摘要:

iAnalytical Chemistry Journalsfrom The Royal Society of ChemistryJournal ofAnalyticalAtomicSpectrometryUAAS)The Analyst AnalyticalAbstractsAn International Journal onthe Development andApplication of AtomicSpectrometric TechniquesJournal of Analytical Atomic Spectrometry(JM) 1s a new international journal,launched in February 1986, which containsoriginal research papers, short papers,communications and letters concernedwth the development and analyticalapplication ot atomic spectrometrictechniques. JAAS is published bunonthlyand includes comprehensive reviews onspeclfic topics, general information andnews of interest to analytical atomicspectroscoplsts, including information onforthcoming conferences and bookreviews. Special lssues o f j .will bepublished, devoted to subjects highlightedby particular symposia. Also included inJAAS are the literature reviews previouslycovered in Annual Reports on AnalyticalAtomic Spectroscopy.JAAS has a style and format similar to thatof the well-established journal The Analyst.JAAS provides an improved publicationservice to support the growing researchefforts in, and applications of, atomicspectrometric techniques.Six issues per m u m (plus index)plus 2 special issues 1987An international journal of h g h reputecontaning original research papers on thetheory and practice of all aspects ofanalytical chemlstry drawn from a widerange of sources. It also publlshes regularcritical reviews of important techniquesand their applications, short papers andurgent communications (which arepublished in 5-8 weeks) on important newwork, and book reviews.12 lssuesper annum; annual author andsubject index.AnalyticalProceedingsAnalytical Proceedings is th news andinformation journal of the AnalyticalDimion.It contains special articles,reports of meetings, extended summariesor original papers, safety articles, details ofrecent legislation, surveys of eq-upment,and many other items of general interest toanalytical chemists both in Britain andoverseas.12 m e s per annum (plus index).Analytical Abstracts endeavours to coverthe whole field of analytical chemlstry,providing more than 12,000 abstracts perannum of papers and books considered tobe of unportance and interest to analyticalchemlsts.A subject mdex 1s provlded ineach lssue and a volume index 1s producedannually.12 m e s per annum (plus index).Analytical Abstracts IS now available onllne-further lnformation obtalnable fromThe Royal Society of ChernsrrY, Sales &Promotion Dept , The Uruversity,Nottmgham NG7 2RDSubscription Rates 1987journal of Analytical AtomicSpectrometry (IAAS)€180.00 ($356.00) Rest of World €202.00RSC Members €36.00The Analyst€160.00 ($315.00) Rest of World €129.00RSC Members €32.00Analytical Abstracts€239.00 ($463.00) Rest of World €263.00RSC Members €62.50Analytical ProceedingsE25.00 ($148.00) Rest of World €84.00RSC Members €11.50The Analyst, Analytical Abstractsand Analytical Proceedings€411.00 ($601.00) Rest of World €455.00The Analyst and Analytical Abstracts€364.00 ($209.00) Rest of World €403.00The Analyst and AnalyticalProceedings€200.00 ($394.00) Rest of World €224.00N.B.The version of Analytical Abstractsprinted on one side of the page only 1s nolonger available.Ordering:RSC Members should send their orders to:Membership Manager, The Royal Society of Chemistry, 30 Russell Square, LondonWC 1 B 5DT.Non-RSC Members should send their orders to:The Royal Society of Chemistry, Distribution Centre, Blackhorse Road, Letchworth,Herts SG6 lHN, UK.lnformationServicesCircle 003 for further informatio11SEEYOUAT“ THEWEEK”!CUT OUTAND MAIL FOR YOUR FREE TICKET!To: Brian Curtis, Organiser, British Laboratory Week,The Hub, Emson Close,Saffron Walden, Essex CBlO lHL, United Kingdom.Please send meprincipal area of interest:Laboratory ’87 0 Medical Laboratory Sciences Bio ’87 0Corn puter-Aided-Sciences 0FREE ticket(s) to “The Week” and information on myTitleCompany/OrganisationAddressAlso, please send me details of the following conferences,Bio ’87 Analyticon Computer-Aided-Sciences 0The British Laboratory Week is the~- major international meeting of itskind in Europe in 1987.There’s no Achema, Analytica,Het Instrument or Salon duLaboratoire this year:They’rebi-ennial and tri-ennial, and theirschedules coincide in ’87.So plan your visit now!As you’ll appreciate, ‘The M e kis an important occasion and, wellworth your time and interest. Simplycomplete and clip the couponbelovl! We’ll make sure all therelevant information is sent to youimmediately it becomes available.The British Laboratory Weekis an umbrella title for FOURexhi bitions and THREEconferences. (See panels above.)Over 750 different companies’products from around the worldwill be on displayHeck! With the exhibits andthe subjects for discussion we’vegot lined up, you’d be crazy to stayawayTHE BRITISHLABORATORY WEEK.THE COUNTRY’S LARGESTSClENTlFlC EVENT13-15 OCTOBER, 1987OLYMPIA, LONDON, ENGLANDThe British Laboratory Week is ,6~\pK\19organised by Curtis Steadman & wotayPartners limited, The Hub,Emson Close, Saffron Walden,Essex CBlO 1HL.UnitedKingdom.Telephone: (0799) 26699.Telex: 81653 INFORM G.Fax: (0799) 26088.121x1Circle 002 for further informatioX Conference on Analytical Atomic Spectrometry (CANES)VII Polish Spectroanalytical Conference5-9 September 1988, Torun, PolandThis conference, to be held at the Nicholas Copernicus University in Torun, Poland, will cover the followingbranches of spectroscopy :Physical aspects of analytical atomic spectrometryOptical emission spectrometry and excitation sourcesAtomic absorption and fluorescence spectrometryX-ray spectrometryInorganic mass spectrometryElectron and ion spectrometryInstrumental neutron activation analysisLasers in analytical atomic spectrometryApplication of spectroscopy in trace analysisSpeciation analysisSurface analysisFurther information can be obtained from Dr.J. Fijatkowski, Institute of Nuclear Chemistry and Technology,ul. Dorodna 16, 03-195 Warsaw, Poland.ElectroFinnAnalysisInternational Conference on Electroanalytical Chemistry6 9 June 1988, Turku (Abo), FinlandThe scientific programme will consist of invited plenary and keynote lectures, submitted research papers andposters, divided into the following sessions:InstrumentationIndustrial ApplicationsPharmaceutical ApplicationsClinical ApplicationsElectrochemical SensorsElectrochemical Flow AnalysisSpeakers will include L. R. Faulkner (USA), J. J. Kankare (Finland), E. Pungor (Hungary), J. Asplund (Sweden),P. M. Bersier (Switzerland), W. F. Smyth (Northern Ireland), J. Reust (Switzerland), W.Simon (Switzerland), A.Lewenstam (Finland), J. Janata (USA), I. Lundstrom (Sweden), E. Hansen (Denmark) and M. Trojanowicz(Poland).There will also be an exhibition of electroanalytical instruments and a comprehensive social programme.Call for Papers. A 250-word abstract should be submitted before 31 December 1987 to the address below.Further information and registration forms can be obtained from Dr. Ari Ivaska, Laboratory of AnalyticalChemistry, Abo Akademi, SF-20500 Turku (Abo), Finland....11IIIIIIIIIIIIr-THE ANALYST READER ENQUIRY SERVICE M AY‘87For 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 stamp.PLEASE USE BLOCK CAPITALS LEAVING A SPACE BETWEEN WORDS Valid 12 months1 NAME2 COMPANY i l l l l l i l l l l i l l 1 1 I 1 H I I I I 1 1 1 1 1 1 1 1 1t I I I I l I I I I I I I i I l l 1 1 1 1 1 I i I I I l 1 ~l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l l l l I I I l l I 1 H Ii l l i l l l l l l l l l l l i l l i l l l I l l l l l l l I I iI l l I l l l I l l l l i l l l l l l i l I l l l l l l C ~ ~ TPLEASE GIVE YOUR BUSINESS ADDRESS IF POSSIBLE.IF NOT, PLEASE TICK HERE 03 STREET4 TOWN5 COUNTYPOST CODE6 COUNTRY7 DEPARTMENT/DIVISION i l l l l l l l l l i l I I I I I I 1 1 I I i 1 1 I I 1 1 1 1 1 1POSITION8 YOUR JOB TITLE/9 TELEPHONE NOOFFICE USE ONLY HEC D PROC 0 LA- 1FOLD HERE1I IIIIIIIIIIIIIIIIIIII1IIIIIIIIIIIIIIIIIIIIIIIIIIPostagewill 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 6WFEngland21 IIIIIIIIIIIIIIIIIIIIIIIia xv)-InUI mx mPIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII!IIIIIIIIIIIIIIIIIIIIIIIIIII1IIIIIII

