摘要:

The AnalystThe Analytical Journal of The Royal Society of chemistryAdvisory Board"Chairman: J. D. R. Thomas (Cardiff, UK)"J. F. Alder (Manchester, UK)D. Betteridge (Sunbury-on-Thames, UK)E. Bishop (Exeter, UK)A. M. Bond (Australia)D. T. Burns (Belfast, UK)G. D. Christian (USA)"N. T. Crosby (Teddington, UK)*L. Ebdon (Plymouth, UK)"J. Egan (London, UK)L. de Galan (The Netherlands)A. G. Fogg (Loughborough, UK)*H. M. Frey (Reading, UK)"C. W. Fuller (Nottingham, UK)V. D. Goldberg (London, UK)T. P. Hadjiioannou (Greece)W. R. Heineman (USA)A. Hulanicki (Poland)I. Karube (Japan)"D. L. Miles (Wallingford, UK)"J. N. Miller (Loughborough, UK)E. J. Newman (Poole, UK)T. B. Pierce (Harwell, UKlE. Pungor (Hungary)J. RBiiCka (USA)"R. M.Smith (Loughborough, UK)W. I. Stephen (Birmingham, UK)M. Stoeppler (Federal Republic of Germany)"G. M. Telling (Bedford, UK)K. C. Thompson (Sheffield, UK)A. M. Ure (Aberdeen, UK)A. Walsh, K.B. (Australia)G. Werner (German Democratic Republic)T. S. West (Aberdeen, UK)J. D. Winefordner (USA)Yu. A. Zolotov (USSR)P. Zuman (USA)*Members of the Board serving on the Analytical Editorial BoardRegional Advisory EditorsFor advice and help to authors outside the UKProfessor Dr. sc. K. Dittrich, Analytisches Zentrum, Sektion Chemie, Karl-Marx-Universitat,Professor L. Gierst, Universite Libre de Bruxelles, Faculte des Sciences, Avenue F.-D.Dr. 0. Osibanjo, Department of Chemistry, University of Ibadan, Ibadan, NIGERIA.Dr. G. Rossi, Chemistry Division, Spectroscopy Sector, CEC Joint Research Centre,Dr.I. Rubeska, Geological Survey of Czechoslovakia, Malostranske 19, 118 21 Prague 1,Professor K. Saito, Coordination Chemistry Laboratories, Institute for Molecular Science,Professor M. Thompson, Department of Chemistry, University of Toronto, 80 St. GeorgeProfessor P. C. Uden, Department of Chemistry, University of Massachusetts, Amherst,Professor Dr. M. Valcarcel, Departamento de Quimica Analitica, Facultad de Ciencias,Professor Yu Ru-Qin, Department of Chemistry and Chemical Engineering, Hunan University,Talstr. 35, DDR-7010 Leipzig, GERMAN DEMOCRATIC REPUBLIC.Roosevelt 50, Bruxelles, BELGIUM.EURATOM, lspra Establishment, 21020 lspra (Varese), ITALY.CZECH 0s LOVA K I A.Myodaiji, Okazaki 444, JAPAN.Street, Toronto, Ontario M5S I A I , CANADA.MA 01003, USA.Universidad de Cordoba, 14005 Cordoba, SPAIN.Changsha, PEOPLES REPUBLIC OF CHINA.Editorial Manager, Analytical JournalsJudith EganEditor, The AnalystJanet DeanAssistant E ditcrsPaul Delaney, Mandy Mackenzie, Harpal MinhasEditorial Office: The Royal Society of Chemistry, Thomas Graham House, Science Park,Milton Road, Cambridge CB4 4WF.Telephone 0223 420066. Telex No. 818293 ROYAL.Fax 0223 423623.Advertisements: Advertisement Department, The Royal Society of Chemistry, BurlingtorHouse, Piccadilly, London, W I V OBN. Telephone 01-437 8656. Telex No. 268001.~The Analyst (ISSN 0003-2654) is published monthly by The Royal Society of ChemistryBurlington House, London W I V OBN, England.All orders accompanied with payment shoulcbe sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse RoadLetchworth, Herts. SG6 1 HN, England. 1989Annual subscription rate UKf200.00, Rest of Worlcf230.00, USA$460.00. Purchased with AnalyticalAbstractsUKf432.50, Rest of World f490.00USA $963.00. Purchased with Analytical Abstracts plus Analytical Proceedings UK €51 0.00Rest of World €580.00, USA$I 142.00. Purchased with AnalyticalProceedingsUKf254.00, Resof World €292.00, USA $584.00. Air freight and mailing in the USA by Publications ExpeditincInc., 200 Meacham Avenue, Elmont, NY 11003.USA Postmaster: Send address changes to: The Analyst, Publications Expediting Inc., 20(Meacham Avenue, Elmont, NY 11003.Second class postage paid at Jamaica, NY 11431. Alother despatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Posoutside