time after each addition of titrant before making the measurements. This shortcoming is not present in the FIA I Sensitivity - 600 Decreased sample volume Rapidity Decreased IT1 consumption reagent 1 1 Fig.1. Improvements of the suggested procedure over the conven- tional titration procedure method, which affords a sampling frequency of 90 samples h-1 compared with 6 samples h-1 in the conventional titration. In addition, the titration technique requires the use of anANALYST, FEBRUARY 1986, VOL. 111 169 automatic burette (reading to 0.01 ml) owing to the small reagent volume needed. Analysis of urine samples The same study performed on urine samples with final concentrations of 9.26 pg ml-1 after a 1 + 5 dilution (Table 3) yielded average errors of 2.7 and 4.3% for the FIA and conventional titration methods, respectively, both relative to direct weighing. These errors are slightly smaller than those obtained for synthetic samples.The average error of the FIA method relative to the conventional method is 3.8% in this instance. These results reveal a smaller error with the proposed method compared with the conventional method, for both synthetic and urine samples. When the proposed method is compared with automatic methods such as that of Pelletier and Brassard12 it can be concluded that the proposed method has several advantages over the latter: although the latter technique offers higher sensitivity and similar precision and recovery, the working scheme is very complex (13 channels are necessary as opposed to 3 in the FIA method); maintenance is expensive and time consuming (it requires daily washing with 75 ml of HN03 and distilled water for 40 min); the flow cell must be washed weekly with dichromate solution and the tubing system must be changed after 120 determinations, whilst the FIA manifold only requires washing for 5 min with distilled water after a working day and allows up to 10000 determinations to be performed (90 determinations h-1, 8 h a day, for 14 d) before changing the tubing system.In addition, the maximum sampling rate achievable is 13 samples h-1 (with the use of a sampler) compared with 90 samples h-1 for the suggested method. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Cooke, J. R., and Moxon, R. E. D., in Counsell, J. N., and Hornig, D. H., Editors, “Vitamin C,” Applied Science, Barking, 1981, pp. 167-198. Harris, L. J., and Ray, S. N., Biochem. J . , 1933, 27, 303. Birch, T. W., Biochem. J . , 1933, 27,590. Roe, J. H., and Kuether, C. A., J. Biol. Chem., 1943,147,399. Zloch, Z., and Ginter, E., 2. Klin. Chem., 1970, 8, 302. Pelletier, O., J. Lab. Clin. Med., 1968, 72, 674. Saari, J. C., Anal. Biochem., 1966, 15, 537. Roe, J. H., J. Biol. Chem., 1961, 236, 1611. Egberg, D. C., J. Sci. Food Agric., 1973, 24, 789. Dunmire, D. L., J. Assoc. Off. Anal. Chem., 1979, 62, 64F. Aeschbacker, H. U., and Brown, R. G., Clin. Chem., 1972,18, 965. Pelletier, O., and Brassard, R., Adv. Autom. Anal. Technicon Znt. Congr., 1973, 9, 73. VBlcarcel, M., and Luque de Castro, M. D., “Flow Injection Analysis: Principles and Applications,” Ellis Horwood, Chichester, in the press. LBzaro, F., Rios, A., Luque de Castro, M. D., and VBlcarcel, M., Analyst, 1986, 111, 163. Haddad, P., J. Chem. Educ., 1977, 54, 192. Verma, K. K., and Gulati, A. K., Anal. Chem., 1980,55, 1045. Paper A51244 Received July 8th, 1984 Accepted August 28th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100167

出版商:RSC    

年代:1986

数据来源: RSC