ISSN:0003-2654    

DOI:10.1039/AN98712BP021

出版商:RSC    

年代:1987

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98712FX025

出版商:RSC    

年代:1987

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98712BX027

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 933 Analytical Chemistry of Synthetic Food Antioxidants A Review Kevin Robards” School of Applied Science, Riverina-Murra y Institute of Higher Education, P. 0. Box 588, Wagga Wagga, New South Wales 2650, Australia and Sergio Dilli Department of Analytical Chemistry, University of New South Wales, P.O. Box 1, Kensington, New South Wales 2033, Australia Summary of Contents introduction Recovery procedures for antioxidants Qua ntif icat ion of ant ioxida n ts Colorimetric methods Spectrophotometric methods in the ultraviolet region Paper and thin-layer chromatographic methods Gas chromatographic methods Methods based on high-performance liquid chromatography Miscellaneous techniques Keywords: Review; analysis; antioxidants; foods; fats and oils Free antioxidants Derivatives Introduction Antioxidant technology dates from the early 1940s when studies with the alkyl esters of gallic acid were shown1 to have considerable potential as antioxidants.Later, butylated hydroxyanisole (BHA) was shown to have outstanding antioxidant properties. Combinations of these antioxidants dominated the scene until 1954 when butylated hydroxy- toluene (BHT) was cleared for use in food in the USA. Other synthetic antioxidants include tert-butylhydroquinone (TBHQ) ,2,3 first approved for use in foods in the USA in 1972, and nordihydroguaiaretic acid (NDGA) .4 Of course, many natural substances, including tocopherols5 and various plant extracts,6 exhibit antioxidant activity and several spices are known for their excellent antioxidant properties.In this category, ascorbyl palmitate is often referred to as a natural antioxidant although the ester does not occur naturally. Moreover, it has been claimed6 that ascorbyl palmitate is not a true antioxidant but only exhibits such activity by virtue of its sequestering ability. However, because of synergism,6 it is usually used in combination with tocopherols. In fact, it is a common occurrence in antioxidant technology nowadays to use a more effective combination of two or more antioxidants rather than the equivalent amount of a single compound. As most countries limit the types and amounts of anti- oxidants allowed in a particular food, their quantitative determination is an important consideration. Unfortunately, this poses a number of problems centred on the very low concentration of antioxidants which may be present in a varied and complex matrix.One such problem is that phenolic antioxidants may complex with the naturally occurring fats such as phospholipids,7 although no information is available as to whether these compounds still function as a source of antioxidant. Another problem is that plastics and wrapping materials frequently contain antioxidants capable of migrating into the enclosed fo0d.8~9 Obviously, the concern for “total” * To whom correspondence should be addressed. or “free” antioxidant is therefore an important question. Owing to the lack of adequate specificity in the quantification techniques, separation of an antioxidant is a necessary prerequisite to a comprehensive scheme of analysis.To date, the general approach has involved separation from the sample matrix, usually with pre-concentration, followed by quanti- fication. The literature review which follows is limited to the analytical chemistry applied to the five most common food antioxidants, namely, BHA, BHT, NDGA, propyl gallate (PG)lO and TBHQ (see Table l), and covers the period up to June 1986. No attempt has been made to include other compounds, such as octyl gallate, lauryl gallate and ethoxy- quin, despite their use in some countries. The approach adopted has been to divide the review into two parts dealing with (i) antioxidant recovery procedures and (ii) quantifica- tion procedures. The latter section, as an historical account, contains some dated materials and practices that would either be unavailable or unacceptable today.Recovery Procedures for Antioxidants Because of the dynamic conditions associated with their action, a satisfactory recovery procedure for antioxidants must cover a wide concentration range extending from normal usage6 to the trace level. Recovery is further complicated by the need to consider the antioxidant decomposition products and the bound antioxidant in addition to the free antioxidant. The difficulty with the recovery of free antioxidants is derived from their vastly different properties. Hence, for example, PG is considerably more polar than BHA, which in turn is more polar than BHT which is lipoidal in character.11 Despite these differences, solvent extractionl2-16 has been widely used for the isolation of antioxidants, with or without subsequent clean-up of the extract.Obviously, when dealing with compounds of such diversity in so complex and varied a matrix as foods, a universal approach to recovery is largely impractical. One problem with solvent extraction is that considerable amounts of the matrix can be co-extracted with934 ANALYST, JULY 1987, VOL. 112 ~ ~~~ Table 1. Selected properties of synthetic antioxidants used in foods. The data are from reference 10a TBHQ NDGA Property BHA* BHT PGt CH3 CH3 -QH Structure . . . @C(CH& ( C H 3 ) & , @ X 3 - M 3 H o y f o H H o p c H z - c H - I I CH- CHZ HO OH / / OH OCH3 CH3 COOCBH, Melting-point/"C 48-55 69-70 146-148 126 5128.5 Synergism . . With gallates With BHA but not With BHA - and BHT with gallates Solubilitieslg per 100 ml- Water .. Negligible Negligible 0.35 1 Typical fat or oil 30-40 20-30 1 5-10 RFS- Solvent1 . . 0.48 0.79 0.09 0.30 Solvent2 . . 0.50 0.75 0.32 0.45 184-185 - Negligible 1 0.06 0.34 * Commercial preparations also contain some 2-tert-butyl-4-hydroxyanisole. t Octyl and dodecyl gallate are used less commonly. $ Using silica gel plates and light petroleum - benzene - acetic acid (40 + 40 + 20) (solvent 1) or hexane - acetone - acetic acid (55 + 40 + 15) (solvent 2) as mobile phases. Data taken from reference 14. the aotioxidant and its decomposition products. To overcome this, liquid - liquid partitioningl7-19 and column clean up12>20 of the extracts have been used. The extracting solvents have included acetonitrile ,I5 hexane,7 various alcohols6J1 and aqueous solutions of alcohols.1 2 ~ ~ As trace impurities may cause antioxidant losses, high-purity solvents are mandatory. Moreover, considerable care must be exercised during any extraction steps in order to avoid low recoveries associated with particular types of samples. For example, BHA was recovered22 quantitatively from paper with 80% aqueous ethanol whereas the same procedure extended to lard23 gave non-quantitative yields of the antioxidant. Similarly, low recoveries of BHA from dehydrated potatoes have been attributed24 to the retention of BHA by the amylose com- ponent of starch. In this instance, using solvents of increasing dielectric constant did not improve the recoveries unless the potato granules were first hydrated.Steam distillation,l1725 vacuum distillation26 and vacuum steam distillation27 have also been used to isolate antioxidants but must be recognised as limited in value. Hence, Anglin et aZ. 11 found steam distillation most suitable for the recovery of BHA and BHT, and solvent extraction for the recovery of PG and NDGA. On the other hand, Stuckey and Osborne7 distinguished between high-fat and low-fat foods and treated each differently. Extraction with a hydrocarbon solvent, such as hexane in a Soxhlet apparatus, was proposed for the recovery of antioxidants from high-fat foods. The extract was then analysed directly or subjected to steam distillation. In this way, most high relative molecular mass compounds such as protein and phospholipids remained in the sample residue together with any bound antioxidant.The recovery from low-fat foods was made by direct steam distillation. However, steam distillation has been criticisedll for the time taken and the volume of distillate necessary to obtain quantitative recovery. The addition of salts to increase the temperature of the steam and hasten distillation was also found unsatisfactory owing to the difficulty in maintaining high temperatures long enough to complete the distillation. Predictably, superheated steam gave a quantitative recovery of BHA and BHT. However, the need for super- heated steam was subsequently eliminated by the use of a rapid steam distillation apparatus28 incorporating a mildly alkaline suspension of magnesium oxide to remove any sulphur dioxide in the volatile constituents.In contrast, Sloman et aZ.25 advocated steam distillation through a heated magnesium oxide suspension to remove any natural phenolic and acidic components from the sampIe that could interfere with the colorimetric determination of BHA and BHT. The heating of the suspension was said to improve the recovery of BHT. BHA and BHT have been recovered15 quantitatively from milk fat by distributing the sample on a Florisil column and eluting with aqueous acetonitrile. The extract was purified by solvent partitioning. On the other hand, direct extraction of the milk fat with acetonitrile provided a suitable approach for TBHQ recovery. Purging with nitrogen29 has proved a simple and efficient means of recovering all three antioxidants. Quantification of Antioxidants Antioxidants are used to prevent oxidative deterioration of fatty materials so that the usual sample is therefore an edible oil or fat, a shortening or a baked or dairy product containing many substances able to interfere with most analytical methods.For this and other reasons the earlier spectroscopic techniques (colorimetry and ultraviolet - visible spectro- photometry) have, to a large extent, been replaced by chromatographic methods. Colorimetric Methods The reactions that have been exploited in the colorimetric methods for determining phenolic antioxidants are based on the phenolic and reducing properties of the compounds. Ideally, a reagent would be specific for each antioxidant but of course, in practice, this is not so. For example, the two isomers of BHA generally react at different rates, a fact which complicates the determination of BHA colorimetrically .Three principal reagents have been applied to the analysis of antioxidants and may be classified according to their specificity. The least specific reagent, iron(II1) chloride - 2,2'-dipyridyl, depends on the reducing properties of the antioxidant to produce iron( 11) which gives a characteristic red complex (see equation 1) with 2,2'-dipyridyl: [Fe(H20)#+ + 3 dipyridyl- [Fe(dipyridyl)3]2+ + 6H20 . . (1) Obviously, any other compound capable of reducing iron(II1) will interfere. However, two further complications may ariseANALYST, JULY 1987, VOL. 112 935 with the reagent because BHA accelerates the reaction of BHT and, in addition, BHT reacts rather slowly, making control of the reaction time important.7.23 This reagent has been applied to the determination of BHA in paper and paper-board.22 For this, the antioxidant was extracted in a Soxhlet apparatus for 10 h with aqueous ethanol and the absorbance of the complex measured at 522 nm.A correction for the background was made by preparing a solution from paper containing no BHA. Anglin et al.11 used this reagent for the determination of BHA and BHT in edible oils following steam distillation recovery. BHA was deter- mined separately by reaction with Gibb's reagent (see below). A similar procedure was found suitable28 for the analysis of potato flakes containing BHA and BHT. In this instance, it was found that sodium sulphite, which is frequently added in the processing of potatoes, led to erroneously high results.Gibb's reagent, 2,6-dichloroquinonechlorimine, couples with phenolic substances in the orrho and para positions and consequently does not react with BHT. The reaction, illus- trated for phenol in equation (2), produces indophenols whose alkali metal and ammonium salts have a characteristic blue colour. Any phenolic compound is, of course, a potential interferent if it has a free ortho or para position. Indeed, Stuckey and Osborne7 have reported that naturally occurring substances, presumably phenolic, give false results. Despite this limitation and the dependence of the colour development on pH, the reagent has been used extensively22.25.28~3~34 in earlier work. CI Cl Cl The last of the commonly used reagents is diazotised sulphanilic acid,35,36 which reacts with BHA in ethanolic solution to form [see equation (3)] a red - purple azo dye.However, the reaction suffers from several complications. For example, the slope of the calibration graph is affected by the alcohol concentration of the test solution. In addition, the ratio of nitrite to sulphanilic acid concentrations is critical as too much nitrite causes rapid fading and too little reduces the intensity of the colour. Despite these limitations, uses of the reagentlgJ9 include an official AOAC method.37 + N2 OH I OCH3 I S03H I OH Of the less frequently used reagents, the reaction of BHT with an alcoholic solution of 3,3'-dimethoxybenzidine and nitrous acid18J9yz warrants mention. The chromogen result- ing from this reaction can be extracted into a chlorinated solvent for measurement at 520 nm but the colour is light-sensitive and dependent on the concentration ratio of the two reagents.38 Of a number of antioxidants studied for possible interference effects, only ethoxyquin (6-ethoxy-l,2- dihydro-2,2,4-trimethylquinoline) posed a problem although BHA was subsequently reported39 to enhance the absorbance of BHT.Recently, various other reagents have been proposed for antioxidant analysis, including N, N-dimethyl-p-phenylene- diaminea and p-aminophenol41 for BHA and p-(N-methyl- amino)phenol with periodate42 or permanganate43 for PG and BHA. Spectrophotometric Methods in the Ultraviolet Region In principle, the measurement of UV absorption (see Fig. 1) provides an excellent means for quantifying the phenolic antioxidants which give rise to characteristic UV spectra due to x - x* transitions.However, the method suffers from non-specificity and many substances can interfere in the determination. More recently, derivative spectroscopy has been shown@ to reduce the interference due to light scattering and matrix effects and improve the selectivity in the determi- nation of BHT in polypropylene. The measures adopted in the numerous papers45-50 (see Table 2) employing UV spectroscopy follow predictable patterns. Some endeavour to use only a solution of the sample and introduce corrections for the background absorption. Others separate the antioxidant(s) by a chromatographic separation,123*0 in recognition of the possible interference from other phenolic compounds occurring naturally or present in the samples.Examination of the UV spe~tra21723.~5 shows that methods based on measurements in the UV region are applicable to single pure compounds or, at best, for mixtures of PG and one other in the group BHA, BHT and NDGA. In other early work, Whetsel etaZ.46 described a UV method for the determination of BHA and PG involving the direct dissolution of the sample in propan-2-01 and measurement of the absorption at 232, 241 and 252 nm. WolffM described a similar method for the determination of BHA and gallic acid esters, following partition between a cyclohexane solution of the sample and ethanol at 72 "C, The absorption was 50 I-- - -1 ,T / "\ \ r i - 3OC , I e / >. Y .- > .- Y 2 20 2 a 11 10 0 240 260 280 300 320 Wavelengthhm bCH3 Fig.1. dants: A, BHT; B, NDGA; C, BHA; and D, Fd Absorption spectra in the ultraviolet re ion for the antioxi-936 ANALYST, JULY 1987, VOL. 112 measured at three wavelengths and the BHA and gallates were determined by substitution into a set of equations that included provision for the background absorption of the sample. Wolff49 also reviewed the application of UV and visible spectrophotometry to antioxidant analysis as applic- able at that time. BHT has been determined23 in edible fats following separation on silicic acid deactivated with 13% water. The cyclohexane eluate was examined at 284 nm for the determi- nation of BHT. In the work of Hansen et a1.,50 a chloroform - methanol extract of lard was examined at 270,290 and 310 nm for BHA after precipitation of the bulk of the fat by cooling.However, BHT and tocopherols interfered. Paper and Thin-layer Chromatographic Methods Paper (PC) and thin-layer chromatography (TLC) have been used in the separation and determination of antioxidants more extensively than any other technique. The reasons for this include simplicity, low cost and the mode of presenting the data. Before the advent of more sensitive techniques, separa- tion on layers provided an extremely attractive proposition, especially as a technique that lends itself to simultaneous isolation, identification and quantification. In an early study of the suitability of PC, TLC and colorimetry for antioxidant analysis, Rutkowski et al.51 reported that separation is most precise by PC and most rapid by TLC.However, whereas TLC allows the use of more corrosive spray reagents, reproducibility of RF values is generally poorer than in PC.52 The relative merits of PC and TLC for antioxidant analysis have also been discussed5555 elsewhere. As most antioxidants are highly polar molecules, efficient separation on paper can only be achieved with the use of highly polar mobile phases.56 Consequently, reversed-phase chromatography57-60 or the use of acetylated papers61 (see Tables 3 and 4) are frequently employed to reduce the effect of “tailing.” Probably the most comprehensive TLC scheme is that devised by van der Neut and Maagdenberg62 in which the antioxidants are separated into groups according to RF ranges by a preliminary chromatogram.Based on this preliminary classification, subsequent solvent systems are selected until complete identification is achieved. Nine solvent systems and four detection reagents are specified and the scheme has been applied to over 30 antioxidants. More recently, a scheme for the separation of 11 antioxidants has been described14 involving only two adsorbents and three solvent systems. High-performance TLC has also been used63 for measuring antioxidant activity. Polyamide powder has been widely used as a support in the TLC of antioxidants.64-70 It has been suggested66 on the basis of work involving lipophilic compounds such as antioxidants that the polyamide layer acts as an apolar support or, in special instances, as a kind of polar stationary phase.The nature of the phase is determined by the composition of the mobile phase. Separation on acetylated cellulose,71.72 alumina,73 silica gel and impregnated silica ge174-96 has also been used successfully. Silica gel has been used95 in a collaborative survey of the separation of gallate esters, BHA, BHT and NDGA. The method gave satisfactory results for soybean oil, lard and high-fat cakes, but was unsatisfactory for corn oil. Apart from the approaches outlined above, several multi- technique procedures have been described.97-100 For example, TLC, column and gas chromatography (GC) have aided99 the identification of antioxidants in oils. Various components including antioxidants were identified100 in chewing gum by TLC, IR and NMR spectrometry. In other instances, BHA101 and BHT in addition to various flavours and preservativeslO* were recovered by a vapour separation technique prior to TLC separation.The reagents developed for the colorimetric methods for antioxidants are suitable as spray reagents for antioxidant detection. Van der Heidel03 identified four broad groups Table 2. Summary of ultraviolet methods for the determination of antioxidants Analytical Antioxidant Sample Solvent wavelengthhm Comments Reference BHA, BHT . . . . . . Hydrocarbon Aqueous ethanol BHT . . . . . . . . Oil Cyclo hexane BHA, NDGA, gallates Oil Methanol - ethanol BHT . . . . . . . . Not applied Alkaline ethanol BHA,PG . . . . . . Oil Propan-2-01 BHA . . . , . . . . Oil Ethanol BHA . . . . . . . . Fat Chloroform - methanol 283.5 Florisil column 12 276 Cellulose column 20 21 30 46 242,274,285 Background 48 clean-up clean-up 240,250,274 - 306 - 232,241,252 - absorption must be known interfere 290 BHT and tocopherols 50 Table 3.Summary of selected paper chromatographic methods Antioxidant Solvent system BHA, BHT, PG, NDGA . . . . 80% aqueous methanol, then water BHA, PG, NGDA . . . . . . 5% aqueous methyl or ethyl acetate BHA, PG . . . . . . . . 75% aqueous dioxane BHA, BHT, PG . . . . . . Ethyl acetate - acetone - aqueous BHA, BHT, PG, NDGA . . . . (1) Acetone - ethyl acetate - water sodium acetate (3 + 2 + 75) (3 + 1 + 6) (2) Acetone - water (60 + 40) Detection reagent * Comments Reference soybean oil in ether - Reversed phase on 58 liquid paraffin sesame oil PMA arachis oil PMA (solvent 1) (solvent 2) PMA Reversed phase on 57 - Reversed phase on 59 Reversed phase on 60 Reversed phase 61 Acetylated paper * PMA = phosphomolybdic acid or ammoniacal phosphomolybdic acid.ANALYST, JULY 1987, VOL.112 937 Table 4. Summary of selected thin-layer chromatographic methods Antioxidant BHA, BHT, PG, NDGA, TBHQ BHA,BHT,PG,NDGA . . BHA,PG,NDGA . . . . BHA,PG,NDGA . . . . BHA, BHT, PG, NDGA . . Extensive . . . . . . PG . . . . . . . . . . BHA,BHT,PG . BHA, BHT, PG, NDGA . . Extensive . . . . . . BHA,BHT,PG,NDGA . . BHT . . . . . . . . BHA, BHT, PG, NDGA . . . . . . . . . . . . . . . . . . . . . . . . . . BHA, BHT, PG, NDGA, TBHQ BHA, BHT . . . . . . . . PG . . . . . . . . . . . . BHA, BHT,PG,NDGA . . . . BHA, BHT, PG, NDGA . . . . Extensive . . . . . . . .BHA, BHT, PG, NDGA . . . . Solvent system Light petroleum - benzene - acetic acid (40 + 40 + 20) Hexane - acetone - acetic acid (55 + 40 + 15) Acetonitrile - hexane Methanol - acetone -water Hexane - benzene - (6 + 1 + 3) acetic acid - dimethylformamide (10 + 10 + 5 + 0.25) Chloroform - methanol - hexane Methanol -water (3 + 1) Shellsol - propanol - acetic acid - formic acid (15 + 2 + 1 + 2) Light petroleum - dioxane Benzene then ace t onit rile Hexane - isobutanol - acetic acid (70 + 20 + 10) Cyclohexane - dioxane - acetic acid (80 + 15 + 5 ) Methanol -water - acetic acid (82 + 16 + 2) Chloroform - methanol - acetic acid (90 + 10 + 2) Cyclohexane - ethyl acetate Benzene - acetic acid (15 + 4) Chloroform - methanol - acetic acid (80 + 3 + 10) Hexane Ethyl acetate - cyclohexane Benzene - light petroleum - acetic acid (40 + 40 + 20) Chloroform - methanol - acetic acid (90 + 10 + 2) Chloroform Methanol - water - acetic acid (82 + 16 + 2) Hexane - acetic acid (9 + 1) (10 + 1) (10 + 1) Detection reagent* B B or dianisidine A, C Fe3+ - hexacyanoferrate(II1) A B - NH3 Fe3+ A B - B B B Fluorescence or Fe3+ - dipyridyl Folin - Ciocalteau reagent Cu2+ or Fe3+ B A - Folin - Ciocalteau reagent Adsorbent Silica gel Polyamide powder Polyamide powder Polyamide powder Polyamide powder; silica gel Modified cellulose; resins Acetylated cellulose Alumina Silica gel Silica gel RP-18 Silica gel Silica gel Kieselgel Silica gel Silica gel Silica gel Silica gel Silica gel Silica gel and reversed phase Silica gel Reference 14 64 65 66 67 71 72 73 76 78 80 88 90 91 92 93 95 96 98 106 * Reagents: A, phosphomolybdic acid or ammoniacal phosphomolybdic acid; B, Gibb's reagent; and C, diazotised sulphanilic acid.containing twelve spray reagents and tabulated the colour given with various antioxidants. In addition, a spray reagent, said to be selective and extremely sensitive to antioxidants (1-ng detection limit for BHA), has been described.104 The reaction sequences consisted of a p-ocimene spray followed by a 2,4-dinitrophenylhydrazine overspray. The sensitivity was dependent on the efficiency of the antioxidant in preventing autoxidation of p-ocimene. A similar procedure105 using a tocopherol-stripped soybean solution and exposure to UV radiation relied on the measurement of the relative fluores- cence produced.Antioxidant migration from plastics into oil has been studied by TLC** and, in combination with spectropho- tometry,l06 TLC was used to measure BHA, BHT, PG and NDGA in plastic containers and oils. Despite the serious limitations of TLC compared with GC, particularly regarding detection limits, continued interest in TLC seems assured because of the low capital outlay and its suitability in studying matrix effects. Gas Chromatographic Methods Free antioxidants The attraction of GC lies in its ability to separate, identify and determine nanogram amounts of complex mixtures simul- taneously. As the phenolic antioxidants are relatively non- volatile, low stationary phase loadings are necessary to reduce the retention times to a sensible value. However, the more polar PG and NDGA have generally been derivatised prior to GC in an attempt to reduce the retention times further and to improve the chromatographic performance.On the other hand, BHA, BHT and, to a lesser extent, TBHQ exhibit excellent GC properties and may be determined on either polar or non-polar stationary phases. 107-114 Current practice differs mainly in the use of fewer and lower loadings of stationary phases than was so in earlier work. GC methods (see Table 5 ) have been used to determine antioxidants in fats and oil~,26,*~,83,1121115-12~ dairy pro- ducts,15,122,123 chewing gum,124 potato granules,l3*24 sea- foods, 125,126 breakfast cereals ,397 127-133 waxed paper, 128 meat products134 and vitamin oils.135 As an example of early detailed methodology, Takahashi39 developed a procedure for determining BHA and BHT in breakfast cereals following extraction with hexane.The extract was concentrated and analysed on a glass column packed with 10% silicone grease on Chromosorb P at 160 "C. The recoveries of the added amounts at levels of 0.5-50 p.p.m. ranged from 83 to 120% for BHT and 89 to 104% for BHA. These GC results were compared with those obtained using a colorimetric method.25 In a subsequent modification129 of this procedure, subjected to collaborative survey,130 the antioxidants were extracted by938 ANALYST, JULY 1987, VOL. 112 percolating hexane through the sample packed in a short glass column. The extract was concentrated by removing the solvent at room temperature under nitrogen and examined by GC on a column of 20% DC-200 oil on Chromosorb P coupled to an argon ionisation detector.Resulting further from this collaborative survey, the method was again revised.131 In the final f0rm,132J33,13~ the antioxidants were extracted with carbon disulphide, concentrated and analysed on a column of Table 5. Columns and conditions used for the GC of free antioxidants and their derivatives Antioxidant Sample BHA, BHT . . . . . . Potato granules BHA, BHT . . . . . . Dehydrated mashed potato BHA,BHT,NDGA . . . . Lard BHA, BHT, TBHQ . . . . Soybean oil Column packing Apiezon L on firebrick 3% OV-17 on Gas-Chrom Q 3% GE-XE-60 on Gas-Chrom Q 10% polymetaphenoxylene on 10% silicone grease on 10% GE-Versilube F-50 on 10% SE-30 on Aeropack 30 SE-30, QF-1 or neopentyl glycol succinate on Chromosorb W 10% silicone grease on Chromosorb P 5% DEGS + 1% H3P04 on Chromosorb W DC-200 or Carbowax 20M on Gas-Chrom Q 3% JXR on Gas-Chrom Q 10% polymetaphenoxylene on “Cross-linked 5% Tenax GC Chromosorb P Gas-Chrom Q Tenax GC phenylmethylsilicone fused-silica capillary” DB-1 fused-silica WCOT 5% DEGS + 1% H3P04 on 30% SE-30 on Chromosorb W Chromosorb W Column temperaturePC 220 165 150-250 140-250 Reference 13 24 26 29 39 107 111 113 115 116 117 118 119 120 124 126 127 128 129,130 136 134 135 139 26 107 113 131-133, 118 139 142 144 145 147 148 149 150 151 152 154 BHA, BHT .. . . . . Cereal 160 BHA,BHT . . . . . . Oils 150-210 BHA,BHT . . . . . . Oils BHA, BHT . . . . . . Food extracts 150-215 130-230 BHA, BHT . . . . . . Fats, oils 160 BHA, BHT .. . . . . Shortening 140-170 BHA,BHT . . . . . . Oils 160 BHT.. . . . . . . . . Oils BHA,BHT,TBHQ , . . . Oils 105-250 140-250 BHA,BHT,TBHQ . . . . Oils 50-280 BHA, BHT . . . . . . Chewing gum TBHQ . . . . . . . . Seafood, oils 65-200 196 BHA, BHT . . . . . . Cereals BHA, BHT . . . . . . Cereals, potato products, BHA, BHT . . . . . . Cereals BHA, BHT . . . . . . Cereals waxed paper 125-190 SE-30 on Tween 80 20% DC-200 on Chromosorb P QF-1 or Apiezon L on 5% GE-XE-60 on Gas-Chrom Q 20% SE-30 on Chromosorb W or 3% OV-17 on Gas-Chrom Q 3% GE-XE-60 on Gas-Chrom Q Gas-Chrom Q firebrick 150 160 160 BHA, BHT . . . . . . Meat products BHA, BHT . . . . . . Foodstuffs, vitamin A oil 150 152 15&250 85- 175 BHA, BHT, TBHQ . . . . Fats, oils, dried foods TMS derivative of PG .. . . Lard TMS derivative of BHA, BHT, PG, NDGA Oils 10% GE-Versilube F-50 on Gas-Chrom Q SE-30, QF-1, or neopentyl glycol succinate on Chromosorb W 150-210 TMS derivatives of gallates . . Food extracts 130-230 TMS derivatives of BHA, BHT, PG, NDGA,TBHQ . . . . TMS derivatives of PG,NDGA,TBHQ . . Benzoylated derivative of BHA, BHT, (PG, NDGA) Heptafluorobutyrate esters ofBHA,PG,TBHQ . . Trifluoroacetate ester of BHA, (BHT) . . . . . . Cyclohexadienone derivative of BHT . . . . . . . . TMS derivatives of BHA, gallates . . . . . . . . Various . . . . . . . . Haloacetates of PG, BHA andothergallates . . . . Oils Fats, oils, dried foods Not applied Oils Not applied Not applied Not applied - Not applied 3% JXR on Gas-Chrom Q 105-250 3% OV-225 on Gas-Chrom Q 100-250 OV-73 glass WCOT 120-300 3% OV-3 on Chromosorb W HP 120 5% SE-30 on Chromosorb W AW 160 2% OV-1 on Chromosorb W 150 4% silicone on Diataport Various 150-250 200-250 3% OV-101 on Chromosorb W HP 5% OV-101 on Chromosorb G, H, p 130-260 Methyl ether of PG and NDGA .. . . . . . . Not applied 3% OV-101 or QF-1 or 4% OV-17 on Chromosorb W 20% SE-31 on Celite 545 10% Versilube F-50 on Gas-Chrom Q 205-280 TMSderivativeofPG . . . . Not applied TMS derivative of TBHQ . . Oils 200 190ANALYST, JULY 1987, VOL. 112 939 5% Apiezon L on Gas-Chrom Q and 10% QF-1 on Gas- Chrom Q using 3,5-di-tert-butyl-4-hydroxyanisole as an inter- nal standard. A feature of these columns is that the elution order of BHA and BHT is reversed. In fact, only recently the use of at least two columns has been recommended137 to avoid interferences from hydrocarbons in dried fish during the GC determination of BHA and BHT.BHA and BHT have also been determined128 in breakfast cereals, waxed paper and potato granules following the extraction of the antioxidants in diethyl ether and concentra- tion at 3540 "C under a stream of nitrogen. The concentrated extract was chromatographed on a metal column packed with SE-30 and Tween 80 coated on Chromosorb W with excellent resolution and recovery of both antioxidants. However, extreme care needs to be exercised to avoid loss of some antioxidants during the concentration procedures. The nature of the solvent, temperature and evaporation time deter- mine138 the extent of any antioxidant losses, which can be severe for diethyl ether.Such losses have been attributed13 to the presence of peroxides in the diethyl ether. Another early GC method