Europe. PRINTED IN THE UK.Information for AuthorsFull details of how to submit material forpublication in The Analyst are given in theInstructions to Authors in the January issue.Separate copies are available on request.The Analyst publishes papers on all aspects ofthe theory and practice of analytical chemistry,fundamental and applied, inorganic andorganic, including chemical, physical, biochem-ical, clinical, pharmaceutical, biological, auto-matic and computer-based methods. Papers onnew approaches to existing methods, newtechniques and instrumentation, detectors andsensors, and new areas of application with dueattention to overcoming limitations and to un-derlying principles are all equally welcome.There is no page charge.The following types of papers will be con-sidered:Full papers, describing original work.Short papers: the criteria regarding origin-ality are the same as for full papers, but shortpapers generally report less extensive investi-gations or are of limited breadth of subjectmatterCommunications, which must be on anurgent matter and be of obvious scientificimportance.Rapidity of publication is enhancedif diagrams are omitted, but tables and formulaecan be included. Communications receive pri-ority and are usually published within 5-8weeks of receipt. They are intended for briefdescriptions of work that has progressed to astage a t which it is likely to be valuable toworkers faced with similar problems.A fullerpaper may be offered subsequently, if justifiedby later work.Reviews, which must be a critical evaluationof the existing state of knowledge on a par-ticular facet of analytical chemistry.Every paper (except Communications) will besubmitted to a t least two referees, by whoseadvice the Editorial Board of TheAnalystwill beguided as to its acceptance or rejection. Papersthat are accepted must not be published else-where except by permission. Submission of amanuscript will be regarded as an undertakingthat the same material is not being consideredfor publication by another journal.Regional Advisory Editors.For the benefit ofpotential contributors outside the United King-dom, a Panel of Regional Advisory Editorsexists. Requests for help or advice on anymatter related to the preparation of papers andtheir submission for publication in The Analystcan be sent to the nearest member of the Panel.Currently serving Regional Advisory Editors arelisted in each issue of The Analyst.Manuscripts (three copies typed in double spac-ing) should be addressed to:The Editor, The Analyst,Royal Society of Chemistry,Thomas Graham House,Science Park,Milton Road,CAMBRIDGE CB4 4WF, UKParticular attention should be paid to the use ofstandard methods of literature citation, includingthe journal abbreviations defined in ChemicalAbstracts Service Source Index. Wherever pos-sible, the nomenclature employed should fol-low IUPAC recommendations, and units andsymbols should be those associated with SI.All queries relating to the presentation andsubmission of papers, and any correspondenceregarding accepted papers and proofs, shouldbe directed to the Editor, The Analyst (addressas above). Members of the Analytical EditorialBoard (who may be contacted directly or via theEditorial Office) would welcome comments,suggestions and advice on general policy mat-ters concerning The Analyst.Fifty reprints of each published contribution aresupplied free of charge, and further copies canbe purchased.0 The Royal Society of Chemistry, 1989. Allrights reserved. No part of this publication maybe reproduced, stored in a retrieval system, ortransmitted in any form, or by any means,electronic, mechanical, photographic, record-ing, or otherwise, without the prior permission ofthe publishers

ISSN:0003-2654    

DOI:10.1039/AN98914FX005