for determining BHA and BHT in potato granules within the range 0.5-10 p.p.m. of antioxi- dant has been described." Here, a stainless-steel column, packed with 20% Apiezon L on firebrick and operated at 220 "C, was used. Conditioning of the column for 1 week was said to be esssential for the satisfactory resolution of BHA, which is eluted first. The antioxidant was recovered by a lengthy solvent extraction procedure and the consistently low results were attributed to incomplete extraction. Apparent decomposition of BHA on the column was overcome by the injection of a BHA standard immediately prior to the unknown. However, Choy et al.135 reported on-column decomposition of BHA using this procedure. Although polar columns gave better resolution of the antioxidants, excessive base-line noise was also observed above 200 "C. In a novel step, Hartman and Rose117 used two stationary phases to reverse the elution order of BHA and BHT to verify their presence in vegetable oils following dissolution of the oil in carbon disulphide. A short stainless-steel pre-column containing siliconised glass-wool separated the non-volatile oil components from the antioxidants and an aluminium column consisting of DC-200 or Carbowax 20M on Gas-Chrom Q, operated at 160 or 190 "C for the latter, was used for the analysis. The recoveries of BHA and BHT added to various oils were 97-104% and each analysis took only 20 min to perform.In contrast, McCaulley et al.,26 recognising that the conventional extraction procedures without adequate clean- up could introduce co-extractants capable of interfering in the GC determination of antioxidants, devised an alternative procedure involving vacuum sublimation to recover several antioxidants from lard. The extracts were chromatographed on a column packed with a nitrile silicone gum stationary phase coated on Gas-Chrom Q, using temperature program- ming. The average recoveries of BHA and BHT added to lard ranged from 91 to 96%. The identity of the column eluate was confirmed by UV, IR and mass spectrometry (MS). Following the acetonitrile extraction of BHA and BHT from oil or fat and liquid - liquid partitioning, Schwien et al. 115 found clean- up of the extract on an alumina column to be essential prior to the GC analysis of the antioxidants.More recently, both direct injection of a solution and extraction before GC have been used139 in a single procedure. Again, after extraction and clean-up, BHT in powdered milk has been identified123 by coupled GC - MS. TBHQ has been determined quantitatively by GC in 01ls,~~J19 seafoods126 and dairy products122 as the underivat- ised compound. However, substantial losses have also been reported15 for this antioxidant. Hence, the determination of BHA and BHT added to milk fat gave recoveries between 80 and 110% whereas only 40% of oxygen-sensitive TBHQ was recovered. Fry140 showed that the conversion of TBHQ to recovered. Fry140 showed that the conversion of TBHQ to tert-butylquinone was the cause of the low recoveries in the GC method of analysis.Consequently, when the method for TBHQ in miik fat was modified15 to allow for the rapid extraction and preliminary silylation of the antioxidant, the recovery of added TBHQ exceeded 83%. The dual advantages of reduced activity and retention afforded by the capillary columns141 have been realised in the determination of BHA and BHT in chewing gum124 and BHA, BHT and TBHQ in vegetable oils.12O In both instances, fused-silica wall-coated open-tubular (WCOT) columns improved the resolution and the reproducibility of the retention times and peak areas. Derivatives The GC behaviour of BHA and BHT is adequate for quantitative purposes but pre-column derivati~ation,l18J42-15~ in particular trimethylsilylation,118~148J49 has been employed for many years in the determination of these antioxidants.Further, although trimethylsilylation, methylation149 and acetylation149 may result in an increase in sensitivity, such an increase is secondary and incidental. Trifluoroacetyla- tion,145J49 on the other hand, is a different proposition because of the enhanced sensitivity that can be obtained by use of an electron-capture detector (ECD). Thus, a 10-fold reduction in the detection limit was achieved145 for BHA using the trifluoroacetate ester and an ECD. Moreover, the separation of BHA from BHT was improved by trifluo- roacetylation as BHT remains underivatised under the condi- tions employed, The same considerations apply to the heptafluorobutyric acid derivati~es143114~ although here,143 BHT was also derivatised by prior formation of the corre- sponding cresol.In contrast to BHA and BHT, the polyhydric nature of PG and NDGA reduces their volatility to such an extent that they cannot be gas chromatographed successfully without recourse to derivatisation. Consequently, the GC methods for these antioxidants include derivatisation as a necessary step prior to GC.26~113~11S,144,149-152 For example, attempts to determine NDGA without derivatisation26J53 have been largely unsuc- cessful, whereas Stoddardll8 produced a lengthy but success- ful method for the determination of NDGA by GC, involving formation of the trimethylsilyl ether and separation on a 3% JXR on Gas-Chrom Q column. This approach enabled all five antioxidants, in addition to the isomers of BHA, to be resolved.Although the recoveries of the other four antioxi- dants from peanut oil were greater than 87%, those of the 2- and 3-isomers of BHA were only 76 and 68%, respectively. Nishimoto and Uyetall3 separated several gallate esters as their trimethylsilyl derivatives on any one of three stationary phases. However, in a detailed study152 of the behaviour of the trimethylsilyl ethers of gallates on several stationary phases, the non-polar phases were preferred. The trimethyl- silyl derivatives have also been used148 to separate alkyl gallates including PG from BHA. Finally, the trimethylsilyl ether154 and heptafluorobutyrate ester derivatives144 of TBHQ have also been employed in GC methods. Other derivatives include two haloacetates of the gallates150 and the tetramethyl ether of NDGA.151 In a GC-MS study,lso the trifluoroacetates and heptafluorobutyrates of PG, OG and LG were well separated on OV-101 columns, using temperature-programmed GC and a flame-ionisation detector.Similarly, the trimethyl ether151 of PG, synthesised with diazomethane, was successfully chromatographed using relatively non-polar stationary phases (OV-3, SE-30, OV-17 and QF-1). The retention time was 5 min at 205 "C, with a detection limit of about 1 ng. The tetramethyl ether of NDGA,151 being appreciably less volatile, required higher temperatures. The conditions best suited to this compound were a column packed with 4% OV-17 and a temperature of 280 "C, with a detection limit of approximately 5 ng. The940 ANALYST, JULY 1987, VOL.112 calibration graphs were linear over the range 5-100 p.p.m. for both derivatives. Recently, a method for the determination of nine phenolic antioxidants in oils has been developed162 and submitted to a collaborative survey. 17 The antioxidants, including BHA, BHT, PG, NDGA and TBHQ, were partitioned from a hexane solution of vegetable oil into acetonitrile, concen- trated in vacuo, diluted with propan-2-01 and separated by reversed-phase gradient elution on a LiChrosorb RP-18 column using detection at 280 nm. The over-all mean recoveries of the antioxidants from “spiked” vegetable oils exceeded 90%. In contrast, direct injection and fluorimetric detection have been successfully applied166 to the HPLC determination of TBHQ.Simultaneous detection with fluorescence, UV and electro- chemical detectors has improved the analysis of phenolic antioxidants and related compounds following isocratic elu- tion from a reversed-phase microparticulate c18 column. 159 Normal-phase separations of the antioxidants, as expected, have been used less commonly and, indeed, appear to be restricted to the analysis of TBHQ, BHA and BHT.51167-169 For example, the isomers of BHA have been separated168 by normal-phase chromatography and TBHQ has been analy- sedl67 using a LiChrosorb NH2 column with hexane - ethanol (85 + 15) as the mobile phase. The three antioxidants BHA, BHT and TBHQ have been determined169 in edible oils using either pPorasil or Rad-Pak Cyano columns. The more polar PG and NDGA are adsorbed on the columns.Size-exclusion chromatographyl70J71 holds the promise of Methods Based on High-performance Liquid Chromatography (HPLC) HPLC92J55 offers all the advantages of GC and, in addition, the increased flexibility of a liquid mobile phase. However, the HPLC detectors are generally less sensitive. Hence, for example, the detection limit for the 4-methoxyketone deriva- tive of BHT is 5 pg with an ECD or 0.5 ng by HPLC with detection at 236 nm.147 Despite the lower sensitivity of HPLC, ultraviolet and electrochemical detection based on the reduc- ing properties of the phenolic antioxidants provide adequate sensitivity for many applications.156J57 Further, the polar nature of these compounds, in particular PG and NDGA, is ideally suited to reversed-phase chromato- graphy. Indeed, a variety of reversed-phase chromatographic procedures (see Table 6) have been developed for the analysis of antioxidants,16,17,15~165 mostly utilising UV detection.For example, Hammondlm reported the separation of five antiox- idants including BHA, BHT and PG on a yBondapak column using gradient elution with aqueous acetic acid and methanol solvents. King et aZ.161 also used a pBondapak c18 column to separate BHA, PG and TBHQ extracted from oils and foods. However, in this instance amperometric detection was employed. Table 6 . Systems used for the HPLC of antioxidants Antioxidant Sample BHA,BHT,TBHQ . . Oils Stationary phase (column) Mobile phase Acetonitrile - 0.05 M NaH2P04 (PH 3.0) Reference 6 17 92 142 147 155 157 158 159 160 161 162 164 165 166 167 169 171 173 Detection Amperometric Unisil Q CP and Diasil CN (150 x 4 mm) BHA, BHT, PG, NDGA, TBHQ ., . . . . Oils, lard LiChrosorb RP-18 (250 x 4.6 mm) Corasil I1 (1 m length) Gradient elution 280 nm ? 235 nm BHT . . . . . . . . Peanut oil Benzoylated derivatives of BHA, BHT, PG, NDGA Not applied Heptane LiChrosorb SI 60 or RP-18 (300 x 3 mm) Isocratic and/or gradient elution Cyclohexadienone derivative of BHT . . Not applied BHT . . . . . . . . Not applied Antioxidants . . . . . . Soysauce BHA, BHT, PG, TBHQ . . Fats, oils Hexane - propan-2-01(99 + 1) Cyclo hexane Gradient mixture: - 1% acetic acid in water and 1% acetic acid in methanol methanol 0.05 M LiC104 in aqueous pPorasil Corasil I1 Zorbax CN pBondapak C18 (300 x 3.9 mm) 236 nm uv 280 nm - BHA, BHT, TBHQ .. Not applied Partisil PXS 10/25 ODS-2 uv, fluorescence, electrochemical 280 nm BHA, BHT, PG . . . . Fats, oils pBondapak C18 (300 x 4 mm) pBondapak CIR (150 x 4.2 mm) Aqueous acetic acid + methanol Methanol - 0.1 M gradient elution ammonium acetate (1 + 1) BHA,PG,TBHQ . . . . Oils,foods Amperometric BHA, BHT, PG, NDGA, TBHQ . . . . . . Fats LiChrosorb RP-18 (250 x 3 mm) Gradient elution from water - acetic acid (95 + 5 ) to acetonitrile - acetic acid (95 + 5 ) Acetonitrile -water (2 + 3) 280 nm 228 nm 280 nm Fluorescence 293 nm 280 nm BHA . . . . . . . . Isomer BHT . . . . . . . . Oils, feeds TBHQ . . . . . . . . Oils separation Hypersil ODS Partisil ODS 3 LiChrosorb Si 60 (250 x 4 mm) LiChrosorb NH2 (250 x 4mm) pPorasil or Rad-Pak Cyano 85% aqueous methanol Dioxane - hexane (24 + 76) TBHQ .. . . . . . . Oils, butter Hexane - ethanol (85 + 15) BHA, BHT, TBHQ . . Edible oils Hexane - dichloromethane - acetonitrile (e.g. 85 + 9.5 + 5.5) Tetrahydro fur an Refractometry + UV at 254 nm BHA, BHT, PG, NDGA Fats, oils S-832 gel BHT and decomposition products . . . . . . Polyethylene Partisil(250 X 4 mm) Dichloromethane - hexane gradient elution 242 nmANALYST, JULY 1987, VOL. 112 941 direct injection of complex samples with minimal sample pre-treatment. Nevertheless, Pokorny et al. 171 introduced a preliminary extraction in the determination of BHA, BHT, PG and NDGA in edible fats and oils by size exclusion. More recently, the advantages of direct injection have been realised in the determination170 of BHA, BHT and TBHQ in edible fats and oils. Here, the sample was simply dissolved in chloroform prior to injection.The migration of BHT out of polypropylene (PP) and poly(viny1 chloride) into several spices has been studied by HPLC172 with an extremely high migration from PP into stored cloves being reported. The phenolic antioxidants and their degradation products formed in polyethylene have also been separated173 by HPLC on a Partisil column using gradient elution with hexane - dichloromethane. Indeed, the analysis of antioxidants in, and their migration from, food packaging materials is a pressing problem as antioxidants are frequently incorporated into such materials where they are subject to degradation especially during processing or on exposure to UV radiation.Miscellaneous Techniques The various methods that have found occasional application to the determination of antioxidants include polarography and related electrochemical methods ,243174180 complexometry,181 fluorimetry,l82-185 infrared186 and nuclear magnetic resonance spectrometry187 and kinetic methods.l88>l89 Polarographyl7’ has been used in the determination of antioxidants by anodic oxidation in a basic solution at the rotating graphite electrode.175 A feature of such methods is that no prior extraction of antioxidant is required. 179 The related techniques of chronopotentiometryl74 and voltam- metry24,176>178>180 have had limited success in antioxidant analysis. The latter has been applied176 to the determination of BHA in oil and lard by using a stationary planar vitreous carbon electrode.Differential-pulse voltammetry has been used180 for the selective determination of BHA in flow injection analysis and the sensitivity has been increased178 by employing a rotating glassy carbon electrode. As antioxidants react with peroxide radicals and inhibit oxidation of hydrocarbons, the resulting decrease in fluores- cence has been used182 as a measure of antioxidant activity. Latz and Hurtubise183 studied the luminescence properties of BHA, BHT and PG and based an analytical method for PG in lard on their results. Chloroform, used as a solvent for the sample, quenches the fluorescence of BHA, BHT and the background fluorescence of lard, but not PG. In a subsequent paper,184 spectrofluorimetry was used for the measurement of BHA in lard, cereals and waxed liners.However, the procedure for each type of sample varied considerably. In the simplest instance, a standard additions technique was employed, whereas for more complex samples, a chromatographic separation was necessary prior to spectro- fluorime try. Paris et al. 177 based their method for the determination of NDGA on the fact that the 2,4,6-tri-tert-butylphenoxy radical is capable of extracting the labile hydrogen from antioxidants. It was found that the major interference encountered arose from the diffusion of atmospheric oxygen into the titration vessel, in a procedure where the end-point was detected potentiometrically or spectrophotometrically at 625 nm. Complexometric methods for determining antioxidants have been based on the reduction of silver nitrate to metallic silver and titration with EDTA via the substitution reaction with tetracyanonickelate .I81 PG has been determined by precipitation with mercury(I1) acetate and complexometric titration of the mercury.181 Cerium(1V) sulphate titration in 72% ethanol has also been used,190 the end-point being detected potentiometrically or with a redox indicator.However, the temperature must be controlled because above 50 “C, ethanol is oxidised also. This method was unsuitable for sesame, cottonseed and olive oils, and oxidised oils with peroxide values exceeding 10. The support of the Applications Laboratory, Pye Unicam, Cambridge, UK, is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.1 Oa 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. References Sherwin, E. R., J. Am. Oil Chem. SOC., 1972,49,468. Adhikari, J., and Adhikari, S., Indian Food Ind., 1985, 4(3), 106. Ahmad, S., and Augustin, M. A., J. Sci. Food Agric., 1985,36, 393. Galea, V., and Costache, C., Zgiena, 1965, 14, 299. Piironen, V., Varo, P., Syvaoja, E. L., Salminen, K., and Koivistoinen, P., Int. J. Vitam. Nutr. Res., 1984, 54(1), 35. Sanders, T. A. B., and Coppen, P. P., in Allen, J. C., and Hamilton, R. J . , Editors, “Rancidity in Foods,” Applied Science, Barking, 1983, pp. 67-108. Stuckey, B. N., and Osborne, C. E., J. Am. Oil Chem. SOC., 1965,42, 228. Copius-Peereboom, J. W., in Stahl, E., Editor, “Thin Layer Chromatography,” Second Edition, George Allen and Unwin, London, 1969, pp.631-636. Crompton, T. R., “Additive Migration From Plastics Into Food,” Pergamon Press, London, 1979. Maddox, D. N., Int. Flavours Food Addit., 1976, 7(3), 117. Emmanuel, N. M., and Lyaskovskaya, Yu. N., translated by Allen, K. A., “The Inhibition of Fat Oxidation Processes,” Pergamon Press, London, 1967. Anglin, C., Mahon, J., and Chapman, R., J. Agric. Food Chem., 1956, 4, 1018. Alicino, N. J., Klein, H. C., Quattrone, J. J., and Choy, T. K., J. Agric. Food Chem., 1963, 11, 496. Buttery, R. G., and Stuckey, B. N., J. Agric. Food Chem., 1961, 9, 283. Van Peteghem, C. H., and Dekeyser, D. A., J. Assoc. Off. Anal. Chem., 1981, 64, 1331. Stijve, T., and Diserens, J. M., Dtsch. Lebensm.-Rundsch., 1983, 79(4), 108.Kitada, Y., Ueda, Y., Yamamoto, M., Shinomiya, K., and Nakazawa, H., J. Liq. Chromatogr., 1985, 8 , 47. Page, B. D., J. Assoc. Off. Anal. Chem., 1983, 66, 727. Heidrick, P. L., and Conroy, H. W., J. Assoc. Off. Agric. Chem., 1962,45, 244. Schwien, W. G., and Conroy, H. W., J. Assoc. Off. Agric. Chem., 1965,48,489. Aczel, A., Selmeci, G., Noske, O., and Marik, M., Husipar, 1972,21,222. Salazar Macian, R., and del Pozo, A., Galenica Acta, 1963, 16, 45, Austin, J. J., J. Am. Oil Chem. SOC., 1954, 31, 424. Phillips, M. A., and Hinkel, R. D., J. Agric. Food Chem., 1957, 5 , 379. Beaulieu, F., and Hadziyev, D., J. Food Sci., 1982, 47, 589. Sloman, K. J . , Romagnoli, R. J., and Cavagnol, J. C., J. Assoc. Off. Agric. Chem., 1962, 45, 76. McCaulley, D. F., Fazio, T., Howard, J.W., Di Cuircio, F., and Ives, J., J. Assoc. Off. Anal. Chem., 1967, 50, 243. Shu, C. K., and Mookherjee, B. D., J. Agric. Food Chem., 1985,33, 1107. Filipic, V. J., and Ogg, C. L., J. Assoc. Off. Agric. Chem., 1960,43, 789. Min, D. B., and Schweizer, D., J. Food Sci., 1983, 48, 73. Berger, K. G., Sylvester, N. D., and Haines, D. M., Analyst, 1960, 85,341. Kobayashi, K., Tonogai, Y., and Ito, Y . , Nippon Shokuhin Kogyo Gakkai-Shi, 1986, 33, 134. Mahon, J., and Chapman, R. A., Anal. Chem., 1952,24,534. Raadsveld, C. W., and Kooy, C. G., Ned. MelkZuiveltijdschr., 1961, 15, 282. Davidek, J., and Beniac, F., 2. Lebensm.-Unters.-Forsch., 1966, 131, 32. Pujol Forn, M., and Lopez Sabater, M. C., Circ. Farm., 1984, 42, 3. Laszlo, H., and Dugan, L. R., J.Am. Oil Chem. Soc., 1961,38, 178.942 ANALYST, JULY 1987, VOL. 112 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81 82. 83. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, AOAC, Washington, DC, 1970, p. 330. Szalkowski, C. R., and Garber, J. B., J. Agric. Food Chem., 1962, 10, 490. Takahashi, D. M., J. Assoc. Off. Agric. Chem., 1964,47,367. Komaitis, M. E., and Kapel, M., J. Am. Oil Chem. SOC., 1985, 62, 1371. Rao, K. E., Murthy, K. V., and Sastry, C. S., Zndian Drugs, 1982, 20(2), 58. Sastry, C. S., Rao, K. E., andPrasad, U. V., Talanta, 1982,29, 917. Prasad, U. V., Rao, K.E., and Sastry, C. S., Food Chem., 1985, 17,209. SouEek, J., Jelinkova, E., Analyst, 1982, 107, 623. Seher, A., Fette Seifen Anstrichm., 1958, 60, 1144. Whetsel, K. B., Roberson, W. B., and Johnson, K. E., J. Am. Oil Chem. SOC., 1955,32,493. Wa, M. C., Wang, S. J., and Tsen, H. Y., Chung Kuo Nung Yeh Hua Hseuh Hui Chih, 1984,22, 81. Wolff, J. P., Rev. Fr. Corps Gras, 1958, 5 , 630. Wolff, J. P., Chem. Anal. (Warsaw), 1958,40, 159. Hansen, P. V., Kauffman, F. S., and Wiedermann, L. H., J . Am. Oil Chem. SOC., 1959,36, 193. Rutkowski, A., Koslowska, H., and Szerszynski, J., Rocz. Panstw. Zakl. Hig., 1963, 14, 361. Brodasky, T. F., Anal. Chem., 1964, 36, 996. Seher, A., Nahrung, 1960, 4, 466. Seher, A,, J. SOC. Cosmet. Chem., 1962, 13, 385. Seher, A., Fette Seifen Anstrichm., 1959, 61, 345.Wheeler, D. A., Talanta, 1968, 15, 1315. Mitchell, L. C., J. Assoc. Off. Agric. Chem., 1957, 40, 909. Yasuhara, S., and Masuyama, S . , Kagaku To Kogyo (Osaka), 1957, 31, 399. Kawashiro, I., Kawata, K., and Hosogai, Y., Eisei Shikenjo Hokoku, 1959,77, 175. Subramanian, S. A., Curr. Sci., 1966,35, 309. Dehority, B. A., J . Chromatogr., 1959, 2, 384. van der Neut, J. H., and Maagdenberg, A. C., Plastics ’ (London), 1966, 31. 66. Sano, M., Abe, M., Yoshino, K., Matsuura, T., Sekino, T., Saito, S., and Tomita, I., Chem. Pharm. Bull., 1986, 34, 174. Alvarez-Pino, M. I., Vinagre Leiro, J., and Schmidt-Hebbel, H . , Grasas Aceites (Seville), 1969, 20, 129. Davidek, J., and Pokorny, J., Z. Lebensm.-Unters.-Forsch., 1961, 115, 113. Copius-Peereboom, J.W., Nature (London), 1964, 204, 748. Ishikawa, S., and Katsui, G., Bitamin, 1964, 29, 203. Freimuth, U., Foerster, K., and Ludwig, E., Nahrung, 1965,9, 569. Wzng, R. T., and Chou, S . S., J . Chromatogr., 1969,43, 522. Lehmann, G., and Moran, M., Z. Lebensm.-Unters. -Forsch., 1971, 145, 344. Mancini, P., Delton, V., and Gelsomini, N., Rass. Chim., 1977,29, 107. Salo, T., and Salminen, K., Z. Lebensm.-Unters.-Forsch., 1964, 125, 167. Mathew, T. V., and Mitra, S. N., Zndian J . Technol., 1965,3, 102. Wasa, S., Kawaida, T., Murakami, K., Hitaka, M., Morishita, S . , and Hosoda, K., Shokuhin Eisi Kenkyu, 1985, 35, 921. Wei, R., Song, F., and Gao, H., Shipin Yu Fajiao Gongye, 1985, ( 9 , 32. Sahasrabudhe, M. R., J. Assoc. Off. Agric. Chem., 1964, 47, 888. Salo, T., Makinen,-R., and Salminen, K., Z.Lebensm.- Unters.-Forsch., 1964, 125, 450. Alary, J., Grosset, C., and Coeur, A . , Ann. Pharm. Fr., 1982, 40, 301. Vioque, E., Yoshiro, A., and Martel, J., Grasas Aceites (Seville), 1967, 18, 310. Scheidt, S. A., and Conroy, H. W., J . Assoc. Off. Anal. Chem., 1966,49,807. Davidek, J., Janicek, G., and Davidkova, E., Z. Lebensm.- Unters.-Forsch., 1966, 131, 29. Kramer, R. E., J. Am. Oil Chem. SOC., 1985,62, 111. Pujol Forn, M., and Garrabou Goma, M., Circ. Farm., 1976, 34,419. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. Davidek, J., Janicek, G., and Davidkova, E., Z. Lebensm.- Unters.-Forsch., 1967, 131, 345.Glavind, J., and Holmer, G., J . Am. Oil Chem. SOC., 1967,44, 539. Bajaj, K. L., and Ahuja, K. L., Analusis, 1986, 14, 161. Martelli, A., and Nano, G. M., Farmaco, Ed. Prat., 1967, 22, 660. Woggon, H., Uhde, W. J., and Zydek, G., Z. Lebensm.- Unters.-Forsch., 1968, 138, 169. Vilicic, D., Lui, B., and Mihelic, F., Kem. Ind., 1968, 17, 731. van Dessel, L., and Clement, J., Z. Lebensm.-Unters.-Forsch., 1969, 139, 146. Pujol Forn, M., Grasas Aceites (Seville), 1980, 31, 187. Phipps, A. M., J. Am. Oil Chem. SOC., 1973, 50, 21. Guven, K. C., and Guven, N., Eczacilik Bul., 1973, 15, 46. Bajaj, I., Verma, K. K . , Prakash, O., and Parihar, D. B., J . Chromatogr., 1970,46,261. Guldborg, M., Fresenim Z. Anal. Chem., 1981, 309, 117. Valdehita, M.T., and Vincente, M. C., An. Bromatol., 1969-70 (Pub. 1971), 23, 107. Gertz, C., and Herrmann, K., Z. Lebensm.-Unters.-Forsch., 1983, 177, 186. Nakazato, M., Kanmuri, M., Ariga, T., Fujinuma, K., and Naoi, Y . , Shokuhin Eiseigaku Zasshi, 1980, 21, 64. Nakazato, M., Kanmuri, M., Ariga, T., Fujinuma, K., and Naoi, Y., Shokuhin Eiseigaku Zasshi, 1981, 22, 239. Tessier, A. M., Linard, A., Delaveau, P., Hoffelt, J., and Gaugain, B., Z. Lebensm.-Unters.-Forsch., 1983, 176, 12. Winkelmann, W., and Montag, A., Lebensmittelchem. Gerichtl. Chem., 1977, 31, 87. Laub, E., and Woller, R., Dtsch. Lebensm.-Rundsch., 1976, 72, 276. van der Heide, R. F., J . Chromatogr., 1966, 24,239. Anet, E. F., J . Chromatogr., 1971, 63,465. Chang, W. H., Luu, H. X., and Cheng, A.C., J . Food Sci., 1983, 48, 658. Jayaraman, S., Vasundhara, T. S . , and Parihar, D. B., Mikrochim. Acta, 1976,II, 365. Wyatt, D. M., J. Am. Oil Chem. SOC., 1981, 58, 917. Ito, Y., Toyoda, M., Ogawa, S . , and Iwaida, M., Eisei Kagaku, 1978, 24, 338. Isshiki, K., Tsumura, S., and Watanabe, T., Agric. Biol. Chem., 1980,44, 1601. Maruyama, T., Niiya, I., and Imamura, M., Shokuhin Eisei- gaku Zasshi, 1977, 18,283. Senten, J. R., Waumans, J. M., and Clement, J. M., J . Assoc. Off. Anal. Chem., 1977, 60, 505. Dobarganes Garcia, M. C., and Perez Camino, M. C., Grasas Aceites (Seville), 1983, 34, 183. Nishimoto, T., and Uyeta, M., Shokuhin Eiseigaku Zasshi, 1964,5, 287. Sakurai, K., Yamada, T., Mori, M., Masuda, N., Mori, Y . , and Sudo, T., Mie-kenritsu Eisei Kenkyusho Nenpo, 1979,25, 27; Chem.Abstr., 1985, 95, 167201q. Schwien, W. G., Miller, B. J., and Conroy, H. W., J. Assoc. Off. Agric. Chem., 1966,49, 809. Hayakawa, J., Narafu, T., Takahashi, H., Ishida, Y., and Tsuiki, T., Shokuhin Eiseigaku Zasshi, 1969, 10, 190. Hartman, K. T., and Rose, L. C., J . Am. Oil Chem. SOC., 1970, 47, 7. Stoddard, E. E . , J . Assoc. Off. Anal. Chem., 1972, 55, 1081. Min, D. B., Ticknor, D., and Schweizer, D., J. Am. Oil Chem. SOC., 1982, 59, 378. Yu, L. Z., Inoko, M., and Matsuno, T., J. Agric. Food Chem., 1984, 32, 681. Mariani, C., and Fedeli, E., Riv. Ital. Sostanze Grasse, 1983, 60, 667. Takeba, K., and Matsumoto, M., Eisei Kagaku, 1983,29,329. Fukushima, S., Kashimoto, T., Koyama, K., Konisi, K., and Kobayashi, T., Osaka Furitsu Koshu Eisei Kenkyusho Kenkyu Hokoku, Shokuhin Eisei Hen, 1979,9,19; Chem.Abstr., 1979, 91, 191476~. Greenberg, M. J., Hoholick, J., Robinson, R., Kubis, K., Groce, J., and Weber, L., J . Food Sci., 1984, 49, 1622. Yamamoto, M., Ochi, S . , Konagai, I., Komine, S., Ishikawa, M., Narita, H., Masui, T., and Kitada, Y., Shokuhin Eiseigaku Zasshi, 1985, 26, 285. Toyoda, M., Ogawa, S . , Tonogai, Y., Ito, Y., and Iwaida, M., J . Assoc. Off. Anal. Chem., 1980,63, 1135.ANALYST, JULY 1987, VOL. 112 943 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. Anderson, R. H., and Nelson, J . P., Food Technol., 1963,17, 915. Schwecke, W. M., and Nelson, J.H., J. Agric. Food Chem., 1964, 12, 86. Takahashi, D. M., J. Assoc. Off. Agric. Chem., 1965,48, 694. Takahashi, D. M., J . Assoc. Off. Agric. Chem., 1966,49, 704. Takahashi, D. M., J. Assoc. Off. Anal. Chem., 1967,50,880. J . Assoc. Off. Anal. Chem., 1968, 51, 490. Takahashi, D. M., J . Assoc. Off. Anal. Chem., 1970, 53, 39. Singh, J., and Lapointe, M. R., J . Assoc. Off. Anal. Chem., 1974, 57, 804. Choy, T. K., Quattrone, J . J . , and Alicino, N. J., J . Chromatogr., 1963, 12, 171. Takahashi, D. M., J . Assoc. Off. Anal. Chem., 1968, 51, 943. Yamamoto, M., Gijutsu Joho-Shizuoka-ken Eisei Kankyo Senta, 1984, 2, 1; Chem. Abstr., 1985, 102, 183784a. Dilli, S., and Robards, K., unpublished data. Kline, D. A., Joe, F. L., and Fazio, T., J . Assoc. Off. Anal. Chem., 1978, 61, 513.Fry, J. C., Leatherhead, 1982, 765. Anal. Chem., 1983, 55, 165R. Doering, C. E., Estel, D., Pehle, W., Gaikowski, M., and Seiffarth, K., J . Chromatogr., 1985, 348, 430. Galensa, R., and Schaefers, F. I., Dtsch. Lebensm.-Rundsch., 1982, 78,258. Maruyarna, T., Murakami, C., Kanematsu, H., and Niiya, I . , Shokuhin Eiseigaku Zasshi, 1981, 22, 64. Page, B. D., and Kennedy, B. P., J . Assoc. Off. Anal. Chem., 1976, 59, 1208. Dilli, S., and Robards, K., J. Chromatogr., 1977, 133, 363. Buxtorf, U. P., and Pavesic, V., Z. Lebensm.-Unters.-Forsch., 1977, 163, 104. Mizutani, T., Tajima, K., Okino, N., and Yamamoto, K., J . Chromatogr., 1985,333, 171. Wachs, W., and Gassmann, L., Dtsch. Lebensm.-Rundsch., 1970,66, 37. Kato, S., Shimoda, M., and Sasaki, T., Eisei Shikenjo Hokoku, 1968, 87, 24.Bajardi, M. L., Giannola, L. I., and Giammona, G., Riv. SOC. Ztal. Sci. Aliment., 1982, 11, 105. Dilli, S., and Kwan, M. Y. H., unpublished data. Takemura, I., Bunseki Kagaku, 1970, 20, 61. Fazio, T., Howard, J. W., and Sandoval, A., J . Assoc. Off. Anal. Chem., 1968, 51, 17. Austin, R. E., and Wyatt, D. M., J. Am. Oil Chem. SOC., 1980, 57, 422. Done, J. N., Kennedy, G. J., and Knox, J . H., Nature (London), 1972, 237, 77. Pellerin, F., Delaveau, P., Dumitrescu, D., and Safta, F., Ann. Pharm. Fr., 1982,40, 221. Okamoto, T., and Suezawa, Y . , Kagawa-Ken Hakko Shokuhin Shikenjo Hokoku, 1983, 76, 79; Chem. Abstr., 1985, 102, 22959~. Archer, A. W., Anal. Chim. Acta, 1981, 128, 235. Masoud, A. N., and Cha, Y. N., J . High Resolut. Chromatogr. Chromatogr. Commun., 1982, 5, 299. Harnrnond, K., J. Assoc. Public Anal., 1978, 16, 17. King, W. P., Joseph, K. T., and Kissinger, P. T., J. Assoc. Off. Anal. Chem., 1980,63, 137. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. Page, D. B., J . Assoc. Off. Anal. Chem., 1979, 62, 1239. Bocca, A., and Delise, M., Riv. SOC. Ital. Sci. Aliment., 1982, 11, 345. Berridge, J. C., Kent, J., and Norcott, K. M., J . Chromatogr., 1984, 285, 389. Orsi, F., and Abraham Szabo, A., Elelmiszervizsgalati Kozl., 1985,31,78; Chem. Abstr., 1985, 103, 177056~. Van Niekerk, P. J., and Du Plessis, L. M., J . Chromatogr., 1980, 187,436. Kitada, Y., Tamase, K., Mizobuchi, M., Sasaki, M., Tani- gawa, K., Yamamoto, M., and Nakazawa, H., Bunseki Kagaku, 1984,33, E33. Ansari, G. A., J. Chromatogr., 1983, 262, 393. Indyk, H., and Woolard, D. C., J . Chromatogr., 1986, 356, 401. Doeden, W. G., Bowers, R. H., and Ingala, A. C., J . Am. Oil Chem. SOC., 1979, 56, 12. Pokorny, S., Coupek, J., and Pokorny, J., J . Chromatogr., 1972, 71, 576. Niebergall, H., and Hartmann, M., Lebensm. -Wiss. Technol., 1983, 17,254. Lichtenthaler, R. G., and Ranfelt, F., J . Chromatogr., 1978, 149, 553. Ward, G. A., Talanta, 1963, 10, 261. Franzke, C., Kretzschmann, F., and Beining, K., Fette Seifen Anstrichm., 1968, 70, 472. McBride, H. D., and Evans, D. H., Anal. Chem., 1973, 45, 446. Paris, J. P., Gorsuch, J. D., and Hercules,D. M., Anal. Chem., 1964,36, 1332. Brieskorn, C. H., and Mahlmeister, K., Z . Lebensm.-Unters.- Forsch., 1980, 71, 348. Thunyaudom Tonmanee, N., and Archer, V. S . , Talanta, 1982, 29, 905. Wang, J., and Freiha, B. A., Anal. Chim. Acta, 1983, 154, 87. Sedlacek, B. A., Vopr. Pitan., 1964, 23, 8. Karpukhin, 0. N., Shlyapintokh, V. Ya., Rusina, J. T., and Zolotova, N. V., Zh. Anal. Khim., 1963, 18, 1021. Latz, H. W., and Hurtubise, R. J., J. Agric. Food Chem., 1969, 17, 352. Hurtubise, R. J., and Latz, H. W., J. Agric. Food Chem., 1970, 18, 377. Dilli, S., and Robards, K., Analyst, 1977, 102, 201. Rolea, G., and Tataru, M. S . , Rev. Chim. (Bucharest), 1981, 32, 912; Chem. Abstr., 1982, 97, 84401~. Boughton, 0. D., Bryant, R., and Combs, C. M., J. Agric. Food Chem., 1967, 15,751. Miller, H. E., J. Am. Oil Chem. SOC., 1971, 48, 91. Bors, W. , Michel, C., and Saran, M., Biochim. Biophys. Acta, 1984, 796, 312. Wenger, F., Mitt. Geb. Lebensmittelunters. Hyg., 1954, 45, 185. Paper A61296 Received August 22nd, 1986 Accepted December 30th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200933

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 945 Improved Group Determination of Total N-Nitroso Compounds in Human Gastric Juice by Chemical Denitrosation and Thermal Energy Analysis* Brigitte Pignatelli, Isabelle Richard, Marie Christine Bourgade and Helmut Bartscht Unit of Environmental Carcinogens and Host Factors, International Agency for Research on Cancer, 150 cows Albert Thomas, 69372 L yon, France A procedure for the determination of total N-nitroso compounds (NOC) in human gastric juice was developed by modifying earlier methods. The gastric juice sample, treated with sulphamic acid to remove nitrite, is injected directly into refluxing ethyl acetate containing either acetic acid for determining thermo- and acetic acid labile thermal energy analyser (TEA) responsive compounds (TAC), or into hydrogen bromide for the determination of TAC and NOC.The nitric oxide (NO) levels released are measured using thermal energy analysis with chemiluminescence detection, and the difference between the two determinations represents the concentrations of NOC in gastric juice. This method is not affected by nitrate concentrations of up to 1000 pmol I-’. The method was found to be rapid, reproducible and sensitive (detection limit 0.02 ymol I-’ NOC), requiring only small volumes of gastric juice and no prior extraction. Because the difficulties arising from the system response to the denitrosating agent and variability of NO release by acetic acid from nitrite were eliminated, this improved method can more accurately distinguish NOC from most other TEA-responsive species.Suitable techniques for stabilising gastric juice samples from duodenal ulcer and atrophic gastritis patients and the influence of the time and storage conditions on NOC concentrations have been studied. Keywords: N- Nitroso compound determination; gastric juice analysis; chemical denitrosation; thermal energy analysis; gastric juice stabilisation and storage Carcinogenic N-nitroso compounds (NOC) formed in the stomach have been suggested as causative agents in the pathogenesis of gastric cancer.2J According to the best known hypothesis, originally put forward by Correa et u Z . , ~ gastric hypochlorohydria, by permitting the colonisation of the stomach by bacteria, including nitrate-reducing species, results in increased concentrations of nitrite which yield NOC.A positive correlation between high intragastric pH and both bacterial count and nitrite concentrations has been found in most of the reported studies.4-7 However, recent applications of the two available methods for group determination of total NOCSlO gave rise to contradictory findings, and the reported concentrations of NOC in human gastric juice showed large inter-laboratory variations.47J1-15 Both methods have been subject to criticism.1o316J7 The group-selective procedure for the determination of total NOC899 is based on their chemical denitrosation with hydrogen bromide and chemiluminescence detection of the released nitric oxide (NO). Sequential addition of acetic acid and hydrogen bromide allows distinction between NOC and most of the other potentially NO-releasing substances.Because this technique has some limitations, modified proce- dures have been proposed.10J4 Although they permit the direct analysis of aqueous samples, other interfering NO- releasing compounds cannot be distinguished from NOC. Our aim was to develop a more reliable procedure for determining total NOC in human gastric juice. The results obtained from validation studies and some applications of this procedure are presented in this paper. Experimental Reagents and Standards All reagents used were of analytical-reagent grade. Ethyl acetate was treated with 20% mlV sulphamic acid with * Presented in part at the Ninth International Meeting on N-Nitroso Compounds, Baden, Austria, 1-5 September 1986. t To whom correspondence should be addressed.repeated shakings, kept for a minimum of 3 d and filtered just before use. Hydrogen bromide was prepared as described previously.8 The following NOC standards were purchased from East- man, Fluka, Aldrich and the Chemical Carcinogen standard reference repository of the National Cancer Institute: N-nitro- soproline (NPRO), N-nitrosodiethylamine (NDEA), N-nitro- somethylpentylamine (NMPA), N-nitrosodipropylamine (NDPA) , N-nitrosodiethanolamine (NDELA), N-methy1-N’- nitro-N-nitrosoguanidine (MNNG), N-propyl-N’-nitro-N- nitrosoguanidine (PNNG), N-methyl-N-nitrosourea (MNU), N-methyl-N-nitrosourethane (MNUT), N-propyl-N-nitroso- urethane (PNIrT). The standard aqueous NOC solutions (1-3 pmol l-1) were freshly prepared and their concentrations were verified by the method described under Determination of Total NOC using Procedures A (no decomposition) and B.Aliquots (10-20 yl) of these standard NOC solutions were subsequently used for spiking. Caution-As all of the nitrosamine standards mentioned above, except NPRO, are potent carcinogens, extreme caution should be taken when working with or handling the chemicals. Contact with the skin or inhalation of vapour must be avoided. Determination of Total NOC The operating conditions, glassware assembly and reagents were the same as described previously,g except for the following three modifications. (i) A gas-stream filter (CTR, Thermo Electron, Waltham, MA, USA) was added between the last cold trap and the thermal energy analyser (TEA) input in order to improve the selective passage of NO; (ii) in the assembly for Procedure A (see below), a dry caustic soda trap replaced the three successive traps containing 6 M sodium hydroxide used for procedure B; (iii) because of the variable TEA response of the nitrite standard (observed after its decomposition by acetic acid alone), 0.1% V/V hydrochloric acid was added, which led to excellent reproducibility and sensitivity without further decomposition of NOC [Fig.l(a)]. The NO chemiluminescence detectors were TEA Models 502946 ANALYST, JULY 1987, VOL. 112 (Procedure B) and 543 (Procedure A) (Thermo Electron) connected to Model 3390 A integrators (Hewlett Packard). The gastric juice sample was treated with sulphamic acid or hydrazine sulphate, as described below, to destroy nitrite and analysed by two parallel procedures: (1) Procedure A.The gastric juice sample was injected directly into refluxing ethyl acetate containing acetic acid - 0.1% VlV hydrochloric acid and the thermo- and acetic acid labile TEA-responsive compounds (TAC) were determined. (2) Procedure B. The gastric juice sample was injected directly into refluxing ethyl acetate containing hydrogen bromide (15% mlV in glacial acetic acid) and both TAC and NOC were determined simultaneously. The difference between the two procedures (B - A) represents the concentration of NOC in gastric juice. The NO released was measured by chemiluminescence using the TEA. The concentrations of TAC and TAC + NOC were calculated from the ratio of the peak area of the sample to that of a standard sodium nitrite solution and a standard NDPA solution, respectively.The sample must be vigorously homogenised (Vortex) just before its injection for both procedures. Conditions to Prevent Artefactual Nitrosation Three known inhibitors of N-nitrosation, namely hydrazine sulphate, ascorbic acid and sulphamic acid, were compared for their efficiency as nitrite scavengers. Aqueous sodium nitrite solutions at concentrations ranging from 21.7 to 434 pmol l-1 were treated as follows: 2% mlV of sulphamic acid, pH 1; 0.3 ml of 3.4% m/V hydrazine sulphate solution (adjusted to pH 4) per ml with or without 1% mlV phthalate, pH 4; 1 YO m/V of ascorbic acid, pH 2.9. The remaining nitrite in the aqueous solution was determined by thermal energy analysis as a function of reaction time.Preparation of Samples A total of 73 gastric juice samples, obtained from patients before and after operations for duodenal ulcer (n = 64) or with chronic atrophic gastritis (n = 9) were analysed for NOC. Because of the small sample volumes, some gastric juices with similar pH were pooled. After collection, the gastric juices were frozen and analysed within 3 d. After thawing, they were vigorously homogenised (Vortex) and, when necessary, centrifuged. The samples were then divided into two-one part was treated with sulphamic acid and the other with hydrazine sulphate, under the conditions described previously. For each sample, these two treatments were carried out in parallel. Aliquots were subsequently transferred into stoppered glass tubes with Teflon joints to allow spiking with sodium nitrite (200-230 pmol l-1) or with 2,6-dimethylmorpholine (130-150 pmol l-1) or with standard NOC (6-25 pmol l-1).Multiple preparations were carried out for each assay condition. Aliquots not analysed at once were immediately frozen at -20 "C. Results Determination of Total NOC in Human Gastric Juice A typical example of the analysis of gastric juice for total NOC is shown in Fig. 1. Equimolar solutions of nitrite and NDPA led to identical TEA responses after denitrosation by hydrogen bromide. Therefore, to determine the concentration of NOC, the amount of TAC calculated from nitrite as external standard [Fig. l(a)] can be subtracted from the TEA response arising from denitrosation by hydrogen bromide [Fig. l(b)]. Both standard solutions gave reproducible values for the peak areas, with a coefficient of variation of 2%.The presence of water in ethyl acetate (3-5% VIV) or of sodium nitrate up to 1000 pmol 1-1, with or without treatment by sulphamic acid, and up to one month's storage at -20 "C, did (a) Sodium nitrite, 175 pmol I-' (10 pl) Sodium nitrite, 175 pmol 1-1 (10 pi) I (6) N-Nitrosodipropylamine, N-Nitrosodipropylamine, Hydrogen 200 pmol I-' (5 200 pmol 1-1 (5 pi) bromide 0.91 pmol 1-1 1 Gastric juice (500 pl), Gastric juice (500 pl) / Fid(2; in acetic & 1 duplicate, 0.91 pmol 1-1 Water (500 pI) Fig. 1. Analysis of gastric juice for total N-nitroso compounds (NOC), showing thermal energy analyser (TEA) response. Thermolabile acetic acid TEA responsive compounds (TAC); and ${ TAC + NOC not affect the determination.Higher concentrations of nitrate, although never found in gastric juice samples, led to a false positive TEA response. It was noted that the first injection of water (500 pl) gave a variable detector response (Fig. l), but subsequent injections did not give a similar response at any stage in either analytical procedure. The concentration of nitrite in gastric juice can also be determined by the difference between TEA measurements following injection of gastric juice with and without sulphamic acid treatment according to Procedure A. Different conditions and nitrite scavengers were used to stabilise gastric juice samples in the two available methods for determination of total NOC.i+-1° As sulphamic acid requires a low pH to avoid artefactual nitrosation,8.9 Bavin et aZ.10 claimed that this condition would lead to the denitrosation of some NOC; they recommended the use of hydrazine sulphate at pH 4, a pH value anticipated to stabilise NOC more efficiently. Ascorbic acid was reported to inhibit nitrosation from pH 1 to 4 by reducing nitrite to nitric oxide (NO) but NO in the presence of oxygen could regenerate some of the nitrite.18 A comparison of the efficiency of these three nitrite scavengers applied under the usual pH conditions is illustrated in Table 1.Sulphamic acid was the most effective agent for the removal of nitrite at pH <2. At pH 4, hydrazine sulphate did not react with nitrite as quickly as sulphamic acid; however, when a hydrazine sulphate solution (see under Experimental) with a pH of 1.48 was not adjusted to pH 4 prior to addition to the nitrite solution, the pH fell to 2 and at least 2-3 min were necessary to establish a pH of 4.At pH 2, hydrazine sulphate was as efficient as sulphamic acid in removing nitrite. The above observations could explain why hydrazine sulphate has been reported to be an efficient nitrite trap at pH 4.10 Ascorbic acid is also less effective than sulphamic acid in the removal of nitrite. After the treatment of a nitrite solution (220 pmol l-1) with sulphamic acid (as described under Experimental) and storage at -20 "C for up to 6 weeks, no nitrite was detected. When sodium nitrite (final concentration, 229 pmol 1-1) was reacted with 2,6-dimethylmorpholine (final concentration, 162 pmol 1-1) containing sulphamic acid, no nitrosamine was detectable after 5 min of reaction and after storage at -20 "C for up to one month.Suitability of the Method In order to verify the suitability of this analytical method the efficiency of sulphamic acid and hydrazine sulphate as nitrite scavengers was investigated in nitrite-spiked gastric juice samples and the stability of NOC following these two treatments and the influence of time and storage conditions on NOC concentrations were examined. Artefactual N-nitrosa-ANALYST, JULY 1987, VOL. 112 947 Table 1. Comparison of the efficiency of three nitrite scavengers. Results presented are percentage nitrite recovery Reaction time/min NaN02/ pmoll-1 Treatment* 3 5 10 15 25 35 50 60 70 434 SA - 0.05 NDT HS 83 75 62 46 27 - 7 4 2 HS ' 75 55 30 15 4 1 0.2 0.06 0.01 AA HS - 60 40 25 13 5 2 - ND HS ' - 53 30 13 4 1 - ND AA - 5 2 - 1 - - ND - - 17 10 3 0.1 ND HS AA 11 4 2 - 0.5 ND - 1 - - - 23 6 3 - 43.4 SA ND 21.7 SA - ND * SA, sulphamic acid, 2% mlVpH 1; HS, hydrazine sulphate, 0.3 ml of 3.4% mlVsolution per ml, pH 4; HS', 0.3 ml Of 3.4% mlVsolution per t ND: not detectable, [NaN02] < 0.02 pmol 1-I.ml, 1% m/V phthalate, pH 4; AA, ascorbic acid 1% mlV pH 2.9. ~ ~~ ~~ Table 2. Recovery of N-nitroso compounds (NOC) in gastric juice as a function of storage time at - 20 "C. Results presented are mean f SD percentage recovery of NOC, calculated by taking as 100% the measurement of NOC concentration in 2 ml of water spiked with 10 p1 of freshly prepared NOC solution (to simulate a gastric juice sample). The amount of NOC measured in each unspiked sample was subtracted from values obtained with the the spiked sample after the same storage time.For unspiked gastric juice, the first determination of NOC was taken as 100% Storage timeld NOC* added to gastric juice None . . . . . . . . NPRO . . . . . . . . NA . . . , . . . . MNNG . . . . . . . . MNU . . . . . . . . MNUT . . . . . . . . Concentration/ pmoll-1 - . . . . 14 . . 12-16.4 . . 11-14.4 . . 12.7-22 6-24.7 PH range 3.3-8.9 1.2-3.0 1.5-5.9 1.6-6.3 6.7-7.9 1.5-1.7 3.3-5.9 1.2-2.1 6.2-8 .O 1.2-3.0 n 12 18 3 5 4 4 3 3 3 3 5 5 6 3 10 - Treatmentt SA SA SA SA SA HS SA HS SA HS SA HS SA HS SA SA* HS 0-1 - - 99.0 5 1.0 97.8 2 2.2 94.7 k 6.7 88.7 f 13.8 87.7 f 3.8 98.3 f 2.9 99.0 2 1.4 90.5 k 13.4 89.6 f 6.9 96.0 f 5.8 88.0 f 12.2 94.0 f 1.41 67.2 f 21.6 48.7 f 12.0 91.7 f 5.6 15-19 70.4 k 21.5 67.7 f 29.9 96.3 f 1.5 92.4 k 3.9 73.5 k 5.2 28.5 f 18.0 21.7 k 13.2 11.0 k 3.5 95.0 f 7.1 86.5 f 7.8 40.8 L 27.7 86.6 k 5.8 76.4 f 13.1 85.0 f 5.7 15.0 L 17.1 42.0 k 10.2 81.2 k 11.8 30-34 60.4 f 25.9 61.2 f 33.9 94.3 f 1.5 91.2 k 2.9 57.0 f 6.3 15.5 k 15.9 8.3 f 6.8 8.0 f 6.1 91.5 f 2.1 83.0 k 2.8 22.8 k 25.3 69.6 rt 14.0 70.4 f 15.3 61.5 f 19.1 13.2 k 14.1 41.5 f 10.5 61.6 f 31.8 * NPRO, N-nitrosoproline; NA, N-nitrosodiethylamine or N-nitrosomethylpentylamine; MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; MNU, N-methyl-N-nitrosourea; MNUT, N-methyl-N-nitrosourethane. The NOC solutions were freshly prepared and their concentrations were checked.1- SA, sulphamic acid, 2% mlV, pH 1; SA*, sulphamic acid, 2% mlV, pH 1 during 5-10 min, then adjustment to pH 4; HS, hydrazine sulphate 0.3 ml of 3.4% mlV solution per ml, pH 4.tion was verified after spiking gastric juice samples with 2,6-dimethylmorpholine. For the analysis of gastric juice samples treated with sulphamic acid, the coefficients of variation from duplicate or triplicate TEA measurements (Fig. 1) were 5-10%. The limit of detection was 0.02 pmol l-1. When the gastric juice samples were treated with hydrazine sulphate, the same analytical procedure led to broader peaks with more tailing, especially from NO released by TAC. In addition, the TEA response of standard nitrite [Fig. l(a)] was reduced by up to 50% after the first sample injection, and the response did not improve after the addition of more acetic acid - 0.1 % V/V hydrochloric acid.As a result, the latter measurement was less accurate and duplicates could not be determined. After stabilisation of the gastric juice samples with sulphamic acid, the addition to the samples of nitrite (n = 27) or 2,6-dimethylmorpholine (n = 20) did not produce any artefactual nitrosation. After treat- ment by hydrazine sulphate at pH 4, no additional nitrosation of added 2,6-dimethylmorpholine was found ( n = S), but artefacts (i.e., remaining nitrite and formed NOC) were observed in all samples spiked with nitrite (n = 20). The stability of aqueous solutions of nine reference NOC (13-26 pmol 1-1) was measured following treatment by sulphamic acid or hydrazine sulphate, after storage at 4 "C for 1 h or at -20 "C for one month.The stabilities under these storage conditions and after both treatments decreased similarly in the following order (recoveries 60-99%): NPRO = NDEA = NMPA 2: NDELA > PNNG = MNU >> PNUT > MNUT. As an exception (after one month's storage at -20 "C), 90% MNNG was recovered following sulphamic acid treatment , but none after hydrazine sulphate treatment. It was noted that the stability of homologues of N-nitrosoguanidines or N-nitrosourethanes increased with increasing alkyl chain length, as observed by Haga et al. 19 Compounds that are more susceptible to decomposition (MNU, NMUT, MNNG) and NDEA, NMPA and NPRO were selected for further studies in stabilised gastric juice samples (Table 2) to which each of these compounds had been added.When reference NOC were added to gastric juice and determined within 24 h of storage at -20 "C, the recoveries were all 7&99% (Table 2). In gastric juice samples that were treated with hydrazine sulphate, MNU and MNUT were fairly948 ANALYST, JULY 1987, VOL. 112 ~~ Table 3. Concentrations of thermo- and acetic acid labile thermal energy analyser-responsive compounds (TAC) and total N-nitroso compounds (NOC) in gastric juice samples as a function of pH, and effect of TAC on measurement of NOC. Samples were collected from duodenal ulcer and atrophic gastritis patients. ND = <0.02 pmol 1-l (detection limit) Effect of Samples TAC on pmoll-1 pmoll-1 pmoll-1 TAC, O/o of NOC, TAC/ TAC + NOC/ NOC/ containing measurement No.of mean f SD mean k SD mean k SD (no. of TAC + NOC ) pH range samples (range) (range) (range) samples) ( NOC 1-1.5 11 0.33 k 0.39 1.6-2.3 25 0.39 f 0.70 2.5-3.3 10 0.22 k 0.46 3.6-5.0 4 0.95 f 1.13 5.7-7.0 11 0.88 k 1.86 7.2-8.9 12 0.55 k 1.13 (ND-4.0) (ND-1.18) (ND-2.20) (ND-1.50) (ND-2.20) (ND-6.40) 1.01 k 0.69 (0.19-2.19) (0.85 f 1.22) 0.60 k 1.13 (0.08-3.75) 2.37 k 1.72 2.16 k 3.49 1.00 k 1.16 (ND-4.0) (ND-5.40) (O.Ofj-4.02) (0.05-1 2.4) 0.68 f 0.48 (0.10-1.69) (0.46 k 0.74) (0-3.6) 0.38 f 0.68 (0-2.25) 1.41 k 0.92 (0.06-2.13) 1.28 f 1.71 0.46 k 0.59 (0.05-6.0) (0-2.0) 55 1-4 36 50 50 1-6.5 (6) (9) ( 5 ) 1-2.4 1-1.2 1-8 (2) 64 1-2 (7) (7) 58 N-N itrosod ipropylami ne (200 pmol I-') t f L Gastric juice+ N-methyl-N-nitrosourethane (24.7 pmol I-') Fig.2. Thermal energy analyser response after denitrosation by hydrogen bromide of a gastric juice spiked with N-methyl-N-nitroso- urethane after treatment with (a) sulphamic acid and ( b ) hydrazine sulphate stable, but MNNG was largely decomposed after storage for 15 d at -20 "C. After treatment with sulphamic acid, the TEA peaks were sharper (Fig. 2) and MNNG, MNU and MNUT were reasonably stable in gastric juice samples that had initially a neutral or basic pH. However, MNU was much less stable and MNNG and MNUT were unstable in samples with an initially acidic pH. The adjustment to pH 4 after sulphamic acid treatment and before storage prolonged the stability of MNUT. Remarkably, the unidentified NOC measured as a group in our gastric juice samples ( n = 30) treated with sulphamic acid were less stable than nitrosoamino acids or nitrosamines but more stable than N-nitrosoureas or N-nitro- sourethanes when stored at -20 "C for up to one month.Their recoveries were not dependent on the initial pH of gastric juice samples. Analysis of Human Gastric Juice Samples The means and ranges of TEA responses for the determined TAC and calculated NOC concentrations in the gastric juice samples analysed (stabilised by sulphamic acid) are given in Table 3 for several pH ranges. Although the sample size was small, the following conclusions can be drawn. The concentra- tions (means) of NOC were not significantly different in gastric juice samples with extreme acidic or basic pH (0.6 versus 0.5 ymol 1-1, respectively); in contrast, samples at pH 3.6-7.0 had higher (mean) NOC concentrations (1.3 pmol l-1).The concentrations of TAC were as pH-dependent as the NOC, and the TAC concentrations profoundly influ- enced the total TEA-responsive compounds as the ratio [NOC] : [TAC] was varied for individual samples from 0.2 to 5. Discussion A progressive increase of NOC levels in gastric juice with increasing pH has been reported previously.6~7J5 These results support the hypothesis according to which gastric hypo- chlorohydria permits bacterial overgrowth leading to increased nitrite formation, NOC formation and, hence, an elevated risk of gastric cancer. Other studies, however, failed to show any significant variation between levels of NOC in gastric juice and pH.5.11 Moreover, much lower concentra- tions of total NOC in hypoacidic as compared to acidic gastric juices were obtained using the methodology of Bavin et aZ.10 In these latter studies, the absolute values were up to five times greater than those measured by the method of Walters and co-workers.g.9 In our study, intermediate values for NOC concentration (mean and ranges) were obtained that were closer to those reported by users of Walters' method.The latter procedure requires the extraction of NOC in ethyl acetate (probably with poor yields of some NOC), and the acidic pH (<2) precludes the extraction of basic NOC from gastric juice. The method of Bavin et aZ.10 may overestimate NOC in gastric juices, because TAC are not distinguished from NOC (Table 3). The existence in the 73 gastric juice samples of an optimum NOC concentration at pH 3.6-7.0 (Table 3) may result from two counteracting pH-dependent variables: (i) the available nitrite concentration that increases with rising pH because of higher numbers of nitrate-reducing bacteria in the stomach of achlorohydric subjects and increased nitrite stability; and (ii) the rate of acid-catalysed nitrosation that decreases with elevated pH.The limitations of the sequential method devised by Walters and co-workersg>g come from substantial contributions of the system (i. e., hydrogen bromide itself, trace amounts of NOC impurities in the reagents, the interaction of hydrogen bromide with glassware and connecting tubing of the appa- ratus) to the detector response. Corrective procedures and calculations have been employed.gY20 Because the difficulties arising from the system response and variability of NO release from nitrite by acetic acid were eliminated, our method is more accurate and more sensitive.ANALYST, JULY 1987, VOL.112 949 In conclusion, our findings (Table 3) in general confirm the criticism made by Smith et al. 17 about modified versions10J4 of Walters’ proced~re.8~9 The magnitude of the false positive response from heat-labile (e.g. , pseudonitrosites, nitroso- thiols) and acid-sensitive compounds (e.g., alkyl nitrites) can be very important. In addition, a false positive response may arise from the very first injection of gastric juice without prior injection of water (Fig. 1). In our method, only nitrolic acids, nitrothiols and some aliphatic C-nitroso compounds can give a response under the same conditions as NOC.1517 Our improved method for determining total NOC in gastric juice maintains a great selectivity for NOC and offers additional advantages. It is hoped that its application can resolve current controversies concerning the role of intragastrically formed NOC in gastric cancer.The authors thank Dr. R. Lambert, Dr. B. Moulinier, Dr. Y. Minaire, Dr. J. Forichon and their medical staff at Edouard- Herriot Hospital, Lyon, France, Professor M. Crespi at Regina Elena Institute, Rome, Italy for providing gastric juice samples, Ms. E. Heseltine for editorial help and Ms. M. Wrisez for secretarial assistance. 1. 2. 3. 4. 5 . 6. References Pignatelli, B., Richard, I., Bourgade, M. C., and Bartsch, H., IARC Int.Agency Res. Cancer Sci. Publ., 1987, No. 84, in the press. Correa, P., Haenszel, W., Cuello, C., Tannenbaum, S . , and Archer, M., Lancet, 1975, ii, 58. Mirvish, S. S., J . Natl. Cancer Inst., 1983, 71, 631. Hall, C. N., Darkin, D., Brimbelcombe, R., Cook, A. J., Kirkham, J. S., and Northfield, T. C., Gut, 1986, 27, 491. Kyrtopoulos, S. A., Daskalakis, G., Legakis, N. I . , Konidaris, N., Psarrou, E., Bonatsos, G., Golematis, B., Lakiotis, G., Bliouras, N., and Outram, J . R., Carcinogenesis, 1985,6,1135. Reed, P. I . , Smith, P. L. R., Haines, K., House, F. R., and Walters, C. L., Lancet, 1981, ii, 550. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Ruddell, W. S. J., Bone, E. S., Hill, M. J., and Walters, C. L., Lancet, 1978, i, 521. Walters, C. L., Hart, R. J., and Smith, P. L. R., IARC Znt. Agency Res. Cancer Sci. Publ., 1983, No. 45, 295. Walters, C. L., Downes, M. J . , Edwards, M. W., and Smith, P. L. R., Analyst, 1978, 103, 1127. Bavin, P. M. G., Darkin, D. W., and Viney, N. J . , IARC Int. Agency Res. Cancer Sci. Publ., 1982, No. 41, 337. Bartsch, H., Ohshima, H., Munoz, N., Crespi, M., Cassale, V., Ramazotti, V., Lambert, R., Minaire, Y., Forichon, J., and Walters, C. L., IARC Int. Agency Res. Cancer Sci. Publ., 1984, No. 57,955. Keighley, M. R., Youngs, D., Poxon, V., Morris, D., Muscroft, T. J., Burdon, D. W., Barnard, J., Bavin, P. M. G., Brimbelcombe, R. W., Darkin, D. W., Moore, P. J., and Viney, N., Gut, 1984, 25, 238. Milton-Thompson, G. J., Lightfoot, N. F., Ahmet, Z . , and Hunt, R. H., Lancet, 1982, i, 1091. Dang Vu, B., Paul, J. L., Ekindjian, 0. G., Yonger, J . , Gaudric, M., and Guerre, J., Clin. Chem., 1983,29, 1860. Walters, C. L., Smith, P. L. R., Reed, P. I . , Haines, K., and House, F. R., IARC Int. Agency Res. Cancer Sci. Publ., 1982, No. 41, 345. Walters, C. L., Smith, P. L. R., and Reed, P. I . , IARC Int. Agency Res. Cancer Sci. Publ., 1984, No. 57, 113. Smith, P. L. R., Walters, C. L., and Reed, P. I., Analyst, 1983, 108, 896. Mirvish, S. S . , in Burchenal, J . H., and Oettgen, H. F., Editors, “Cancer 1980: Achievements, Challenges and Pros- pects for the 1980’s, Volume 1, ” Grune and Stratton, Orlando, FL, 1981, p. 557. Haga, J. J., Russell, B. R., and Chapel, J. F., Cancer Res., 1972, 32, 2085. Massey, R. C., Key, P. E., McWeeny, D. J., and Knowles, M. E., Food Addit. Contam., 1984, 1, 11. Paper A61485 Received December 30th, I986 Accepted February 23rd, 1987