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALAO 1 14(2) 125-252 (1 989) February 1989The AnalystThe Analytical Journal of The Royal Society of Chemistry12513313714314915516116516917318118519119520120721 121 722 122522723 123323724 124324525 1CONTENTSExtraction of Geological Materials With Mineral Acids for the Determination of Arsenic, Antimony, Bismuth andImportance of Calibration for Accurate Determination of Vanadium in Soil Samples-Bharti Patel, Koon Hung Chan,Determination of Iron Spevcies in Wine by Ion-exchange Chromatography - Flame Atomic Absorption Spectrometry-Direct Determination of Iron in Urine and Serum Using Graphite Furnace Atomic Absorption Spectrometry-LianFourier Analysis Method for Temperature Compensation of a Microcomputer-controlled Piezoelectric Crystal SulphurPhotolytic Interface for High-performance Chromatography - Chemiluminescence Detection of Non-volatile N-NitrosoStability of Bis(2,4,6-trichlorophenyl) Oxalate in High-performance Liquid Chromatography for ChemiluminescenceIon Chromatographic Determination of Aluminium With Ultraviolet Spectrophotometric Detection-John R.DeanLiquid Chromatographic and Fluorescent Derivative Aerobic Degradation Studies of Dehydroascorbic Acid in AqueousSolution at Elevated Temperatures-David Emlyn Hughes, Sylvia Van DeusenIn Situ Modified Electrodes in Stripping Voltammetry-Khjena Z. Brainina, Albina V. Tchernyshova, Natalya Yu.Stozhko, Lubov N. KalnyshevskayaInvestigation of the Mechanism of the Electrochemical Oxidation of Bamipine Hydrochloride by Voltammetry-I nciBiryol, Melike Kabasakaloglu, Zuhre SenturkSilver Bromide Based Chalcogenide Glassy - Crystalline Ion-selective Electrodes-Yuri G.Vlasov, Leonid N. Moskvin,Evgeni A. Bychkov, Dmitri V. GolikovDetermination of Thiosulphate at Trace Levels Using an Iodide Ion-selective Electrode-Tomozo Koh, KatsunobuKitamiPolycyclic Aromatic Hydrocarbon Solute Probes. Part II. Effect of Solvent Polarity on the Fluorescence Emission FineStructures of Coronene Derivatives-Riaz Waris, Michael A. Rembert, David M. Sellers, William E. Acree Jr.,Kenneth W. Street Jr., John C. FetzerSolvent Extraction of Titanium(lV), Zirconium(1V) and Hafnium(lV) Salicylates Using Liquid Ion Exchangers-N. M.Sundaramurthi, V. M. ShindeDetermination of Microgram Amounts of Platinum as Dithizonate in the Presence of Palladium by Second-derivativeSpectrophotometry-Sta nislaw KuS, Zyg rn u nt Ma rczen koStopped-flow - Photometric and Kinetic - Fluorimetric Methods for the Determination of Thyroid Hormones inTa blets-Ma ri na Toleda no, Ma.Carmen G ut ierrez, Ag ust i na Gomez-Hens, Dolores Perez-Bend itoSpectrophotometric Method for the Determination of Sorbic Acid in Various Food Samples With Iron(ll1) and2-Thiobarbituric Acid as Reagents-Oi-Wah Lau, Shiu-Fai Luk, Richard K. M. LamTitrations in Non-aqueous Media. Part XIV. Redoximetric Titrations of Aldehydes With Manganese(ll1) Acetate inGlacial Acetic Acid-Turgut Gunduz, Esma KIII~, S. Gul Ozta9SHORT PAPERSTitrations in Non-aqueous Media.Part XV. Redoximetric Titrations of Anthracene and Anthracene Derivatives WithTitrations in Non-aqueous Media. Part XVI. Redoximetric Titrations of Hydrazine and Some Hydrazine Derivatives withFourier Transform Infrared Microscopy for the Determination of the Composition of Copolymer Fibres: AcrylicIon-selective Electrodes in Organic Analysis-Determination of Alcohols via In Situ Generation of Xanthates UnderSpectrophotometric Determination of Certain Cephalosporins Using Molybdophosphoric Acid. Part II. DeterminationDetermination of Chlorpheniramine Maleate in Tablets by Second-derivative Absorption Spectrophotometry-Chu ng-Spectrophotometric Determination of Periodate With Salicylaldehyde Guanylhydrarone. Indirect Determination ofBOOK REVIEWSCUMULATIVE AUTHOR INDEXS e I en i u m by H y d ride G e n e r a t i o n