ISSN:0003-2654    

DOI:10.1039/AN9871200945

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 951 Potentiometric-Determination of Heavy Metal Sulphide Solubilities Using a pH2S (GlasslAgo, Ag2S) Electrode Cell* Stefan Peiffert and Tonnies Frevert Chair of Hydrology, University of Bayreuth, P.O. Box 10 12 51, 0-8580 Bayreuth, FRG pH2S values (-log [H2S]) were calculated from the electrochemical reaction of a glass(Ag0, Ag2S electrode cell in equilibrium with metal sulphide precipitates at pH d 5 and compared with pH2S values predicted by the law of mass action. The results obtained indicated that the pH2S electrode cell has a Nernstian response at very low total sulphide concentrations. Measurements of the ion activity products in these solutions verified literature values. However, they also indicated that sulphide solubilities are influenced by the adsorption of hydrolysed metal species, leading to deviations from the theoretically predicted solubilities.As a general result, a detection limit for total sulphide activities corresponding to pH2S 18.2 was obtained. The electrode cell proved to be a suitable tool for the determination of the ion activity products of metal sulphides in difficult media (marine or estuarine waters) with a precision of f0.3 logarithmic unit. Keywords: Sulphide determination; potentiometry; sulphide solubilities The presence of hydrogen sulphide in limnetic and marine systems is significant for the following reasonsl? it has a toxic effect on the respiratory system of higher organisms; it implies anoxic conditions; and it causes the precipitation of many heavy metals.Investigations of the appearance and disappearance of hydrogen sulphide have usually been performed by the measurement of the e.m.f.s of sulphide-sensitive electrodes in combination with conventional reference electrodes in alka- line solutions.~5 The application of glass and sulphide electrode half-cells6.7 eliminates the interference from mixed potentials that can occur when reducing agents are used in alkaline solutions.8 A combined glass (Ag2S electrode cell (pH2S electrode cell, Ingold, FRG) satisfactorily provides a simple and precise analytical technique.9 The cell can be represented as Ago, AgII0.2 M HIlglass(solutionlAg2S, Ago, with Ago and AgI in 0.2 M HI as the internal reference system in the glass half-cell (pH = -0.7). The construction details of the cell are shown in Fig.1. The residual potential difference of the electrode cell is a direct function of the logarithm of the activity of hydrogen sulphide.9 Consequently, the pH2S value was introduced to represent the electrochemically detected variable of the concentration (activity) of H2S in aqueous so1utions.Y The electrochemical response of the glass(Ag2S electrode cell depends only on H2S at pH d 5 and is insensitive to further decreases in pH.9.10 Knowledge of the pH thus provides a means of determining the total sulphide in situ.11-13 The accuracy of this method is higher than that of the methylene blue method, with a precision of +1 mV.14 No effect was found on sensitivity, selectivity and slope up to an ionic strength of ca. 0.7.14 The electrode cell was employed in warm tropical sea water15 and under ice cover16 and no effect on reproducibility was observed.However, a temperature depen- dence of 6% per 10°C has been reported.14 According to Peters et al. 13 the calibration is carried out by a known standard addition procedure with 10-3, 10-4 and 10-5 M Na2S.9H20 solutions in buffers of pH d 5. With this method the buffer solutions have to be de-aerated by lengthy N2 bubbling (ca. 1 h) with the pH2S electrode immersed. The corresponding operational detection limit was reported to be lo-’ M.13 It was the purpose of this study to review and prove experimentally the theoretical implications of the measure- ments obtained with the cell with special respect to the measurement of low total sulphide activities.* Presented at the International Symposium on Electroanalysis and Sensors in Biomedical, Environmental and Industrial Sciences, Cardiff, UK, April 6-9, 1987. t To whom correspondence should be addressed. u i n Internal reference, solidified Ag2S-coated Ag ring b+---- Glass membrane Fig. 1. pH2S combined electrode cell (Ingold, FRG, Model Ag- 275-85-6329) Theoretical Theoretically, the lower detection limit can be defined by the electrochemical reaction of the silver half-cell AgOeAg+ + e- . . . . . . (1) A corresponding consideration of the glass half-cell can be neglected as in aqueous solutions X(H30+ + OH-) can never drop below 10-6.7 M. As silver ion-selective electrode cells are reported to have a Nernstian response down to Ag+ activities of 10-2O M , ~ ~ J ~ the response of the pH2S electrode cell may be linear down to extremely low pH2S values when buffered by heavy metal sulphide precipitates.19 In a first approach the e.m.f.s can be correlated with pH2S values from the operational calibration graphs as with e.m.f.O = e.m.f. corresponding to the known pH2SOvalue of the standard solution (0.1 M Na2S.9H20, pH 5 ) , e.m.f. =952 ANALYST, JULY 1987, VOL. 112 e.m.f. corresponding to the pH2S values in unknown solutions and S = slope (= 2.303 RT/2F). In the presence of metal sulphide precipitates, which act as sulphide buffers according to the metal buffer concept ,20,Z1 ion activity products can be calculated in the following way: pIAP = mpMe"+ + pKAl + pKL2 + pH2S - 2pH (3) where pMen+ = -log aMen+, pKAl and pKL2 = negative logarithm of the mixed acidity constants of H2S due to the ionic strength and temperature of the test solution and m = stoicheiometric coefficient.In a second approach, the e.m.f.s of the pH2S electrode cell can be calculated from the standard potentials of the glass and silver half cells (25"C, in mV), as the stoicheiometric composition of the solid silver sulphide membrane used in the electrode can be presumed to contain no excess of sulphur in the crystal lattice? e.m.f.calc, = E:lass - 59pH - Ag+ + 59pAg+ (4) - Ag+ is The value of aH30+ is defined by the given pH, taken from the literature and aAg+ is calculated as pAg+ = t [pCKsp(Ag2S) - PCKsp (MemS) + pMen+ - (rn + n) py] . . . . . . (5) where pcKsp is the non-ideality correction of pKsp for a given ionic strength.can be determined experimentally by the use of an external reference half-cell considering the following equa- tions (in mV, 25°C): EJass = e-m*f.glass - ref. + 59pH + &ref. &ref. = &go - Ag+ - 59[PKsp(AgCl) - Pc1-1 Thus, pH2S values due to the electrochemical reaction can be calculated by inserting e.m.f.calc, into equation (2). Further, pH2S and pS2- can be obtained from the law of mass action: pH2Seq = pS2- + 2pH - pKA1 - pKL2 . . (7) (8) ps2- = PC&p(Me,S) - mpMe"+ + rnpyMe"+ + pys2- where y? is the mean activity coefficient of Me,S(aq). Contrary to pH2Seq., values depend on both theoretical data and the properties of the electrode cell. If pH2Seq is equal to PH~S,,,.~., then the electrochemical response of the electrode cell corresponds to equation (1) and the response, given by pH2Smeasured, is proportional to the total sulphide activity (pH d 5 ) in the presence of metal sulphide precipitates.Experimental The e.m.f.s of the pH2S electrode cell (Ingold, FRG, Model Ag-275-85-6329) and of a pH measurement cell were deter- mined in solutions equilibrated to solid sulphide precipitates of some heavy metals. The precipitates were produced by adding 1 ml of 10-1 M Na2S.9H20 standard solution (stan- dardised iodimetrically) to 100 ml of 0.1 M ZnS04, PbC12, SnC12, CdS04 and AgCl with I = 0.1 and pH d 5. The pH was adjusted by the solution itself ( e . g . , by the formation of hydroxides), or was decreased, if necessary, by the addition of the de-aerated acid of the anion of the metal salt (H2S04 or HC1).If sparingly soluble metal salts [PbC12, Sn(OH)2, AgCl] were formed, the ionic strength was adjusted to I = 0.1 by adding 0.1 M NaCl. The reactions were carried out over 1 h in de-aerated and anoxically sealed reaction vessels (see Fig. 2). In view of the high and constant ionic strength, constant Fig. 2. Reaction vessel diffusion potentials (Ediff.) could be assumed at the liquid junction of the pH reference half-cell.22 Although equilibration was usually reached after several hours, the final potential readings were taken only after 4-10 d, during which time the e.m.f.s were monitored continuously on a two-channel recorder (Type 7020 Linseis) in the constantly stirred solutions. Ion activity products were calculated from equation (3) and the measured pH2S values; uMen+ was calculated from [Me"+] with respect to ionic strength (Davies approximation23), the common ion effect, pH [e.g. , Sn(OH)2] and complex forma- tion (e.g., AgC1). pKal and pKa2 (6.92 and 13.9 at 25 "C), were introduced as given by Broderius and Smith24 and Smith and Marte11,25 respectively, and were corrected to the ionic strength using the Guntelberg approximation.23 (B = base) i.e., pKA1 = 6.80, pKi2 = 13.54 for I = 0.1, T = 25°C. The non-ideality corrections of pKsp values taken from Smith and Martell25 were applied by the use of the Davies approximation23 VT l + V 7 PKA = PK, + O.~(Z& - &) - pCKsp = pKsp - (nz& + mzk) where n and m are the stoicheiometric coefficients and zM and ZN are the charge numbers of the species M, N.All experiments were carried out at constant temperature (24 k 0.5 "C). Seven replicates were used for PbS and the measurements were extended to decreased pHs by adding de-aerated acid to the equilibrated solutions of CdS and Ag2S. Results The experimental (conventional) calibration graphs were found to be mV+ 120 30 (electrode 1) . . pH2S = 4 - (9)ANALYST, JULY 1987, VOL. 112 953 Table 1. Results of pH2S measurements in the presence of PbS precipitates calculated from equation (9). S = initial value; F = final value Experiment number 1 2 3 4 5 6* 7* Equilibration time/h 70 46 31 49 58 64 71 Statistical results: X S PHS 4.2 4.2 4.2 4.2 4.2 4.3 4.3 PHF 4.2 4.1 4.2 4.0 4.2 4.2 4.2 e.m.f.s -410 -421 - 420 -411 -413 - 443 - 433 e.m.f.F - 420 -431 - 427 -416 - 427 - 448 - 438 4.2 4.2 - 0.08 ,424t 6.06 PH2SS 13.9 13.8 13.7 14.0 13.8 14.5 14.2 PH2SF 14.0 14.0 13.9 14.2 14.2 14.7 14.4 14.2 0.28 14.2 k 0.33 t s 3 f -(a = 1%) 4.2 f 0.09 -424 f 10.2 fi mV + 117 * Electrode cell had to be re-calculated: pH2S = 4 - ~ 31 ' t Calculated without experiments 6 and 7.Table 2. Results of pH2S measurements in the presence of metal sulphide precipitates and their ion activity products calculated from equations (3) and (10). S = Initial value; F = final value Metal salt CdS04* Measurement ZnS04 Equilibrationtime/d . . . . . . 10 5 4 pMenf . . . . . . . . . . 1.54 1.1 1.1 pHS . . . . . . . . . . 3.6 4.8 3.3 pH" . . . . . . . . . . 3.7 4.7 3.3 E .m. f .S/mV . . . . . . . . - 272 - 440 -361 E.m.f.F/mV . . . . . . . . -282 - 444 -363 PH2S2easured .. . . . . . . 9.3 14.6 12.1 PH2SLeasured . . . . . . . . 9.6 14.8 12.2 pIAPMe,S . . . . . . . . 24.1 26.8 27.0 * For CdS04 and AgCl the measurements were carried out at various pH. t Mean values (cf. Table 1). PbC12t SnC12 5 3.4 1.6 1.7 4.2 1.8 -283 - 424 -284 9.7 14.2 f 0.33 9.7 29.5 f 0.33 28.0 AgCl * 4 8.9 4.4 3.7 - 598 - 554 19.7 18.2 48.9 4 8.9 2.7 2.7 - 496 - 496 16.4 16.4 49.1 Table 3. pH2S values calculated from the electrochemical response of the electrode cell [PH~S,.,.~., equation (6)] and the law of mass action [pH2Se,, equation (7)] (for I = 0.1, T = 24 f 0.5"C) and pH2S values measured in this study. pcKs, values from Smith and Martell.25 Metal sulphide Measurement ZnS(a) pc Ksp . . . . . . . . . . 24.0 pMen+ . . . . . . . .. . 1.54 pAg+ . . . . . . . . . . 13.6 pH . . . . . . . . . . . . 3.7 E.m.f.,,,,./mV . . . . . . . . -285 pH2Se,,,f, . . . . . . . . 10.4 pH2Seq. . . . . . . . . . . 10.4 ~HzSrneas . . . . . . . . 9.6 * Calculation at various pH for CdS and Ag2S. t Mean values (cf., Table 1). CdS* 26.3 1.1 12.2 4.7 15.0 15.1 14.8 - 427 - 26.3 1.1 12.2 3.3 12.4 12.3 12.2 ,344 PbSt 26.8 3.4 13.1 4.2 12.4 12.3 14.2 -329 SnS 25.2 1.6 13.0 1.8 8.1 7.1 9.7 - 208 49.6 8.9 8.9 3.7 18.5 19.5 18.2 -562 49.6 8.9 8.9 2.7 16.6 17.5 16.4 .503 for the seven replicates with PbS, and mV + 105 31.5 pH2S = 4 - (electrode2) . . for the other four investigated metal sulphides, for which another electrode cell was used. The results of the measure- ments after insertion into equations (3), (9) and (10) are given in Tables 1 and 2.The precision of pH2S in the presence of PbS is 14.2 k 0.3 (Table 1). The detected change in pH2S due to the change of pH corresponded precisely to equation (7) (Table 2). Qlass was -55 mV for electrode 1 and -70 mV for electrode 2, respectively. Subsequently, pH2S,.,.f, values could be calculated using equations (4), (6), (9) and (10). The experimental conditions used are given in Table 3 together with the results obtained including pH2Seq. The latter corre- sponded well to PH~S,.,,~, values. Both, however, deviate considerably from the measured pH2S as determined in the presence of the metal sulphide precipitates for PbS and SnS. This difference can also be observed when comparing pIAP values of this study with accepted pKsp values.25 Discussion The calculation of the electrochemical response of the electrode cell according to the Ag+ activity in solution provides pH2Se,m,f.values which are in good agreement with pH2Seq values predicted from the law of mass action. The observed differences can be explained by the non-theoretical slope of the electrode cell [31.5 mV, equation (lo)], which may result from the calibration procedure (re-diffusion of954 ANALYST, JULY 1987, VOL. 112 Table 4. Accepted p‘K,, values for ZnS, PbS, SnS, CdS and Ag2S and pIAP values as determined in this study. I = 0.1, T = 24 f 0.5 “C Metal sulphide Smith and Martel125 This study ZnS(cx) 24.0 24.1 PbS 26.8 29.5 k 0.33* SnS 25.2 28.0 CdS 26.3 26.8,27.0 AgS 49.6 48.9,49.1 * Mean value (cf., Table 1). oxygen into the solution of 10-5 M total sulphide). Extrapola- tion of the calibration graph to such low concentrations (activities) as encountered in this study will increase the calibration error.However, no serious effect of the non-ideal behaviour of the electrode cell on the extrapolation is to be expected. The Nernstian behaviour of the pH2S electrode cell at high pH2S values is thus established and the validity of equation (1) is confirmed. The dissolution of sparingly soluble salts is a function of the area and the nature of the surface of the precipitate. As the precipitation of the metal sulphides was rapid, the surface area is much greater than that of pure crystals.26 This increase in surface area gives rise to the adsorption of metal salts, which has been reported to occur during the formation of metal sulphides precipitated from excess metal ions.19 Jean and Bancroft27 discussed a monolayer adsorption model for the adsorption of hydrolysed metal species to the sulphide group of sulphide mineral surfaces.where s denotes the sulphide group of the precipitate. The adsorption of hydrolysed metal species increases in the order Ag < Cd < Zn < Pb < Sn, according to the first hydrolysis constants at a given pH.27 However, only Sn and Pb form hydroxo complexes at the experimental pHs (1.8 and 4.2, respectively). Reaction mechanisms may exist that will alter the surface of the SnS and PbS precipitates and therefore decrease the amount of sulphide in solution. The measured ion activity products therefore deviate from solubility pro- ducts of pure crystals (Table 4). The differences in experimen- tal measurements of ion activity products range from 26.6 to 29.4 for PbS and from 25.0 to 26.9 for SnS.28 - S + Me(OH),(2 - -+ s - Me(OH),(2 - m)+ Conclusions The results of this study allow the following conclusions: (1) a pH2S value of 18.2 has been verified by an experimental measurement technique and (2) ion activity products of heavy metal sulphides can be measured under conditions of changing temperature, ionic strength and chemical composition of the solution with a minimum disturbance of the chemical equilib- rium.The precision is about half a logarithmic unit. The electrode cell thus provides a suitable tool for analytical chemists, geochemists and marine chemists. The support of the State of Bavaria, Ministry of Culture and Education for this study is gratefully acknowledged.We are also grateful to the Yigal Allon Kinneret Limnological Laboratory, Tiberias, Israel for steadily supporting our investigations. 1. 2. 3. 4. 5. 6. 7. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. References Wetzel, R. G., “Limnology,” Second Edition, W. B. Saunders, Philadelphia, PA, 1981. Riley, J. P., and Skirrow, G., Editors, “Chemical Ocean- ography,” Academic Press, London and New York, 1965. Berner, R. A., Geochim. Cosmochim. Acta, 1963,27, 563. Champs, D. R., Gulens, J., and Jackson, R. E., Can. J . Earth Sci., 1979, 16, 12. Thomas, J. D. R, Znt. J . Environ. Anal. Chem., 1985,20, 167. Gulens, J., Herrington, H.D., Thorpe, J. W., Mainprize, G., Locke, M. G., Dal Bello, P., and Macdougall, S., Anal. Chim. Acta, 1982, 138, 55. Gulens, J., Water Res., 1985, 19, 201. Frevert, T., and Galster, H., Schweiz. 2. Hydrol., 1978, 40, 199. Orion Research Inc., US Pat., 4 131 148 and 4 131 428, 1978. Eckert, W., and Frevert, T., in Pungor, E., Editor, “Fourth Scientific Session on Ion-selective Electrodes, Miitrafiired, Hungary,” Hungarian Academy of Sciences, Budapest, 1984. Bergstein-BenDan, T., Frevert, T., and Cavari, B., Water Res., 1985, 19, 983. Peters, K., Huber, G., Netsch, S . , and Frevert, T., Gas Wusserfach, 1984, 125, 386. Frevert, T., Schweiz. 2. Hydrol., 1980,42, 255. Frevert, T., unpublished results. Diireth, S . , Herrmann, R., and Pecher, K., Water Air Soil Pollut., 1986, 28, 131. Schwarzenbach, J., and Widmer, L., Helv. Chim. Acta, 1966, 49,111. Cammann, K., “Das Arbeiten mit ionenselektiven Elek- troden,” Springer Verlag, Berlin, 1977. Buck, R. P., in Freiser, H., Editor, “Ion-selective Electrodes in Analytical Chemistry,” Volume 1, 1978, p. 67. Ringbom, A., “Complexation in Analytical Chemistry,” Inter- science, New York, 1963. Hansen, E. E., Lamm, C. G., and Rfiiieka, J., Anal. Chim. Acta, 1972, 59, 403. Hseu, T. M., and Rechnitz, G. A., Anal. Chem., 1968, 40, 1054. Stumm, W., and Morgan, J. J., “Aquatic Chemistry,” Second Edition, Wiley-Interscience, New York, 1981. Broderius, S. I., and Smith, K. K., Jr., Anal. Chem., 1977,49, 424. Smith, R. M., and Martell, P. E . , “Critical Stability Constants, Inorganic Complexes,” Volume 4, Plenum, New York, 1976. Meites, L., “An Introduction to Chemical Equilibrium and Kinetics,” Pergamon Press, Oxford, 1981. Jean, G. E., and Bancroft, G. M., Geochim. Cosmochim. Acta, 1986, 50, 1455. SillCn, L. G., and Martell, A. E., “Stability Constants of Metal Ion Complexes,” Special Publication No. 17, Chemical Society, London, 1964. Paper A6/416 Received October 30th, 1986 Accepted December 11 th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200951