At o m i c Absorption S p ect r o m e t r y- A r n o I d K u I d v e reStephen J.Haswell, Roman GrzeskowiakRadmila Ajlec, Janez StuparLiang, Patrick C. D'Haese, Ludwig V. Lamberts, Marc E. De BroeDioxide Sensor-R. D. Snook, P. E. ZaftCompounds-James J. Conboy, Jopseph H. HotchkissDetection-Noriko Imaizumi, Kazuichi Hayakawa, Motoichi Miyazaki, Kazuhiro lmaiManganese(ll1) Acetate in Glacial Acetic Acid-Turgut Gunduz, Esma Kilrc, S. Gul OztaSManganese(ll1) Acetate in Glacial Acetic Acid-Turgut Gunduz, Esma Kiliq, Adnan Kenar, S. Gul OztaSFibres-Girish C. PandeyPhase-transfer Catalysis-Wing Hong Chan, Albert Wai Ming Lee, King Sum Lam, Chi Lam Tseof Cefadroxil, Cefapirin, Ceforanide and Cefuroxime-Prodromos B. lssopoulosPui Leung, Che-Keung LawSome Organic Compounds Using the Malaprade Reaction-Juan Jose Berzas Nevado, Pablo Valiente GonzalezTypeset and printed by Black Bear Press Limited, Cambridge, Englan

ISSN:0003-2654    

DOI:10.1039/AN98914BX007

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 114 125 Extraction of Geological Materials With Mineral Acids for the Determination of Arsenic, Antimony, Bismuth and Selenium by Hydride Generation Atomic Absorption Spectrometry* Arnold Kuldvere Geological Survey of Norway, P. 0. Box 3006 Lade, N-7002 Trondheim, Norway Data are presented for the hydride generation atomic absorption spectrometric determination of arsenic, antimony, bismuth and selenium in stream sediments and a certified reference material (CPB-1 Lead Concentrate) using six different extraction media, viz., hydrochloric acid; nitric acid; 7 M nitric acid; nitric acid - sulphuric acid (9 + 1); aqua regia; and reversed aqua regia (HN03 + HCI, 3 + 1). Extraction with 7 M nitric acid was performed in test-tubes placed in a drilled out heating block (110 "C, 3.5 h).All other extractions were carried out overnight in 50-ml calibrated flasks placed in a water-bath. Bismuth was extracted quantitatively by all the extraction media. Arsenic and selenium were extracted completely by each of the media used, except for hydrochloric acid. Antimony was extracted completely by only two of the media, viz., aqua regia and reversed aqua regia. Selenium is converted to the +4 oxidation state on extraction with aqua regia and reversed aqua regia, i.e., selenium(V1) is reduced to selenium(lV), and the lower oxidation states are oxidised to selenium(1V). Hence the extracts can be analysed for selenium without pre-reduction to selenium(lV), which is necessary for hydride generation atomic absorption spectrometry. The nitric acid - hydrochloric acid mixtures used bring arsenic, antimony, bismuth and selenium quantitatively into solution and are, therefore, suitable for the sequential or simultaneous determination of these elements in a single sample solution.The speciation of selenium in stream sediments and surface waters is discussed. Keywords: Arsenic, antimony, bismuth and selenium determination; hydride generation atomic absorption spectrometry; decomposition of geological materials with mineral acids; selenium speciation in stream sediments Arsenic, antimony, bismuth and selenium are interesting elements in geochemical exploration and their determination in a single sample solution is desirable. In this respect antimony and selenium are the most important elements.Antimony and its compounds dissolve in concentrated nitric acid to form antimonic acid, which, however, is precipitated as the hydrated pentaoxide (Sb205.H20) even in concentrated acid.' Selenium must be in the +4 and antimony in the +3 oxidation state when they are to be determined by hydride generation techniques. Antiniony(V) is easily pre-reduced to antimony(II1) with potassium iodide. In contrast the pre-reduction of selen- ium(V1) to seleniuni(1V) is much more difficult and compli- cated. Heating