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 955 Reductive Galvanic Stripping Determination of Lead Sambamoorthy Jaya, Talasila Prasada Rao and Gollakota Prabhakara Rao Cen tra I Electrochemical Research Institute, Ka ra iku di- 623 006, Tam iln a d u, India The feasibility of the determination of trace elements by anodic pre-concentration and subsequent reductive galvanic stripping is demonstrated using the determination of lead as an example. The technique is suitable for the determination of lead in the pg I-' range. Keywords: Reductive galvanic stripping analysis; lead determination; platinum electrode Cathodic stripping voltammetry (CSV) involves the anodic pre-concentration of analyte species followed by a linear cathodic potential scan. 1 During the cathodic reduction, the current flow between the working and counter electrodes is measured.However, this current consists of both capacitance and Faradaic components, of which only the latter is useful for analytical purposes. As in anodic stripping voltammetry (ASV), attempts to isolate the Faradaic current from the capacitance current have resulted in the introduction of sophisticated electroanalytical techniques such as normal pulse (NP) CSV,2 differential pulse (DP) CSV3 and staircase stripping (SS) CSV.3 Recently Jagner4 described the technique of potentiometric stripping analysis (PSA), which is based on the chemical stripping of pre-concentrated elements using mercury(I1) as an oxidant. This technique has been shown to be reliable and has several advantages over conventional ASV. Lead and manganese have been determined by anodic pre- concentration on to a platinum electrode, followed by chemical stripping using hydroquinone or pyrogallol.5 This technique was termed reductive potentiometric stripping analysis (RPSA).This paper reports the development of an electroanalytical technique, reductive galvanic stripping analysis (RGSA), which involves the pre-concentration of the analyte by anodic oxidation followed by stripping in an open- circuit position (OCP). An extension of the anodic stripping technique using a galvanic stripping method has been reported previously.6 In OCP mode, anodically pre-concentrated lead is stripped by a localised galvanic couple formation in the absence of any external electrical input or chemical reductant. The determination of lead by anodic deposition as lead dioxide and cathodic stripping voltammetry has been de- scribed by several workers.7-lo In this paper the determina- tion of lead using a platinum working electrode is discussed as an illustrative example of the principles of the technique.This technique has also been found to be useful in the determina- tion of manganese.11 Experimental Reagents All solutions were prepared by dissolving analytical-reagent grade reagents in conductivity water. A lead(I1) solution (0.01 M) was prepared by dissolving 0.3312 g of lead nitrate in water and diluting to 100 ml. This stock solution was diluted with water as required. Acetate buffer (1.0 M) was prepared by dissolving 12.6 g of sodium acetate and 6 ml of glacial acetic acid in 80 ml of water and diluting to 100 ml.The DH was adiusted to 5. Apparatus A conventional three-electrode cell with a platinum disc working electrode (JMC, 0.38 cmz), a bright platinum foil (12 cm2) counter electrode and a standard calomel reference electrode (SCE) was used. A Wenking Model LB 75 M potentiostat coupled with a Wenking scan generator (Model VSG 72) was used. An X - Y/t recorder (Digitronic Model 2000) was used to record the potential - time ( E - t ) profiles. All the potentials were expressed with respect to SCE and the experiments were carried out at about 60 "C. Procedure To a 50-ml calibrated flask containing 5 X 10-*-1 X 1 0 - 5 ~ lead, add 5 ml of acetate - acetic acid buffer (pH 5) and dilute to the mark with water. Transfer the solution into the voltammetric cell and de-aerate with high purity nitrogen for 10 min.Electrolyse the solution at + 1.4 V vs. SCE for 4 min at 60 "C (stirring the solution with a magnetic stirrer) in order to pre-concentrate the lead on to the platinum disc electrode as lead dioxide. Immediately after the completion of electrolysis, plot the E - t profiles by recording the working electrode potential vs. SCE with time by switching over to the OCP (the E, mode of the potentiostat). Draw a calibration graph by plotting the time required for galvanic stripping against the concentration of lead in solution. Results and Discussion Characteristics of E - t Profiles The feasibility of obtaining a reductive stripping signal for lead dioxide deposited on a platinum working electrode is demon- strated in Fig.1. The typical E - t profiles indicated in the 40 30 v) . ; 20 i= 10 0 1.4 0.9 El V 0.4 Fig. 1. Potential ( E vs. SCE) - time profiles for A, 0.1 M acetate buffer and B and C, 5 x 10-6 and 1 x 10-5 M lead in acetate buffer. Ed = 1.4 V; td = 4 min; pH = 5956 ANALYST, JULY 1987, VOL. 112 1 .o 0 El V Fig. 2 . Potential ( E vs. Pt counter electrode) - time profiles for A, 0.1 M acetate buffer and B and C, 5 x 10-6 and 1 x 10-5 M of lead in acetate buffer. Ed = 1.4 V vs. SCE; td = 4 min; pH = 5 figure were recorded after an anodic pre-concentration at +1.4 V (vs. SCE) for 4 min in 0.1 M acetate buffer (pH 5). Lines A, B and C in the figure represent the E - t profiles obtained when the potential changes of the working electrode vs.SCE were recorded in the E, mode of the potentiostat for 0.0, 5 and 10 p~ concentrations of lead. It can be seen from Fig. 1 that the potential of the working electrode falls rapidly from 1.4 to 0.9 V (first break-point). A plateau is observed on the time axis, the length of which is proportional to the concentration of lead in solution. From the moment at which all the Pb02 is reduced (the second break-point) the potential slowly drops further to the open circuit potential of the system (a Pt working electrode in 0.1 M acetate buffer, pH 5), i.e., +0.09 V. The time over which the plateau lasts is the time required for the stripping of the lead dioxide deposited earlier. The potential over which the plateau occurs (0.77-0.70 V) correlates well with the peak potential of the PbOz - Pb2+ couple (0.74 V vs.SCE) obtained from the cathodic stripping voltammogram recorded under similar conditions in a sep- arate experiment. This clearly indicates that Pb02 dissolves as Pb2+ during the reductive stripping process, either in the presence of an applied potential sweep or in its absence. Fig. 2 depicts the E - t profiles obtained by following the changes in the potential of the working electrode (after PbO2 deposition) against the counter electrode, under conditions similar to those described in Fig. 1. These graphs were recorded by connecting the working and counter electrodes directly to an X - t recorder after disconnecting the potentio- stat immediately after pre-concentration for 4 min at 1.4 V. As seen from Fig. 2, the potential of the working electrode drops immediately from 1.74 to 1.0 V.The E - t graphs obtained in this instance are identical with those in Fig. 1 in that the plateau of the graphs are of the same length for both 5 and 10 p~ lead. Significantly, when the working and counter elec- trodes were connected externally through a sensitive ammeter, a current flow between the counter and working electrodes was observed during the plateau region of the stripping curves where reductive dissolution of Pb02 occurs. The following four observations, made in the absence of any externally added reagents that could react with Pb02, suggest the formation of a galvanic cell. (1) The attainment of the E - t profiles of the working electrode either vs. SCE (Fig. 1) or the platinum counter electrode (Fig.2) described above, in the absence of external polarisation. (2) The occurrence of a plateau on the time axis of the E - t profiles. (3) A finite and constant current flow between the counter and working electrode in the plateau region. (4) The total charge passed during the reductive galvanic stripping is equal to the total charge passed during the anodic deposition of lead. This suggested galvanic cell is analogous to that set up during the dissolution of deposited cadmium from a mercury film6 and is presumed to provide the driving force for the observed E - t responses as discussed below. The dissolution of anodically deposited Pb02 can also be thought of as being due to either the presence of reducing impurities or a chemical reaction with water.The former is ruled out by the fact that the magnitude of the stripping signal is unaffected by addition of 10-4 M hydroquinone and a change of rotation rate. Similar E - t profiles are obtained in an electrolytically purified supporting electrolyte (i. e. , by pre- oxidation). The possible chemical reduction of Pb02 by water is insignificant under the experimental conditions employed as Pb02 deposited anodically on a lead substrate remained stable even after 5-10 min in the OCP. A similar observation was reported by Kinard and Propst12 during their studies of the OCP dissolution of Pb02 deposited on a conducting glass electrode. As noted earlier, the reaction occurring at the working electrode is the reduction of Pb02 to Pb2+ in the plateau region. The coupling reaction of this reduction is identified as oxygen evolution at the platinum counter electrode (see Fig.2), analogous with the hydrogen evolution reaction observed in galvanic stripping analysis.6 Thus, although the general E - t profile pattern with the characteristic plateau region is indicative of a reductive galvanic stripping phenomenon in operation, the similar results shown in Fig. 1, in which the working electrode potential is recorded vs. SCE needs further investigation to be understood, as in this instance, the platinum counter electrode, the site for the galvanic coupling reaction, is disconnected from the circuit. Galvanic coupling can be envisaged even in this example by the formation of localised galvanic cells on the surface of the platinum working electrode (i.e., between the Pb02 centres and the uncovered platinum surface sites constituting the necessary electrode pair for the galvanic cell), which provide the basis for the observed E - t response.This proposition gains support from the facts that, (i) Pb02 deposits as a non-uniform film; and (ii) a high impedance of the order of 1011 B exists between the working and SCE electrodes in the E, mode (thereby ruling out the possibility of current leakage through the reference elec- trode). On the other hand, when the working electrode remains coupled to the counter platinum foil (as in Fig. 2) a galvanic cell is preferentially set up between them, thus shifting the site of anodic reaction to the counter electrode. This galvanic cell is due to the larger area of the counter electrode as compared with the uncovered area of platinum on the working electrode deposited with Pb02.The performance of RGSA, i . e . , the selectivity, accuracy, precision and detection limit, depends on a number of experimental parameters such as mode of electrode surface regeneration, temperature, the pH of the medium and the plating potential. The effect of these parameters on the procedure for the determination of lead are discussed below. Electrode Regeneration In order to obtain reproducible results for the determination of lead using RGSA, any oxide layer formed on the platinum electrode must be removed completely before each experi- ment. Various methods were attempted to achieve this. It was observed that treating the electrode with nitric or hydrochloric acid at room temperature for about 10 min was insufficient to regenerate the electrode surface.The recommended pro- cedure involves the cathodic polarisation of the electrode at about 200 mV anodic to the hydrogen evolution potential in the given medium.5 The regeneration procedure was per- formed immediately after the stripping procedure until the cathodic current dropped to a stable value. This usually occurs after about 2-3 min. Effect of Temperature The re-dissolution of Pb02 in the reductive galvanic stripping mode at room temperature is slow, with no clear break-points,ANALYST, JULY 1987, VOL. 112 957 PH 4.5 5.0 5.5 6.0 I I 1 ' 12.5 9. \ I O H 1.4 1.15 0.9 Deposition potentialN Fig. 3. Effect of pH (A) and plating potential (B) on the RGSA signal of M lead.Conditions: pH 5, 0.1 M acetate buffer, Ed = 1.4 V vs. SCE, rd = 2 min and temperature = 60°C [Hydroquinone]-1/103 M I 1 1 1 0 2.5 5.0 7.5 0 25 50 75 100 125 [Hydroquinone]/lO-5 M Fig. 4. Effect of hydroquinone (H2Q) on the stripping signal of 1 0 - 6 ~ lead. Conditions: pH 5, 0.1 M acetate buffer, Ed = 1.4 V vs. SCE, td = 2 min and temperature 60 "C. A, stripping signal vs. [H2Q] and B, stripping signal vs. [H2Q]-l and the reproducibility of the stripping signal is also poor. At temperatures in the range 4&70"C, the quality and the reproducibility of the RGSA signal are improved significantly. The magnitude of the stripping signal reaches a maximum at a temperature of 50°C and remains constant with further increases in temperature up to 70°C.Hence subsequent experiments were carried out at ca. 60 "C. Effect of pH The effect of pH on the deposition of Pb02 and the subsequent reductive galvanic stripping analysis was studied using 1 0 - 6 ~ PbII, adjusting the pH of the acetate-buffered medium. The pH dependence of the RGSA signal vs. time of stripping is shown in line A of Fig. 3. It is clear from this figure that a constant and maximum signal is obtained in the pH range of 4.7-5.5, which decreases either on decreasing or increasing the pH beyond this range. Hence a pH 5.0 acetate buffer (0.1 M) was used in subsequent experiments. Effect of Deposition Potential The magnitudes of the RGSA signals obtained by the described procedure for a 1 0 - 6 ~ lead solution in 0.1 M acetate buffer (pH 5,60 "C) at various deposition potentials (2 min each) are shown in line B of Fig.3. It is clear from Fig. 3 that the deposition of Pb02 begins only when the deposition potential is above 1.0 V, i.e., 0.25 V more positive than the dissolution potential of the PbOz - Pb2+ couple. The deposition reaches a maximum when the deposition potential is maintained above 1.2 V and then remains unaltered on further increases up to 1.4 V. Hence, for maximum deposition, the potential should 0 50 100"100 500 900 Concentration of lead/l0-8 M Fig. 5. Calibration graph consisting of two linear segments obtained for lead in 0-10-6 M solutions 1.4 1.15 0.9 0.65 0.4 PotentiaW Fig. 6. Typical stripping profiles recorded for A, blank solution, B, solution containing 1 x 10-8 M and C, 5 x 10-6 M lead in solution be at least 0.45 V more positive than the potential of Pb02 dissolution.Effect of Hydroquinone Concentration Hydroquinone has been used as a reductant in RPSA.5 Hence, a study was carried out to investigate the effect of the addition of different amounts of hydroquinone on the reductive stripping signal. The results of the investigation ( e . g . , line A, Fig. 4) show that in the presence of 2 1 0 - 4 ~ hydroquinone, chemical stripping of Pb02 predominates. In more dilute solutions of hydroquinone, however, the presence of hydro- quinone does not affect the stripping signal, as evidenced from the fact that identical signals were recorded both in the presence and absence of hydroquinone. This observation suggests that stripping of Pb02 occurs by galvanic couple formation in these solutions.It is also seen from line B, Fig. 4, that the stripping does not occur entirely by chemical means even in the presence of 10-4-10-3 M hydroquinone, as evidenced by the non-linearity of the graph. Interference Studies The effect of various cations, either singly or in combination, in the RGSA of PbII was examined by the addition of a 100-fold excess of each of these ions to 10-6 M PbIr. Of the ions studied, i. e . , Cd", CoII, TlI, FeIII, ZnlI, CuII, Mn" and HgII, only HgII and MnII were found to interfere. Detection Limit, Precision and Calibration Range The detection limit depends on the deposition time. Thus, a 2-min pre-concentration time for 10-8 M lead on Pt produces a significant signal, i .e . , twice the standard deviation of the background signal. The precision of the technique was obtained by carrying out five successive experiments, each with a 2-min deposition time958 and 1 0 - 6 ~ solutions of lead(II), as described under Pro- cedure. The relative standard deviation was 1.5%. The calibration graph (Fig 5.) was linear in the range 5 X 10-S- 1 x 10-5 M Pb2+ in solution. Stripping profiles in typical dilute solutions of lead are shown in Fig. 6. The authors thank Prof. K. I. Vasu, the Director of the Central Electrochemical Research Institute, Karaikudi, for his keen interest in this work and his kind permission to publish the results. References Florence, T. M., J. Electroanal. Chern., 1979,97, 219. Shimizu, K., and Osteryoung, R. A., Anal. Chern., 1981, 53, 584. Christie, J. H., and Osteryoung, R. A., Anal. Chern., 1976,443, 869. Jagner, D., Analyst, 1982, 107, 593. 5. 6. 7. 8. 9. 10. 11. 12. ANALYST, JULY 1987, VOL. 112 Christensen, J. K., and Kryger, L., Anal. Chirn. Acta, 1980,53, 118. Jaya, S . , Prasada Rao, T., and Prabhakara Rao, G., Anal. Lett., 1985, 18A, 1441. Miwa, T., Motosugi, H., and Mizuike, A,, Bunseki Kagaku, 1971, 7, 846. Monien, H., and Zinke, K., Fresenius Z. Anal. Chem., 1968, 240, 32. Hrabankova, E . , Beran, E., and Dobzal, J., J. Electroanal. Chem., 1969, 22, 303. Monien, H., and Zinke, K., Fresenius Z. Anal. Chem., 1970, 250, 178. Jaya, S . , Prasada Rao, T., and Prabhakara Rao, G., in preparation. Kinard, J. T., and Propst, R. C., Anal. Chem., 1974,46, 1106. Paper A51339 Received September 12th, 1985 Accepted October 16th, I986