with hydrochloric acid in closed systems (autoclaves)' or boiling of aliquots of the final sample solutions with dilute hydrochloric acid (4-6 M) has been recommended .3.3 The pre-reduction of seienium(V1) depends on the acid concentration and the reaction time and also on the sample volume and the nature of the sample.4 Losses of selenium have been reported using such pre-reductions.It is thought that these losses depend on the type of material from which the reduction vessels are made.2 Pre-reduction of selenium(V1) with hydrochloric acid is time consuming for routine analysis. Therefore, in the present work digestion procedures for geological materials have been developed which convert all the selenium in the sample [including selenium(VI)] to selenium(1V). thus making pre-reduction unnecessary. More than 30 years ago Schoeller and Powell5 reported that oxidation of selenium and native selenides by nitric acid or aqua regia proceeds only to the +4 state, giving selenious acid. It is shown here that aqua regia also acts as a pre-reducing agent for selenium(VI), reducing it to selenium(1V) only, and that reversed aqua regia (HN03 + HCl, 3 + 1) acts similarly.Procedures using mixtures of nitric and hydrochloric acids * Presented at Euroaiialysis VI. Paris, September 7-1 1, 1987 for the pre-reduction of selenium(V1) to selenium(1V) have found little application. The reason for this might be that chlorine, which is generated in this mixture, is known to be a strong oxidising agent and can oxidise selenium(1V) back to se 1 en i u ni ( VI ) according to the f o 11 ow i n g e q u a t i on 1 .h.7: H7Se03 + H20 + CI2=SeO42- + 3H+ + 2C1- . . (1) However. in the procedures used in this work the chlorine is driven off during the decomposition process, and, because o f the very high concentration of H+ and C1- ions, the equilibrium is forced completely to the left.This paper also discusses the speciation of selenium in stream sediments. In these waters, selenium occurs mostly in its highest oxidation state. For example, Sinemus et a1.2 found that all the selenium in the water of Lake Constance was present in the +6 oxidation state. Although selenium is normally in its highest oxidation state in surface waters, this is not necessarily so in groundwaters.8 Roden and Tallman9 analysed groundwater samples and found that the pre- dominant selenium species was selenium(V1) with negligible levels of selenium(1V). A comprehensive study of the redox processes that govern the ratio of selenium(1V) to selenium(V1) in waters does not appear to be available, although the following reaction is known to occur10: K = 1()fl S K = 10' ' H2Se03 + iOl = SeOJ'- + 2H+ .. (2) It can be seen that the oxidation is feasible if the reaction milieu is not too acidic. Normally. this is the case in water and, therefore, the above equilibrium explains the preponderance of selenium(V1) in surface waters. The speciation of selenium in stream sediments has received little attention and there is no explanation in the literature as to why selenium does not appear to exist in its +6 oxidation state in this material despite being submerged beneath selenium(V1) containing surface waters. This topic is dis- cussed in this paper and some light is thrown on the speciation of selenium in stream sediments.ANALYST, FEBRUARY 1989. VOL.114 126 Table 1. Instrument parametcrs Atomic nhsorplion spectrometer - Light sourcc . . . . . . . . Wavclengthlnm . . . . . . . . Slit (spectral slit widthinm) . . . . Recorderfull-scaleiA . . . . . . Recorder response (time constantis) Chartspecdhmmin-I . . . . RangeimV . . , . . , . . , . Recorder - MHS- 1 - Program rn e . . . . . . . . Reduction solution (NaBH, in 1.3% m/VNaOH)/ml . . . . Concentration of NaBH,, */o miV . . Pre-reduction solution (50% d V KI . . . . . . . . . . . . . . . . . . . . in 1 “/o m1V ascorbic acid solution jiml . . Sample dilution (150-ml Erlenmeyer flask, 4% V/V HCl)/ml . . . . . . Temperature o f silica celli”C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . As EDL 197.2 0.25 4 (0.7) 1 (‘/a) 20 10 Hyd I 2.5 5 - 2 0 900 Sb EDL 217.6 3 (0.2) 1 (Y,) 0.25 20 10 Hyd I 2.5 7 1 20 900 Bi EDL 223.0 3 (0.2) 1 (!A) 0.25 20 10 Hyd I 2.5 5 - 20 900 Se EDL 196.0 3 (0.2) 1 (‘4) 0.25 20 I0 Hyd I1 2.5 7 - 20 900 Experimental Apparatus A Perkin-Elmer Model 403 atomic absorption spectrometer, equipped with an MHS-1 mercury - hydride system and Model 056 recorder, was used.The light sources for arsenic, antimony. bismuth and selenium were electrodeless discharge lamps with an external power supply. Argon was used as the purge gas and 150-ml Erlenmeyer flasks with standard 29/32 taper joints were used as the reaction vessels in which evolution of the hydrides occurred. The operating parameters for the instrument and accessories are given in Table 1. Reagents and Solutions With the exception of sodium tetrahydroborate(III), all chemicals were of analytical-reagent grade.Sodium trtrahydrohorate(ZII). A 7% mlV solution was used for the antimony and selenium determination and a 5% m/V volution for the arsenic and bismuth determination. This solution was prepared from sodium tetrahydroborate(II1) powder (Merck-Schuchardt) by dissolving it in I .3% mil/ sodium hydroxide solution. The resulting cloudy liquid was vacuum filtered through a 0.2-pm porous membrane. Argon was bubbled through the solution for 5 min at least once every working day to free it from small bubbles of hydrogen. The solution was stored in a refrigerator when not in use. One preparation normally lasted for 3 d without deterioration. Potassium iodide solution, 50% mim. A 50-g mass of potassium iodide (Merck) was dissolved in 49 ml of water in which 1 g of ascorbic acid had been dissolved previously.Hydrochloric acid, 4% VIV. An 80-ml volume of the concentrated acid (Merck) (sp. gr. 1.19) was diluted to 2 1 with water. Working standards of arsenic( V), antimony(IZI), his- muth(ZlZ) and srfenium(ZV), 0.1 mg I-’. These solutions were prepared by serial dilution of the corresponding stock solutions (Titrisol, Merck) and contained 10% VIV reversed aqua regia (HNO? + HCl, 3 + 1). The preparation of a single multi-element working standard solution of the above elements is not recommended because of the well known mutual interference of volatile hydride- forming elements. 11,1? Selenium(VZ) stock solution, 1 g 1-1. A 1.841-g mass of selenic acid (BDI-I) was dissolved in 1 1 of 10% VlVnitric acid.Appropriate working standards were prepared from this solution. maintaining the concentration of nitric acid. Potassium permanganate solution, 5% mlm. A 5-g mass of potassium permanganate (Merck) was dissolved in 95 ml of water. Hydrochloric acid (sp. gr. 1.19), nitric acid (sp. gr. 1.40) and sulphuric acid (sp. gr. 1.84) (all from Merck) were used for digestion of the samples. Procedures All the microvolumes (1 ml or less) specified in the procedures below were added with micropipettes, which had an accuracy of at least t3%. “Rapid method” of decomposition using 7 M nitric acid A 1-g mass of the <180 pm fraction was heated with 5 ml of 7 M HNO? at 110 “C for 3.5 h in open test-tubes (30 ml). The mixture was then diluted to 20.3 ml (the sample residue occupies a volume of 0.3 ml) and the solution was filtered through a 0.02-mm nylon cloth (Seidengazefabrik, Thal, Switzerland) or, alternatively, centrifuged and decanted.This procedure has been used for many years at the Geological Survey of Norway (NGU) for the extraction of heavy metals from stream and river sediments for their simultaneous determination by inductively coupled plasma atomic emission