ISSN:0003-2654    

DOI:10.1039/AN9871200955

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 959 Square-wave Voltammetric Behaviour and Automated Determination of Cephalothin by a Novel Sample Handling Approach Dorit Peled and Chaim Yarnitzky Department of Chemistry, Technion-Israel Institute of Technology, Haifa 32000, Israel and W. Franklin Smyth* Department of Pharmacy, The Queen's University of Belfast, Medical Biology Centre, Belfast BT9 7BL, UK The voltammetric behaviour of the cephalosporin cephalothin has been studied by rapid-scan square-wave voltammetry at single mercury drops and at the static mercury drop electrode (SMDE). A study of the variation of the peak current with solution variables such as pH, supporting electrolyte, concentration of cephalothin and instrumental variables such as scan speed, pulse width, pulse height and delay time has resulted in optimisation of the reduction signal for analytical purposes.This reduction signal can be used to monitor the degradation of cephalothin in acidic and basic media, and has also been applied to the rapid, accurate and precise determination of cephalothin in a pharmaceutical preparation using an automated sampling handling approach controlled by a polarographic analyser. Concentrations of cephalothin in the range IO-7-10-8 M can be determined using this automated approach and involving rapid-scan square-wave voltammetric determination after adsorptive accumulation of cephalothin at the SMDE. The applicability of this latter technique to the determination of cephalothin in body fluids is assessed. Keywords: Cephalothin determination; square-wave voltammetry; automated cell Cephalothin (I), 7-(2-thienylacetamido)cephalosporanic acid, together with cephaloridine and cephalosporin C , were first studied polarographically by Jones et al.1 These antibiotics, commonly used for the treatment of infections caused by Gram-positive cocci and Gram-negative bacilli, were each found to exhibit one reduction wave and were used for qualitative and quantitative analysis. Hall et a1.2 have carried out a more exhaustive electrochemical study on cephalosporin C derivatives which included cephalothin. They found a d.c. current at 10-4 M concentrations which was constant in height over the pH range 0-4 and decreased to zero by pH 7-8. They reported that there was no observable reduction wave for the cephalothin anion in buffered solutions up to pH 8 but that a wave at -1.80 V existed in 0.1 M Et4NC104 supporting electrolyte.They proposed the following reduction mechan- ism for such substituted 3-methylcephalosporins to produce A3-deacetoxy and 3-exomethylene cephalothin compounds which are geometric isomers: [ A c O C I T -I ads HOOC H :.3? + w3< +e- + Hf fast T Yamana et a1.3 have studied the kinetics and mechanism of degradation of cephalosporins in aqueous solutions and have shown that 3-acetoxymethylcephalosporins such as cephalo- thin degrade by the following general mechanism over the whole pH range: CH20H I COOH Ik1 Products I COOH Deacetyl intermediate ~~~ * Lady Davis Visiting Professor to Technion 198&87. Lactone ik5 Products960 ANALYST, JULY 1987, VOL.112 kl, k3 and k5 correspond to rates of P-lactam cleavage in the nucleus; k4 is negligible at neutral and basic pH but is significant at pH < 4. Fogg et aZ.4 have used differential-pulse polarography at the dropping-mercury electrode to follow the degradation of 3.4 x 10-4 M cephalothin in pH 2 Britton - Robinson buffer at 50 “C. A ti of 4.5 h was reported and after 18.5 h no cephalothin or other degradation products could be detected by this technique. This suggests electro-inactivity of 0-lactam cleaved products and the deacetyl and lactone intermediates in the above scheme. Benners has made the only electroanalytical determination of cephalothin in a body fluid. He proposed a 2-h assay for the oscillopolarographic determination of cephalothin and other cephalosporins and penicillins in serum following ultrafiltra- tion with a UM-2 filter.Cephalothin has also been determined in body fluids by high-performance liquid chromatography (HPLC). Cooper et aZ.6 have used ion-pair extraction prior to HPLC determination of cephalothin and its deacetyl meta- bolite in urine, with a recovery of ca. 80%. Buhs et aZ.7 have determined cephalothin and cefoxitin and their deacetylated metabolites in urine by HPLC. Serum proteins have been precipitated by dimethylformamide and the supernatant analysed by HPLC with UV detection at 254 nm8; the serum concentration of cephalothin and its deacetyl metabolite have been detected at concentrations of ca. 1 pg ml-1 with linear calibration in the range 5-50 pg ml-1. Nygard and Khalidg have precipitated plasma proteins with acetonitrile and injected 10 pl of the supernatant on to a reversed-phase C18 column, using a mobile phase of 6-11% acetonitrile in 0.01 M NaH2P04 and UV detection at 254 mm.The limit of detection was <1 pg ml-1 and linearity was found in the concentration range 5-500 pg ml-1 of plasma, of which corresponds to ca. 10-5-10-3 M. This method was found to be particularly selective when the capacity factors for co-administered antibiotics and other drugs were calculated. Such techniques for the measurement of serum and plasma concentrations of cephalosporins are necessary in the clinical environment to ensure that adequate drug levels can be maintained while avoiding toxic concentrations, found parti- cularly with renal impairment and combination therapy with aminoglycosides .This paper is concerned with a study of the voltammetric behaviour of cephalothin using the particularly rapid and sensitive techniques of square-wave polarography at single mercury drops and square-wave voltammetry (SWV) at the SMDE. The reduction signal of cephalothin was optimised by variation of relevant instrumental and solution parameters and then applied to a study of the degradation of cephalothin in acidic and basic media. An automated sample handling approach, invented by Yarnitzkylo and controlled by a modern polarographic analyser equipped with a microproces- sor, was then used to determine the content of cephalothin in pharmaceutical preparations with excellent rapidity, accuracy and precision.Cephalothin is also adsorbed on to the mercury drop over a wide range of potential and this phenomenon was put to analytical advantage in the design of an adsorptive stripping method for the determination of cephalothin at low p.p.b. levels, i.e., 10-8-10-7 M concentrations. Its applicabil- ity to the determination of cephalothin in urine and blood plasma was also evaluated. Experimental Apparatus Rapid-scan square-wave polarography at single mercury drops was performed by an instrument designed and built at the Technion, Haifa. Rapid-scan square-wave voltammetry at the SMDE was carried out using an EG & G PAR 384B polarographic analyser which controlled the operation of a Model 309 automatic voltammetric electrode for rapid and automated sample handling.All potentials were measured with reference to a Ag - AgCl electrode. Reagents Stock solutions of ca. 10-3 M cephalothin sodium salt were made up weekly in trebly distilled water and stored in the dark and under refrigeration in order to minimise decomposition. The pharmaceutical formulations used in this study were Keflin, manufactured by Eli Lilly, and consisted of ampoules which contained 1 g of cephalothin and 30 mg of NaHC03. The solutions of these formulations were also made up in trebly distilled water to the appropriate concentrations. All buffer and supporting electrolyte solutions were made up from AnalaR chemicals with trebly distilled water. Techniques The voltammetric behaviour of cephalothin was investigated by rapid-scan square-wave polarography at single mercury drops and at the SMDE after application of the automated sample handling approach invented by Yarnitzky.10 Using this approach, controlled by a polarographic analyser such as the Technion instrument and the EG & G PAR 384 B instrument, the sample to be analysed (8-10 ml) is drawn through the sample inlet by means of the Venturi effect caused by a stream of nitrogen.This is effected by activation of the nitrogen pressure valve and the sample feed pump at the same time. Once in the nebuliser chamber, the sample is instantanecusly de-aerated and arrives in the polarographic cell where it collects. Once the solution in the cell reaches the indicator electrode, the stream of nitrogen and feed pump are stopped automatically. This is accomplished by measuring the conductivity between the working and counter electrodes, after switching the counter electrode from the potentiostat to +12 V and the working electrode from the current to potential converter to a resistor connected to the ground.A potential across this resistor exceeding - 10 V indicates a proper solution level and the electrodes are switched back to the potentiostat and the converter, respectively. The first portion of the sample, for the Technion instrument, is used for cell flushing. The scavenging pump is then activated and the sample pumped out of the cell through a mercury valve with a glass bulb of 20-ml volume. The waste, pumped into this bulb, continues to the waste bottle by gravity flow when the pumping action is stopped. The nebuliser is then re-activated and a fresh portion‘ of the sample is brought into the polarographic cell.For the link-up of the EG & G PAR 384 B polarographic analyser with this sample handling approach , cell flushing was accomplished by using at least one blank containing trebly distilled water. The polarographic or voltammetric analysis was then carried out by the rapid-scan square-wave technique, suggested by Osteryoung “from the analytical point of view, most likely to dominate pulse voltammetry in the future.”11 The study of the degradation of cephalothin in acidic and basic media was carried out at concentrations of ca. 10-5 M cephalothin in 1 M HC1, 1 M HC104 and 1 M NaOH, respec- tively, at ambient temperature (25 “C). Aliquots of the reaction mixture could be examined without dilution in the voltammetric cell.Alternatively 1-ml aliquots of these reac- tion mixtures were sampled at selected times after initiation of the degradation reaction, diluted to 10 ml with trebly distilled water or the appropriate supporting electrolyte and imme- diately examined by rapid-scan square-wave voltammetry after application of the automated sample handling approach described above. In this way, disappearance of the current due to cephalothin and appearance of any new peaks over a wide range of potential was followed with time. Adsorptive and cathodic stripping voltammetry using adsorption or deposition at pre-selected potentials for time periods up to several minutes were also used to investigate whether any of the degradation products gave rise to comparatively large strip- ping signals for use in the development of sensitive electro- analytical methods based on initial degradation of cephalo- thin.ANALYST, JULY 1987, VOL.112 961 Pharmaceutical preparations containing cephalothin as the sodium salt (1 g) and 30 mg of NaHC03 were initially dissolved in trebly distilled water to give stock solutions of the order of 10-3 M. They were further diluted to give concentra- tions in the range 10-6-10-5 M prior to square-wave voltam- metric determination of their cephalothin content by refer- ence to either standard solutions of cephalothin or to pre-plotted calibration graphs of ip vs. concentration. The rapidity and precision of these square-wave voltammetric determinations when used with the automated sample hand- ling approach described earlier was assessed by evaluation of the throughput of samples and by standard deviation measure- ments, respectively. Adsorptive accumulation of cephalothin was optimised after a study of the effect of accumulation potential and time, and constitution of supporting electrolyte on the square-wave voltammetric reduction signals produced by lo-7-10-5 M concentrations of cephalothin.Determination of Cephalothin in Urine and Plasma Concentrations of cephalothin in the range 10-4-10-6 M were spiked in urine and plasma and subjected to direct determina- tion by application of the sample handling - rapid-scan square- wave voltammetric approach. In the absence of a discernible signal emerging from these studies, spiked 10-4 M concentra- tions in the biological matrices were diluted 100-fold in order to attempt to reduce the biological interferences and still have a 10-6 M concentration of cephalothin which can be detected in aqueous solutions by rapid-scan square-wave voltammetry with or without prior adsorptive accumulation.Finally, for the spiked plasma, the proteins were precipitated with both acetonitrile and HC104, followed by centrifugation and separation of the supernatant, which was assayed for cephalo- thin with and without prior adsorptive accumulation. Results and Discussion Rapid-scan Square-wave Polarography at Single Mercury Drops In single drop square-wave polarography , the dropping-mer- cury electrode (DME) is held at a constant potential, Ei, for a period of time, td, before applying a potential scan consisting of a staircase waveform superimposed on a square wave. If cephalothin is adsorbed at Ei, the surface concentration of cephalothin, To, is given by the Koryta equation? ro = 0.739 ~ p c * ti12 .. . . . . (1) where C* is the bulk concentration of cephalothin, Do is its diffusion coefficient and t is the time from the beginning of the drop life. Equation (1) applies under the condition where all cephalothin molecules arriving at the surface are adsorbed ( i e . , strong adsorption); hence the bulk concentration must be below that at which the surface saturates in the time in question. Under these conditions cephalothin molecules continuously diffuse to the DME from the beginning of the drop life. The solution concentration of cephalothin near the electrode is very low, whereas the surface concentration continues to increase, the drop acting as a pre-concentration site.When the sweep is applied and reduction occurs, almost all current is supplied by adsorbed cephalothin with little contribution from diffusion [Fig. l ( a ) ] . Increase of the bulk concentration increases the contribution of adsorption where- as the diffusion contribution remains small until surface saturation is reached. At this point, the concentration of cephalothin near the surface is no longer negligible and diffusion begins to contribute to. the reduction current. For any further increase in bulk concentration, the contribution of adsorption remains fixed whereas that of diffusion now increases until the reaction appears to be totally controlled by diffusion [Fig.l(b)]. This is graphically illustrated in Fig. 2 for I I T 10 nA 1 T 3 nA 1 -1.1 -0.9 -0.7 -0.5 -1.2 -1.0 -0.8 -0.6 EN and ( b ) 3.5 X 10-5 M Fig. 1. Square-wave voltammogram for cephalothin, pH 2.0. (a) 1 x 1100 990 880 770 660 a 8 550 440 330 220 110 1 I I I l l I 5 10 15 20 25 30 35 40 [Cephalothin)/lO-6 M C,u:.liration graph in SWV for cephalothin in 0.3 M KC1 - HC1, Fig. 2. pH 2.0 cephalothin in the supporting electrolyte 0.3 M KCI with HC1 at pH 2.0. The variation of peak current of cephalothin normalised to the electrode area with pH is given in Fig. 3 for a concentration of 5 X 10-6 M (adsorption control). This figure and a similar one obtained under conditions of diffusion control (7 x 10-5 M) illustrate that the optimum analytical signals under either conditions of adsorption or diffusion control are obtained at a pH of approximately 2.0 and that the cephalothin anion gives a relatively small reduction current at pH 2 7.0.The variation of E, with pH is given in Fig. 4, for these two concentrations of cephalothin, i.e., 5 X 10-6 and 7 x 10-5 M. The fractional surface coverage is given by 8 = A,NTo . . . . . . . . in which A , is the area occupied by a molecule of adsorbed cephalothin and N is the Avogadro constant. On substitution of equation (1), equation (2) becomes 8 = 0 . 7 3 9 ~ , ~ ~ p c* t 1 / 2 . . . . (3)962 ANALYST, JULY 1987, VOL. 112 1 6 i \ I I 1 I I I I 0 2 4 6 8 1 0 1 2 PH Fig. 3. Peak current in SWV normalised to the electrode area for 5 X 10-6 M cephalothin as a function of pH - 1.20 -1.16 -1.12 h -1.08 - 1.04 -1.00 0 2 4 6 8 1 0 1 2 PH Fig.4. Peak potential in SWV as a function of pH for cephalothin. A, 5 x 10-6; and B, 7 x 10-5 M which applies under the same conditions as (1). Substituting the relevant values into this equation under conditions of surface saturation, A, = 160 8L2 which suggests that adsorp- tion of cephalothin in acidic solutions of pH 2.0 occurs through more than one site of the molecule. For a reversible, diffusion controlled single drop square- wave polarographic experiment , with all parameters except electrode area held constant, Ramaley and Tan13 have shown that i,= KIA,= K2tp2'3 . . . . . . (4) where i, is the peak current, A, is the electrode area and tp is the time during the drop life at which i, is measured.K1 and K2 are proportionality constants. For a system in which the reactant is adsorbed, e.g., cephalothin, and when this provides the major contribution to the reduction current, Ramaley et al. 14 have illustrated for the adsorption of picolinic acid complexes of CdII and Pb" that i, = Kdpro = K4(tp2/3) (tp1/2) = Kgtp7/6 . . ( 5 ) This assumes that the peak current is proportional to both electrode area and to surface concentration. The dependence of log i, on log tp is shown in Fig. 5 for concentrations of cephalothin of 5 x 10-6,4 x l O - 5 , 8 X 10-5 and 1 x 10-4 M and the corresponding gradients are 1.33,1.07, 0.85 and 0.80, respectively, showing the transition from adsorption to diffusion control.-1.1 I I I I 1 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Log (rds) Fig. 5. Peak current in SWV as a function of tp for cephalothin. A, 5 x 10-6; B, 4 x 10-5; C , 8 X 10-5; and D, 1 x M. pH 2.0 Rapid-scan Square-wave Voltammetry at the SMDE Cephalothin gives rise to well defined peaks in acidic media (pH 0-2) in the concentration range 10-4-10-7 M on applica- tion of rapid-scan square-wave voltammetry at the SMDE, e.g., at a frequency of 100 Hz (scan rate 200 mV s-1) and an initial potential of -0.8 V, a solution of 5 x 10-6 M cephalothin gives a 350-nA peak at - 1.05 V in 0.1 M HCl and a 612-nA peak at -1.075 V in 0.075 M HC104. In 0.1 M HC1, peak height is linearly related to concentration in the range 5 X 10-7-10-5 M (slope, 645 nA per order of magnitude concentration between 10-6 and 10-5 M) and is limited in height at concentrations higher than 10-5 M.In 0.075 M HC104, peak height is linearly related to concentration in the range 10-7-10-5 M (slope, 960 nA per order of magnitude concentration between 10-6 and 10-5 M) and is also limited in height at concentrations higher than 10-5 M. These results were obtained using the EG & G PAR 384 B polarographic analyser in the square-wave mode with a 20-mV pulse height and 2-mV scan increments, The Model 309 automatic voltam- metric electrode was used in the SMDE mode of operation with a large mercury drop. The significantly higher reduction current of cephalothin in HC104 solutions as compared with HC1 solutions recommen- ded that the former be used as a supporting electrolyte. A concentration of ca.10-1 M was chosen to minimise acid decomposition of cephalothin during the analytical pro- cedures. The optimum instrumental conditions were then chosen from a study of the variation of the peak current of 5 x 10-6 M cephalothin in 0.075 M HC104 with frequency, scan speed, starting potential and pulse height. The highest ip values were found with high frequencies of the order of 100 Hz (200 mV s-1). The peak half-width values remained constant at 65 mV down to a frequency of 10 Hz (20 mV s-1). Initial or starting potentials over the range -0.4 to -0.9 V were investigated and the results are given in Table 1. Although slightly higher currents were observed with less negative initial potentials, best peak definition was found with a starting potential of -0.8 V.As adsorption of cephalothin has a significant effect on the over-all current observed at concentrations < l O - 5 M, it was felt that analytical reproduci- bility at these trace concentrations would be best served by minimising co-adsorption of interferences at lower negative potentials than -0.8 V. Higher currents of approximately aANALYST, JULY 1987, VOL. 112 963 Table 1. Variation of i, of 5 x 10-6 M cephalothin with starting potential (frequency 100 Hz, pulse height 20 mV, 2-mV scan increments) Starting potential i,hA (vs. Ag - AgCl)/V 612 -0.9 616 -0.8 652 -0.6 666 -0.4 Table 2. Variation of 0.01 M HCl supporting electrolyte currents with deposition time at -0.9 V in the presence of 1.7 X 10-7 M cephalothin (not detectable using rapid-scan square-wave voltammetry at the SMDE); frequency 25 Hz; scan rate 50 mV s-1 Current at Current at Deposition time at -0.9 V/ -1.2Vl -1.3Vl S nA nA 0 15 30 60 170 300 500 850 700 1700 2200 3000 factor of 2 were given by increasing the pulse height from 20 to 100 mV without loss of resolution, i.e., the peak half-width of cephalothin remained at 65 mV and its peak width also remained unaltered at 150 mV. Study of the Degradation of Cephalothin in Acidic and Basic Solutions by Rapid-scan Square-wave Voltammetry at the SMDE A solution of 3.5 x 10-5 M cephalothin in 1 M HCl gave rise to a reduction peak due to cephalothin at -1.058 V of height 1250 nA (frequency 100 Hz, scan speed 200 mV s-1) at the start of the degradation reaction.This peak had totally disappeared 20 h later and no new peaks were observed in the potential region -0.600 to -1.400 V.A solution of 1.36 X 10-5 M cephalothin was subjected to degradation in 0.9 M HC104 and the reaction followed by rapid-scan square-wave voltammetry at 100 Hz after dilution of 1-ml aliquots of the reaction mixture to 10 ml with trebly distilled water. The resulting reduction peak of cephalothin occurred at - 1.060 V and had a peak height of 160 nA at the start of the degradation reaction. A tli2 of 7 h was determined. Accumulation at the SMDE of products of the reaction of mercury ions at a low negative potential (i.e., -0.05 V) with any of the electroinac- tive degradation products of cephalothin in acidic media with subsequent cathodic stripping was also attempted but was unsuccessful.A 10-5 M concentration of cephalothin was subjected to degradation in 1 M NaOH at ambient temperature and the reaction followed in situ by rapid-scan voltammetry over the potential region -0.600 to - 1.900 V. The supporting electrolyte, 1 M NaOH, gave rise to interfering peaks at approximately -0.84 and -1.14 V, which also increased in height on accumulation at the indicator electrode at a potential of -0.600 V for 2 min. Over a time period of 20 h, no electroactive degradation products of cephalothin were obser- ved over the potential region -0.600 to -1.900 V. In another experiment on the degradation of cephalothin in 1 M NaOH, 1-ml aliquots of the reaction mixture were diluted 10-fold with either trebly distilled H20 or with 0.1 M acetate buffer of pH 4.6 and examined at 1.36 x 10-6 M concentrations by rapid-scan square-wave voltammetry at the SMDE over the potential range -0.050 to -1.800 V.If the product of degradation of cephalothin in basic media was to behave like degraded benzylpenicillin, which gives cathodic stripping peaks due to mercury and copper complexes of penicill- amine,ls then one might expect a similar phenomenon for cephalothin. After a 2-h reaction time for cephalothin in 1 M NaOH, which is sufficient to significantly degrade cephalo- thin, the voltammogram over the potential region -0.250 to -1.800 V, with 60-s deposition time at the initial potential, showed no new peaks other than interferences from the 0.1 M sodium hydroxide supporting electrolyte at ca. -0.50, -0.79, -1.20 and -1.46 V.A 1-ml aliquot of the reaction mixture after 2-h reaction time was also diluted 10-fold with 0-1 M acetate buffer of pH 4.6 and the resulting solution scanned from -0.150 to -1.500 V with 60-s deposition time at the initial potential. No stripping peaks were detected at ca. -0.3 to -0.4 V with and without a 10-fold molar excess of Cu" as shown by Forsmanls for benzylpencillin degraded in alkaline solution. It is therefore deduced that the alkaline degradation of cephalothin under these experimental conditions yield electroinactive products which are incapable of forming partially insoluble mercury salts at low negative potentials which are then amenable to cathodic stripping voltammetry. Determination of Cephalothin in a Pharmaceutical Preparation Using the Sampling Handling Approach - Rapid-scan Square- wave Voltammetry Keflin (registered by Eli Lilly, Indianapolis, IN, USA, and manufactured by Eli Lilly Italia, Florence, Italy) is manufac- tured in ampoules which contain 1 g of cephalothin and 30 mg of NaHC03.Approximately a 30-mg sample of the ampoule contents is accurately weighed and made up with trebly distilled water to 100 ml in a calibrated flask. Complete dissolution occurs within seconds to give a cephalothin solution which is ca. 0.7 x 10-3 M. This must then be diluted 100-fold so that the final solution (<lO-5 M) gives a peak current which is on the linear portion of the i, vs. concentra- tion calibration graph. Twelve aliquots of this final solution could be subjected to the sample handling approach- rapid- scan square-wave voltammetry at the SMDE in a period of 30 min using two rinse cycles between samples.The Model 384 B polarographic analyser controlled all the mechanical and electrochemical operations in these analyses and printed out the p.p.b. concentrations of cephalothin in the respective samples when compared with a reference standard of cephalo- thin in addition to the voltammograms. A coefficient of variation of 2% was calculated for the repetitive analysis of 12 aliquots of the same final solution. Adsorptive Stripping Voltammetry (AdSV) of Cephalothin at the SMDE Initially AdSV of cephalothin was carried out in HC1 supporting electrolytes. Approximately M concentrations of cephalothin were not detectable in 0.1 or 0.01 M HCl and on application of accumulation times of several minutes at a potential of -0.9 V, the current of the supporting electrolyte decay now steadily increased with accumulation time (Table 2).This would tend to suggest hydrogen evolution from the supporting electrolyte catalysed by adsorbed cephalothin on the SMDE. The same concentration of cephalothin in 0.075 M HC104 (i.e., 1.7 X 10-7 M) was detectable by rapid-scan square-wave voltammetry at the SMDE owing to a significantly higher reduction current in this medium and the peak height of cephalothin was found to increase, together with the current related to the supporting electrolyte decay, for deposition times up to 60 s. At deposition times in excess of 60 s, the peak due to cephalothin merged with the supporting electrolyte decay.Again, hydrogen evolution was catalysed by adsorbed cephalothin and eventually masked the adsorptive stripping signal of cephalothin at accumulation times >60 s using these concentrations of ca. 10-7 M.964 ANALYST, JULY 1987, VOL. 112 Table 3. Variation of adsorptive stripping current of a 5 X solution of cephalothin in 0.075 M HC104 with accumulation time M Deposition time/ Peak potential Peak current*/ S ( VS. Ag - AgC1)N nA 0 - 1.076 666 2 - 1.086 818 5 -1.092 1203 7 -1.102 1170 10 -1.106 1300 * Calculated by the tangent-fit mode of the polarograph. Concentrations of cephalothin in the range lo-8-10-7 M can therefore be determined by this adsorptive stripping tech- nique in perchloric acid supporting electrolytes using accumu- lation times of up to 1 min.Higher concentrations of cephalothin, e.g., 5 x 10-6 M, again in HC104 supporting electrolyte, also show this adsorp- tive stripping phenomenon for accumulation times of up to 10 s (Table 3). Surface saturation at this concentration is reached quickly at an accumulation time of ca. 5 s, according to these results. The applicability of these voltammetric techniques of rapid-scan square-wave voltammetry with and without adsorptive accumulation were then assessed with respect to the determination of cephalothin in biological fluids such as urine and plasma. The direct determination of cephalothin in urine even at concentrations as high as 10-4 M was rendered impossible owing to the presence of very high background currents in the vicinity of the reduction peak of cephalothin. No doubt hydrogen evolution was again responsible, this time catalysed by a combination of endogenous molecules in the urine and any adsorbed cephalothin on the SMDE. A similar situation was encountered with plasma, both in direct determi- nation and with protein precipitation with acetonitrile and perchloric acid followed by separation of the supernatant prior to voltammetric investigation. The determination of cephalo- thin by these rapid-scan square-wave voltammetric techniques with and without adsorption pre-concentration in urine and plasma therefore appears to require a prior chromatographic or solvent extraction step. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Jones, I. F., Page, J. E., and Rhodes, C. T., J. Pharm. Pharmacol., 1968,20,45. Hall, D. A., Berry, D. M., and Schneider, C. J., J. Electroanal. Chem., 1977,80, 155. Yamana, T., and Tsuji, A., J. Pharm. Sci., 1976,65, 1563. Fogg, A. G., Fayad, N. M., Burgess, C., and McGlynn, A., Anal. Chim. Acta, 1979, 108, 205. Benner, E. J., Antimicrob. Agents Chemother., 1970,5,201. Cooper, M. J., Anders, M. W., and Mirkin, B. C., Drug Metab. Dispos., 1973, 1, 659. Buhs, R. P., Maxim, T. E., Allen, N., Jacob, T. A,, and Wolf, F. J., J. Chromatogr., 1974, 99, 609. Nilsson-Ehle, I., Yoshikawa, T. T., Schotz, M. C., and Guzel, B., Antiomicrob. Agents Chemother., 1978, 13,221. Nygard, G., and Khalid, S. K. W., J . Liq. Chromatogr., 1984, 7 , 1461. Yarnitzky, C. N., Anal. Chem., 1985, 57, 2011. Borman, S. A., Anal. Chem., 1982, 54,698A. Koryta, J., Collect. Czech. Chem. Commun., 1953,18, 206. Ramaley, L., and Tan, W. T., Can. J. Chem., 1981,59,3326. Ramaley, L., Dalziel, J. A., and Tan, W. T., Can. J. Chem., 1981, 59, 3334. Forsman, U., Anal. Chim. Acta, 1983, 146,71. Paper A711 7 Received January 20th, 1987 Accepted February 23rd, 1987

ISSN:0003-2654    

DOI:10.1039/AN9871200959

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 965 Voltammetric Behaviour of Nitrofurazone, Furazolidone and Other Nitro Derivatives of Biological Importance Alfonso Morales, Pablo Richter and M. lnes Toral Department of Chemistry, Faculty of Sciences, University of Chile, Las Palmeras 3425, P.O. Box 653, Santiago, Chile In pyridine - formic acid buffer and tetramethylammonium chloride solution of pH 4.5 at a dropping mercury or a glassy carbon electrode, nitrofurazone, furazolidone and nitrofurantoin are reduced in a single six-electron wave, while chloramphenicol and other structurally related nitro derivatives are reduced in only one four-electron wave, the nitro group being reduced to the amine or to the hydroxylamine, respectively. The electrochemical behaviour of these compounds depends mainly on the nature and position of the substituents.Reduction to the primary amine occurs when the substituents possess available JC electrons to conjugate with the nitro group of the aromatic ring, which determines the transformation of the hydroxylamine into the aminevia formation of a highly reducible intermediate imine or a quinonoid structure. In contrast, if the formation of the intermediate imine is made impossible by an adverse effect of the substituent, the hydroxylamine does not undergo further reduction. Cyclic voltammograms were recorded at different pH values and at different scan rates in order to identify certain relatively unstable species. The effect of pH on the diffusion-limited current and on the €&values of the polarographic waves was also studied and the results obtained were compared with those obtained by cyclic vol ta m met ry.On this basis, and according to the polarographic and cyclic voltammetric data, a reduction mechanism for the nitrofuran derivatives is suggested, in which the importance of the homogeneous chemical reactions associated with the electron-transfer steps is examined. Keywords: Nitrofurazone, furazolidone and chloramphenicol; nitro derivatives of ?,4-benzodiazepines; polarography; cyclic voltammetry Nitrofurazone (I) , furazolidone (11) and nitrofurantoin (111) are structurally related, with a nitro group at the 5-position on the furan ring. The electrochemical behaviour of these nitrofurans and other aromatic nitro derivatives such as chloramphenicol, clonazepam, nitrazepam, flunitrazepam and parathion is based on the ease of reduction of the nitro group at a dropping mercury or solid electrode.0 H H II C=N-N -C-NHp Nitrofurazone Polarography has been widely used in order to elucidate the reduction mechanism at a dropping mercury electrode and to investigate the resemblance with the metabolic pathway for the biological degradation of these nitro derivatives. These compounds are generally metabolised in vivo to the corre- sponding amines via nitroso and hydroxylamine interme- diates. However, the polarographic techniques cannot deter- mine these metabolites owing to the ill-defined reduction wave of the hydroxylamine, the inability of the amine group to react at the dropping mercury electrode and because the reduction of the nitroso group to hydroxylamine occurs at a more positive potential than the reduction potential of the nitro group to hydroxylamine, and consequently that reduc- tion wave is not detected.Further, the variable number of reduction waves for each compound, depending on the supporting electrolyte, pH and maximum supressors employed, illustrates the complexity of the electrode processes involved. In Britton - Robinson buffer at pH values below 5 , the nitro group of the nitrofurantoin is reduced to hydroxylamine in a four-electron process and subsequently to the amine in a two-electron process. 1 Similarly, nitrofurazone shows a two-step reduction wave in different supporting electrolytes, the second wave being attributed to a simul- taneous reduction of the hydroxylamine and the azomethine group .2 The determination of some nitrated heterocyclic com- pounds containing similar types of reduction sites to that of nitrofurazone was studied earlier by Vignoli et al.,3 using a Britton - Robinson buffer of pH between 1.81 and 11.98.They observed two waves for the reduction of the nitro group and three waves at lower pH values when the compounds had a substituted imino group. A polarographic method has been used to determine furazolidone in feed pre-mixes,4 but little attention has been paid to the electrochemical behaviour of this compound. In previous polarographic work5 in which a solvent - buffer system containing pyridine and formic acid in conjunction with tetramethylammonium chloride solution was used as the supporting electrolyte, it was found that the nitro group of the nitrofurantoin shows only one reduction wave corresponding to a six-electron process, and that chloramphenicol and other structurally related compounds are reduced in a single four-electron wave.The aim of this work was to study the electrochemical behaviour of molecules having the same electroactive group in order to elucidate the effects of the nature and position of the substituents on the reduction. Significant aspects involved in966 ANALYST, JULY 1987, VOL. 112 ~~ Table 1. Voltammetric data for reduction of I and 11 in pyridine - Scan rate/ Compound Vs-l i&A idCd 0.196m~I . . . . 0.020 0.050 0.100 0.200 0.300 0.400 0.476m~I . . . . 0.020 0.050 0.100 0.200 0.300 0.400 5.60 202 9.20 210 13.20 213 17.60 201 20.80 194 23.60 191 14.4 214 22.4 211 33.2 220 43.2 203 51.6 198 59.2 197 formic acid with TMAC as supporting electrolyte Scan rate/ Compound V s-1 i&A 0.196m~II ., . . 0.120 5.8 0.050 9.2 0.100 12.8 0.200 18.0 0.300 21.6 0.400 24.2 0.476m~II . . . . 0.020 14.4 0.050 22.4 0.100 32.0 0.200 44.2 0.300 55.2 0.400 60.4 iJCd 209 210 207 205 201 195 214 21 1 213 212 212 201 Table 2. Polarographic data for the reduction of nine nitro compounds (0.124 mM) in pyridine - formic acid with TMAC as supporting electrolyte Compound p-Nitrophenol . . Nitrofurantoin . . Nitrofurazone . . Furazolidone . . Chloramphenicol Nitrazepam . . Flunitrazepam . . Clonazepam . . Parathion . . . . No. of runs . . . . 7 . . . . 8 . . . . 7 . . . . 7 . . . . 9 . . . . 5 . .. . 6 . . . . 7 . . . . 8 E4JV -0.50 -0.16 -0.18 -0.17 -0.41 -0.34 -0.28 -0.28 -0.32 i&A 2.38 2.27 2.30 2.29 1.62 1.62 1.55 1.58 1.64 the reduction were examined, together with the homogeneous chemical reactions accompanying the electrode process. Experimental Reagents All chemicals were of analytical-reagent grade unless stated otherwise. Nitrofurazone (I) , furazolidone (11) and nitrofurantoin (111) (Sigma Chemical, St. Louis, MO, USA) were used for the basic studies. Standard solutions (1.0 x 10-2 M) were prepared by dissolving the appropriate amount of each drug in dimethylformamide (DMF). Gelatine solution (0.5%) was used as a maximum suppressor. The supporting electrolyte was prepared by diluting 20 ml of pyridine (12.3 M) and 10 ml of formic acid (98-100°/~) with 120 ml of 0.1 M tetramethylammonium chloride (TMAC) solution.The resulting solution had a pH of 4.5. On varying the ratio of formic acid to pyridine the pH could be varied over the range 2.65.1.5 Other nitro derivatives, such as chloramphenicol, p-nitro- phenol , nitrazepam , flunitrazepam, clonazepam and para- thion (EPA Research, Triangle Park, NC, USA), were also dissolved in DMF. Apparatus Polarographic assays were performed using a Polariter PO4 instrument (Radiometer, Copenhagen, Denmark). A drop- ping mercury electrode was used as the working electrode and a saturated calomel electrode (SCE) as the reference elec- trode. Cyclic voltammetric experiments were performed using a CV-27 voltammograph (Bioanalytical Systems, Lafayette, IN, USA). A three-electrode assembly was used for all measure- ments.Glassy carbon was employed as working electrode, an SCE as the reference electrode and a platinum coil as the counter electrode. An Orion Research Digital Ion-Analyzer 701 with glass and SCE electrodes was used for pH determinations. Techniques Aliquots of the standard solutions were diluted with the supporting electrolyte, de-oxygenated with oxygen-free nit- rogen and analysed using the d.c. polarographic mode. The mercury flow-rate, m, and the drop time, t, were determined at various heights of the mercury column, h. The diffusion- controlled character of the current and the dependence of the diffusion-limited current on the depolariser concentration were established. Cyclic voltammetric experiments were carried out under identical experimental conditions.All measurements were performed at 25 & 1 "C. Dissolved air was removed from the solutions by bubbling oxygen-free nitrogen through the cell for 10 min, then passing it over the solution during the electrolysis. Voltammograms were recorded at scan rates between 0.02 and 0.4 V s-1. The current function i,/Cvi was found to be fairly constant for I and I1 (Table 1). pH Studies The effects of pH on the half-wave potentials and diffusion- limited current for I and I1 at a concentration 0.124 mM were studied over the pH range 1-14. The corresponding voltam- mograms were recorded under identical conditions. Results and Discussion Under the experimental conditions described above , the polarographic reduction of some aromatic nitro compounds of biological importance was found to give rise either to a single well defined wave corresponding to a six-electron process or to a single wave corresponding to a four-electron process.The electrochemical behaviour depends on the nature of the aromatic ring and on the nature and position of the substitu- ents. When a solvent - buffer system (pH 4.5) containing pyridine and formic acid in conjunction with tetramethyl- ammonium chloride solution was used as the supporting electrolyte, compounds containing a nitro-substituted furan ring behave as p-nitrophenol and are reduced to the corre- sponding amine in a six-electron reaction in a single step. Reduction of such compounds, except p-nitrophenol, occurs at relatively lower negative potentials than that of other nitro compounds, which indicates some nitroso character (Table 2).When equimolar solutions of 1-111 were polarographed using the supporting electrolyte mentioned above, the ratio of wave heights was approximately 1.00 +_ 0.04, indicating a reduction process similar to that for the corresponding primary amine (Fig. 1). Using controlled potential electrolysis, Mishra and Gode2 recently demonstrated that the ultimate reduced product of nitrofurazone is the primary amine. Reduction by six electrons in a single step occurs only if the substituents possess available x electrons to conjugate withANALYST, JULY 1987, VOL. 112 967 4 . *.’ C 2 3 u T I I I 0.0 -0.2 -0.4 E N Fig. 1. Polarographic reduction waves of nitrofurazone, furazoli- done and nitrofurantoin, each at 0.124 mM.(A) Nitrofurazone: E+ -0.18 V and id 2.3 PA. (B) furazolidone: E+ -0.17 V, id 2.29 PA. (C) Nitrofurantoin: Ei -0.16 V, id 2.27 FA. Vi = 0.00 V the nitro group of the aromatic ring, which makes the transformation of the hydroxylamine into the imine or a quinonoid structure possible. This substituted donor group must be located at the 2-position on the furan ring or in apara or ortho position on the benzene ring in order to achieve the interaction of the n; systems of the aromatic ring and of the substituents with the intermediate hydroxylamine to give the corresponding imine, which is then reduced to the primary amine. Electrochemical reduction of 1-111 in a single wave can be explained regardless of the electron transfer process by the extremely fast chemical reactions occurring, owing to the presence of the moiety >C=N-N<. The following scheme represents the mechanism of formation of the corresponding imine from the intermediate hydroxylamine.H H- \ /’, C=N-N- HO H nu u p-Nitrophenol, p-nitroaniline, nitrosophenols and other aromatic nitro compounds containing similar types of reduc- tion sites and a donor group substituent show analogous polarographic behaviour.Sg 0. H ( a ) A , B I I 0.6 0.0 -0.6 -1.2 0.6 0.0 -0.6 E N 3, 0.6 0.0 -0.6 Fig. 2. Cyclic voltammograms of (a) nitrofurazone; ( b ) furazoli- done; and (c) nitrofurantoin, each at 0.476 mM. pH, 4.5. Glassy carbon electrode. Scan rate. 0.1 V s-1 Cyclic voltammograms of the nitrofuran derivatives (Fig. 2) were recorded under identical conditions, in order to identify intermediate species.In all instances the scan is initiated in a negative direction from 0.0 V. The initial reduction peak A corresponds to a six-electron reduction of the nitro group to the amine derivative, as shown below. H 2 N e C = N - N - + 2H20 I The amine thus produced is subsequently oxidised in the reverse scan at peak B to the imine or quinonoid structure intermediate. This imine is hydrolysed to a quinone deriva- tive, which is neither oxidised nor reduced at these potentials, as shown in Scheme 1. It can be observed that the electron transfer precedes the chemical reaction (EC reaction). Similar behaviour has been shown to occur in the oxidation of p-aminophenol at a platinum electrode in aqueous solutions.1 0 ~ 1 Wave clipping, that is, reversal of the scan direction before peak A, causes peak B to disappear, indicating that peak B is the oxidation product of the primary amine previously formed in A. At scan rates higher than 0.3 V s-1 a second cathodic peak C appears (Fig. 3), indicating that a reversible reduction of the imine derivative occurs, and that the hydrolysis of this imine is too slow to affect the reduction process. In other words, if the scan rate is very high relative to k , very little imine will be lost to the succeeding hydrolysis reaction and the electrochemical process will be reversible (see Fig. 3). Conversely, if the scan rate is low relative to k , the chemical reaction will be essentially over before the voltage scan is reversed, and the electrode process will appear totally irreversible.The reduction potential of the highly reducible intermediate imine is more positive than the reduction potential of the nitro group to amine (peak A) and consequently this wave is not observed in normal d.c. polarography. On the other hand, it was observed that chlorampheni- c0l~~J2-1~ nitrazepam,l5 flunitrazepam and parathion show a different voltammetric behaviour to the nitrofuran deriva- tives. The former are reduced in a single four-electron wave + HN C-N=N- + 2e- + H+ Scheme I968 ANALYST, JULY 1987, VOL. 112 A 0.6 0 -0.6 EN Fig. 3. Cyclic voltammogram of nitrofurazone. Scan rate, 0.3 V s-l; other conditions as in Fig. 2 A 0.6 0.0 -0.6 -1.2 EN Fig. 4. Cyclic voltammogram of chloramphenicol. Conditions as in Fig.2 0 -0.2 -0.4 -0.6t I , , , , I 0 2 4 6 8 1 0 1 2 1 4 PH Fig. 5. Variation of E with pH. (a) Nitrofurazone and (b) furazolidone, each at 0.124 mM corresponding to the reduction of the nitro group to the hydroxylamine. This hydroxylamine derivative does not undergo further reduction because the formation of the highly reducible intermediate imine is blocked by an adverse effect of the substituents. Fig. 4 shows a typical cyclic voltammogram for these compounds in pyridine - formic acid. The three peaks observed correspond to the following well known electrode process: Peak A: ArN02 + 4e + 4H+ + ArNHOH + H20 Peak B: ArNHOH + ArNO + 2e + 2H+ Peak C: ArNO + 2e + 2H+ + ArNHOH Kissinger and Heineman16 showed that the three peaks observed in the voltammogram of chloramphenicol in an acetate buffer system and using a carbon paste electrode involve more than a simple electron transfer. The irreversibil- ity of peak A is due to the slow electron transfer occurring in the step nitro + nitroso derivative.12-15 In addition to these aspects, the buffer constituents of the supporting electrolyte also seem to have a significant effect on the shape and number of the reduction waves reported.We have observed that in a well buffered medium such as that constituted by pyridine and formic acid, in the acidic range single waves were always obtained in the reduction of all the nitro derivatives studied. On the other hand, in poorly buffered media or in Britton - Robinson buffer, nitrofuran derivatives are reduced to hydroxylamine, which is further reduced to the amine in a second separate wave.1-3 According to Hess,12 chloramphenicol is reduced in two steps in phthalate buffer of pH 4 and a similar reduction in two separate waves in acetate media has been reported.13 pH Studies In d.c.polarography, the half-wave potentials for I and I1 are pH dependent and are shifted cathodically with increasing pH. The E+ versus pH graph (Fig. 5) shows three linear portions. The equations that describe the variations of E4 with pH were deduced from the graph and are given in Table 3. The diffusion-limited current for both compounds is also pH dependent (Fig. 6). The slight decrease in the wave height below pH 3 is probably associated with an acid - base equilibrium as previously reported for nitrofurantoin.175 Above pH 5.0 the wave slowly begins to decay for I and 11, and at this pH the first break on the Ei versus pH graph occurs, representing the pH at which the hydroxylamine intermediate in the reduction of the nitro group is no longer protonated and therefore cannot be easily reduced. At approximately pH 8.5 for I and at pH 8.8 for 11, each wave falls sharply and breaks up into two waves. This fall is accompanied by a change in the slope of the E+ versus pH graph, indicating that a different electrode process occurs. Therefore, for all these compounds the best defined and differentiated waves for analytical purposes were obtained at 3 < pH < 5. The scission of the polarographic wave and the change in the slope of the E+ versus pH graph observed at pH 8-9 can be related to the cyclic voltammetric behaviour.Possibly an increase in pH increases the dissociation constant of the protonated species and these factors affect the protonation rate and consequently the E+ values of the reduction wave are shifted to more Table 3. Equations of the half-wave potentials of I and I1 at different pH values. The first term of each equation is the slope of the line and the second is the intercept of the line with the potential axis. Substitution of the appropriate pH value will give the E, value at that pH I I1 PH Equation PH Equation 0.0-5.12 E+ = -0.051pH + 0.050V 0.0-5.5 E* = -0.045pH + 0.024V 5.12-8.43 E*= -0.075pH + 0.173V 5.5-8.79 E*= -0.061pH + 0.112V 8.43-14 E4 = -0.0089pH - 0.384V 8.79-14 Eb = -0.0080pH - 0.354VANALYST, JULY 1987, VOL.112 969 3 - .-D 1 I I I I 1 I “‘I --\ anion radical formed in C. The existence of this relatively stable anion radical in alkaline media has been recently reported in the reduction of nitrobenzene when platinum, gold and glassy carbon electrodes were used.18 The formation of the nitro anion radical can be explained by delocalisation of the electrons in the aromatic ring due to the low proton activity in the bulk of the solution. Detection of the nitro anion radical is possible only through cyclic voltammetry. Fig. 7. Cyclic voltammogram of 0.476 mM nitrofurazone at pH 10. D was obtained when the sweep was reversed after the first reduction peak. Other conditions as in Fig. 2 A f 2 2 3 u I I I 1 negative potentials. Such a situation occurs when protonation precedes electron transfer.17 Cyclic voltammetric studies of the effect of pH on the reduction of the nitrofurans were carried out in order to relate them to the polarographic data obtained. In the pH range 1-8.5 no difference was observed in the shape of the cyclic voltammetric waves (Fig. 2), except that the potentials were shifted cathodically as the pH increased, indicating a similar electrode process over the pH range. However, above approximately pH 8.5 a noteworthy difference appears, indicating that a different process is occurring. For 1-111 a one-electron reduction peak appears at -0.5 V, correspond- ing to the reversible reduction of the nitro group to a nitro anion radical. This peak precedes the main reduction wave, which is shifted, at these pH values, to a more negative potential (Fig.7). Wave clipping at a potential between peaks C and A (see Fig. 7) causes peaks A and B to disappear, indicating that peak D corresponds to the oxidation of the the supporting electrolyte used, as shown by the i,,,h-t and i,/CvJ relationships. The irreversibility of the electrode process was verified by logarithmic analysis of the wave. The slope of the E versus log (did - i ) graph exceeds appreciably 59/n mV and the numerical value of Ea - E2 exceeds 54.6/n mV.19 The an, values (where a is the transfer coefficient) and the number of protons (p values) corresponding to the rate- determining step were calculated for I and I1 at selected pH values. At pH 4.5 the an, values for I and I1 were found to be 1.23 and 1.15, respectively, indicating that two electrons take part in the rate-determining step of the reaction.dE4 - -0.059 P ---. dpH an, p was found to be 0.93 and 0.88 for I and 11, respectively, showing that one proton is involved in the rate-determining step of the reaction over the pH range 1-8. The participation of the hydrogen ion in the rate-determining step is due to protonation of the nitro group to form a more readily reducible species, which is reduced to the nitroso intermediate (CE reaction). A similar stoicheiometry of the rate-determin- ing step for the reduction of nitroxazepine hydrochloride has been reported recently (Scheme 2).20 The nitroso intermediate group is rapidly reduced to the hydroxylamine, which is stabilised at this stage whenever the fast chemical reactions that would allow the transformation of this hydroxylamine into the highly reducible intermediate imine are inhibited.This inhibition takes place when the substituent does not have donor properties, in which event a four-electron reduction occurs. This was observed in com- pounds such as chloramphenicol, parathion, nitro derivatives of 174-benzodiazepines and other structurally related sub- stances (CEE reaction). The nitroso - hydroxylamine revers- ible couple can be detected by cyclic voltammetry only when the hydroxylamine is the ultimate reduction product (Fig. 4). 0 0 0 -(/ 2e-, H- H H<o\irN=O + H20 -N-N=C <o;-Y-N=C I H n 2e-,2H- Scheme 2970 ANALYST, JULY 1987, VOL. 112 It must be stressed that the nitroso intermediate is not observed when normal d.c.polarography is used because the reduction potential of the nitroso group (E2) is more positive than that of the nitro group. On the other hand, in the reduction of the nitrofuran derivatives due to the effect of the donor properties of the substituent by a EC reaction, the hydroxylamine undergoes very fast chemical reactions, protonation and loss of a water molecule, giving rise to the highly reducible intermediate imine, which is then reduced to the primary amine in a reversible process (Fig. 3). This reversible couple was observed by cyclic voltammetry at high scan rates. At slow scan rates this reaction appears to be totally irreversible owing to the hydrolysis of the imine to give an inert product (Figs. 2 and 3) as indicated above.The reactions are shown below. - N - N = E e /” H+ \ I ”\ -‘ OH I I w It must be noted that the potential E3 is more positive than El and therefore only one wave is observed. It may be concluded that the hydrolysis of the imine derivative is too slow to affect the polarographic wave. The above-mentioned processes, as already stated, take place in a well buffered medium of pH 4.5. In the pH range 9-14 the potentials of I and I1 are shifted cathodically and one reversible couple appears, at more positive potentials, corresponding to the reversible reduction of the nitro group to a nitro anion radical derivative (Fig. 7). Conclusions The voltammetric behaviour of some aromatic nitro com- pounds of biological interest, usually prescribed as therapeutic agents, has been studied by polarography and cyclic voltam- metry.In pyridine - formic acid buffer, in the acidic range, the reduction of the nitro group occurs in a single step indepen- dently of the structure of the compound. The best defined and differentiated d.c. polarographic waves, for analytical pur- poses, were obtained at 3 < pH < 5. As is apparent from Table 2, the difference in the half-wave potentials makes the simultaneous determination of different nitro derivatives possible. Cyclic voltammetry was used to identify certain interme- diates, metabolites and final products when reducing, under similar conditions, nitro derivatives having different substitu- ents. Hence the formation of the intermediate nitroso group in the reduction of the nitro compounds, which are reduced to hydroxylamine, has been clearly demonstrated, together with the formation of the intermediate imine when the nitro compounds are reduced to amines.The donor properties of the substituent drive the reduction completely to the primary amine, whereas a substituent that is not a donor promotes reduction to hydroxylamine. Cyclic voltammetry also proved to be useful for the diagnosis of the electrode reactions that are coupled with homogeneous chemical reactions. Support from the Department of Investigation (DIB) of the University of Chile is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Burmicz, J. S., Smyth, W. F., and Palmer, R. F., Analyst, 1976, 101, 986. Mishra, A. K., and Gode, K. D., Analyst, 1985, 110, 1373. Vignoli, L., Cristau, B., Gouezo, F., and Fabre, C . , Chim. Anal. (Paris), 1963, 45, 439. Slamnik, M., Talanta, 1974, 21, 960. Morales, A., Toral, M. I., and Richter, P., Analyst, 1984, 109, 633. Chodkowski, J., and Gralewska-Ludwicka, D., Pol. J. Chem., 1980, 54, 567. StoEesovB, D., Collect. Czech. Chem. Commun., 1949,14,615. Testa, A. C., and Reinmuth, W. H., J. Am. Chem. SOC., 1960, 83, 784. Nicholson, R. S., and Shain, I., Anal. Chem., 1965, 35, 190. Shearer, C. M., Christenson, K., Mukherji, A., and Papariello, G. J., J. Pharm. Sci., 1972, 61, 1627. Bard, A. J., and Faulkner, L. R. “Electrochemical Methods. Fundamentals and Applications,” Wiley, New York, 1980. Hess, G. B., Anal. Chem., 1950, 22, 649. Fossdal, K., and Jacobson, E., Anal. Chim. Acta, 1971, 56, 105. Van Der Lee, J. J., Van Bennekom, W. P., and De Jong, H. J., Anal. Chim. Acta, 1980, 117, 171. Halvorsen, S., and Jacobsen, E., Anal. Chim. Acta, 1972, 59, 27. Kissinger, P. T., and Heineman, W. R., J. Chem. Educ., 1983, 60,702. Zuman, P., “The Elucidation of Organic Electrode Processes,” Academic Press, New York, 1969. Rubinstein, I., J. Electroanal. Chem., 1985, 183, 379. Meites, L., Editor, “Polarographic Techniques,” Interscience, New York, 1965, p. 289. Mishra, A. K., and Gode, K. D., Analyst, 1985, 110, 31. Paper A61415 Received October 30th, 1986 Accepted February 12th, 1987

ISSN:0003-2654    

DOI:10.1039/AN9871200965

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, JULY 1987, VOL. 112 971 Polarog Pharma raphic Determination of Nitrofurazone and Furazolidone in ceutical Formulations and Urine Alfonso Morales, Pablo Richter and M. lnes Toral Department of Chemistry, Faculty of Sciences, University of Chile, Las Palmeras 3425, P.O. Box 653, Santiago, Chile In pyridine - formic acid buffer and tetramethylammonium chloride solution over the pH range 0-8.5, nitrofurazone, furazolidone and other nitrofuran derivatives are reduced in a six-electron process, giving rise to a well defined polarographic reduction wave at a dropping mercury electrode. The current is diffusion controlled and proportional to the concentration from about 5.88 x 10-3 M to the limit of detection of 1.24 x 10-6~. The proposed method permits these drugs to be determined, without any prior separation or extraction, in pharmaceutical formulations and in urine at levels at which the unchanged drugs are excreted. Other drugs commonly used as therapeutic agents, such as nitro derivatives of 1 ,4-benzodiazepinesr chloramphenicol, metronidazole and tinidazole, are reduced in a single four-electron process at more negative potentials, which makes simultaneous determinations possible.Keywords : Nitro fu razone de te rm ina tion; fu razolidone determination; pola rog rap h y; pha rmaceu tical formulations; urine Nitrofurazone [5-nitro-2-furaldehyde semicarbazone] (I) and furazolidone [3-(5-nitrofurfurylideneamino)-2-oxazolidone] (11) are synthetic nitrofuran derivatives with a nitro group at the 5-position on the furan ring.They have been widely used in the treatment of caecal coccidiosis in chickens and necrotic enteritis in swine. They are generally added to animal feeds to prevent various poultry and swine diseases. In man these drugs are therapeutically effective as antibacterial and bacteri- cidal agents. Structures of the compounds are shown in the preceding paper, p. 965. The exact mechanism of action of nitrofurazone is not known. It appears, however, that the drug acts by inhibiting bacterial enzymes involved in carbohydrate metabolism. Furazolidone is bactericidal owing to its interference with several bacterial enzyme systems, possibly including preven- tion of acetylation of coenzyme A. Furazolidone also acts as a monoamine oxidase inhibitor. These compounds have been determined by spectro- photometry ,l conductimetry,2 gas - liquid chromatography,3 high-performance liquid chromatography4 and iodimetry.5 A survey of the literature indicates that very little attention has been paid to the polarographic determination of these drugs.69 This paper reports a method for the polarographic determination of nitrofurazone and furazolidone in phar- maceutical formulations and urine.Preliminary experiments showed that , in well buffered solutions and using a donor-active solvent, the reduction of aromatic nitro compounds in a single four- or six-electron process depends on the nature and position of the substituents on the aromatic ring, as discussed in the preceding paper.10 Experimental Reagents All chemicals were analytical-reagent grade reagents.Stock solutions (10-2 M) of nitrofurazone and furazolidone (Sigma Chemical, St. Louis, MO, USA) were prepared by dissolving the appropriate amount of each drug in dimethyl- formamide (DMF). Tablets containing these drugs were dissolved in DMF and assayed polarographically. Gelatine solution (0.5%) was used to eliminate the polarographic maxima. The supporting electrolyte used contained 0.1 M tetramethylammonium chloride solution and pyridine - formic acid buffer (pH 4.5) and was prepared as described pre- viously. 11 Apparatus and Conditions for Polarographic Analysis Polarographic analysis was performed using a Polariter PO4 instrument (Radiometer, Copenhagen, Denmark). A satu- rated calomel electrode (SCE) was used as a reference electrode together with a dropping mercury electrode (DME) as the working electrode.Dissolved air was removed from the solutions by bubbling oxygen-free nitrogen through the cell for 10 min. All measurements were performed at 25 k 1 "C. Procedure Aliquots of the stock solutions were diluted in 15 ml of supporting electrolyte, 1 ml of gelatine solution was added and the solution was purged with oxygen-free nitrogen for 10 min and subjected to polarography in the d.c. mode. pH Studies The effect of pH in the range 1-14 on the limiting current for I and I1 at a concentration of 0.124 mM was studied. From these polarographic data, the optimum pH for the determination of the drugs was chosen. Calibrations Graphs Aliquots of the pure drugs dissolved in DMF were diluted in the supporting electrolyte and polarographed. The standard additions method was employed in all instances and the results were used in the preparation of limiting current versus concentration graphs.Determination of I and I1 in Pharmaceutical Formulations An ophthalmic solution containing a nominal 20 mg per 100 ml of I was used for analysis. Portions of 1.0 ml of the ophthalmic solution were transferred into separate 10-ml calibrated flasks and diluted to the mark with the supporting electrolyte and assayed polarographically . Tablet formulations containing a nominal 100 mg of I1 in a total mass of approximately 378 mg were analysed in order to examine the applicability of the proposed method. Ten tablets were thoroughly ground and mixed. Samples of approxi- mately 20 mg of I1 were accurately weighed, dissolved in972 ANALYST, JULY 1987, VOL.112 Table 1. Effect of nitrofurazone and furazolidone concentrations on id values Concentration (C)/ lop6 M 1.24 2.44 3.61 12.4 24.4 36.1 58.8 124 361 588 Nitrofurazone id/pA 0.024 0.048 0.066 0.23 0.45 0.675 1.10 2.30 6.84 10.9 id/C 19.35 19.67 18.28 18.54 18.44 18.70 18.71 18.60 18.94 18.53 Furazolidone - id/pA idle 0.023 18.55 0.045 18.44 0.067 18.56 0.225 18.30 0.45 18.44 0.65 18.00 1.08 18.37 2.29 18.46 6.51 18.03 10.5 17.85 DMF, transferred into separate 10-ml calibrated flasks and diluted to the mark with DMF. The contents of the flasks were shaken for 20 min and then allowed to settle. A 0.1-ml aliquot of the clear supernatant liquor was diluted to 25 ml with the supporting electrolyte and a portion of this solution was subjected to polarography.By reference to calibration graphs the concentrations of the I and I1 in each sample were calculated. Determination of I1 in Urine A calibration graph was constructed in accordance with the limits at which the unchanged drug is excreted. Various amounts of I1 were added to a fixed volume of urine and aliquots of these spiked urine samples were diluted with the suppoting electrolyte and polarographed. The calibration graph in the range 3-60 pg ml-1 was a straight line passing through the origin. Results and Discussion Polarograms of nitrofurazone and furazolidone recorded in the proposed supporting electrolyte (pH 4.5) exhibit a single well defined wave with half-wave potentials of -0.18 and -0.17 V, respectively.The mercury flow-rate, m, and drop time, t, were deter- mined at various heights of the mercury column, h. This was repeated for several concentrations and the value of ih-4 was found to be constant, indicating that the current is diffussion controlled. The relationship between limiting current and concentration of each depolariser was found to be linear over a wide range of concentration (Table 1). The limits of detection were 0.24 and 0.28 pg ml-1 for I and 11, respectively. Above 1 X 10-3 M of either drug the id/c relationship was not linear, probably owing to adsorption of the drugs at the mercury drop surface. At concentrations below 1.5 yg ml-1 the supporting electrolyte was used as a blank. No maximum was observed at concentrations of the drugs below 1 X 10-5 M and no gelatine was added.The precision of the polarographic method was tested with solutions of similar concentration and the standard deviation of the ratio of diffusion current to sample mass was found to be satisfactory. For I and 11, the best defined and differentiated waves for analytical purposes were obtained at 3 < pH < 5 in the proposed supporting electrolyte. 10 On the basis of these results, I and I1 were determined in pharmaceutical formulations. Ten polarographic assays on an ophthalmic solution containing 20 mg per 100 ml of I gave a mean value of 19.78 mg per 100 ml with a relative standard deviation of 1.39%. Tablets containing a nominal 100 mg of I1 were also analysed. Ten assays were carried out, giving a mean value of 98.66 mg per tablet with a relative standard deviation Of 1.05%. f 2 ?2 3 u I 0 -0.3 -0.6 E N Fig.1. Polarographic reduction waves of nitrofurazone and chlor- amphenicol, each at 0.122 mM. (A) Nitrofurazone, E4 -0.18 V, id2.10 PA. (B) Chloramphenicol, E* -0.41 V, id 1.42 p4. Vi = 0.0 V I I 0 -0.3 -0.6 E N Fig. 2. Polarographic reduction waves of furazolidone and chlor- amphenicol, each at 0.122 mM. (A) Furazolidone, Et -0.17 V, id 2.09 PA. (B) Chloramphenicol, E+ -0.41 V, id 1.42 PA. Vi = 0.0 V 0 -0.2 -0.4 E N Fig. 3. Polarographic reduction waves of furazolidone and clona- ze am, each at 0.122 mM. (A) Furazolidone, E, -0.17 V, id 2.07 FA. (I37 Clonazepam, E$ -0.28 V, id 1.43 pA. Vi = 0.0 V The precision of the method for the determination of I1 in urine was checked using different spiked urine samples over a wide concentration range. The results showed a relative standard deviation of 3.2% at the lowest concentration levels and 1.92% at concentrations of 10 yg ml-1 and higher.Urine samples were obtained from patients at specific time intervals during single- and multiple-dose administration. It was observed that after administration of a single oral dose of 200 mg the apparent drug concentration in urine increases until it reaches 8% of the initial dose at 2 h and then begins to decrease. On administration of 200 mg three times daily, approximately 15-20% of a total daily dose is excreted as unchanged drug. After 8-10 h the polarographic signal disappears and only non-nitro metabolites are excreted. The main non-nitro metabolites are voltammetrically active amines that exhibit well defined anodic waves.The differencesANALYST, JULY 1987, VOL. 112 973 between polarography and voltammetry arise because polaro- graphy determines only the nitro moiety whereas voltammetry permits the determination and detection of other metabol- ites.10 Because of the carcinogenicity and other adverse effects of nitrofurazone when administered in large oral doses, the determination of this drug in normal urine samples was not possible. However, the calibration graph constructed using spiked urine samples was linear in the range 3-60 pg ml-1. When a solvent - buffer system containing pyridine and formic acid in conjunction with tetramethylammonium chloride solution is used as the supporting electrolyte, these nitrofurans are reduced in a single six-electron wave whereas other nitro derivatives are reduced in a single four-electron wave.Reduction of these nitrofurans occurs at lower negative potentials than for other nitro compounds commonly used as therapeutic agents, such as chloramphenicol, tinidazole, metronidazole and some nitro derivatives of 1,4-benzodia- zepines. The large difference in E+ values makes simultaneous determinations possible, as can be seen in Figs. 1-3. Support from the Department of Investigation (DIB) of the University of Chile is gratefully acknowledged. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Herrett, R. J., andBuzard, J. A.,Anal. Chem., 1969,32,1676. Egerts, V., Simanska, M. V., and Hillers, S., Latv. PSR Zinat. Akad. Vestis, Kim. Ser., 1963,2, 177. Ryan, J. J., Lee, Y. C., Du Pont, J. A., and Charbonneau, C. F., J. Assoc. Off. Anal. Chem., 1975, 58, 1227. Sudgen, E. A., MacIntosh, A. I., and Vilim, A. B., J. Assoc. Off. Anal. Chem., 1983, 66, 874. Rao, G. R., Raghuveer, S., Murthy, S. S. N., and Bajrangrao, B., Indian Drugs, 1979, 17, 50. Vignoli, L., Cristiau, B., Gouezo, F., and Fabre, C., Chim. Anal. (Paris), 1963, 45, 439. Vignoli, L., Cristiau, B., Gouezo, F., and Fabre, C., Chim. Anal. (Paris), 1963, 45, 499. Mishra, A. K., and Gode, K. D., Analyst, 1985, 110, 1373. Slamnik, M., Talanta, 1974, 21, 960. Morales, A., Richter, P., and Toral, M. I., Analyst, 1987, 112, 965. Morales, A., Toral, M. I., and Richter, P.,Analyst, 1984, 109, 633. Paper A61430 Received November l l t h , 1986 Accepted December 23rd, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200971

出版商:RSC    

年代:1987

数据来源: RSC