spectrometry (TCP-AES). In the present work, data are presented which demonstrate that this pro- cedure is also suitable for the extraction of arsenic, bismuth and selenium from stream sediments for their subsequent determination by hydride generation atomic absorption spectrometry (AAS). “Overnight method” of decomposition using specified acids or mixtures thereof To a mass of 1 g or less of the S180 vm fraction in a 50-ml calibrated flask were added 5 ml of freshly prepared aqua regia (or, alternatively, some of the other specified mixtures or acids) and the mixture was allowed to stand overnight in a water-bath (95-100 “C).The flasks were then cooled to room temperature, the solution was made up to the mark with water, mixed and filtered through S & S folded filters (No. 595i) (Schleicher and Schuell, FRG) into 50-ml polyethylene flasks fitted with screw-caps. Blanks were run throughout the procedures. No significant problems with the blank were encountered.ANALYST, FEBRUARY 1989, VOL. 114 127 General procedure for the determination of arsenic, antimony, bismuth and selenium An appropriate volume of the final sample solution (0.05-5 ml) was transferred into an Erlenmeyer flask (the reaction vessel).A 20-ml aliquot of 4% V/V hydrochloric acid was then added and the flask was purged of air using argon from a hose. The system was closed and the programme started (Table I ) . Peak-height readings were taken and compared with those given by the standard working solutions of the corresponding element. Background correction was applied, but was found to be unnecessary for all four elements studied. Arsenic was determined as arsenic(V). In instances where it was not known with certainty that arsenic was in the +5 oxidation state, 1 drop of potassium permanganate solution was added to the measuring solution before the determination of arsenic.Potassium permanganate oxidises arsenic(II1) to arsenic(V) instantaneously. The determination of arsenic with the MI-IS-1 system has been described elsewhere. Unfortu- nately, in that paper13 there is a typographical error in Fig. 2; “!rg As” should read “ng As.” Antimony(V) and selenium(V1) must be pre-reduced as described below. Pre-reduction of selenium By using the “overnight method” with aqua regia or reversed aqua regia (HN03 + HC1, 3 + 1) digestion, all the selenium is converted to the +4 oxidation state. Hence the final solutions obtained with these two digestion procedures could be used directly for the determination of total selenium. If the oxidation state of selenium is not known prior to the determination. then reduction of the sample is necessary for the determination of total selenium, because selenium(V1) gives almost no signal with hydride generation techniques.11 Various treatments with hydrochloric acid have been pro- posed to convert all the selenium in a sample into selen- ium(IV).’-4 In the present work, however, the following treatment with aqua regia was found to be satisfactory. Treatment with aqua regia To 1 ml or less of the final sample solution in an Erlenmeyer flask (the reaction vessel) were added 2 ml of aqua regia. The mixture was swirled gently and allowed to stand overnight at room temperature after which the flask was transferred into a water-bath for 10 min. Then, at room temperature, 20 ml of 3% V/V hydrochloric acid were added to the flask and the total selenium was determined as described under General proce- dure f o r the determinution of arsenic, antimony, bismuth and selenium.Blanks and standards of selenium(1V) were run similarly. However, no significant differences in the results were found even if the standards were not subjected to the 0.200 1 0.150 a, V c 4 0.100 8 II Q 0.050 10 ng - SeV1 10 ng Selv h 25 ng SeVl 25 ng SeiV 50 ng SeV1 50 ng SeiV :i I Time -.+ O L Fig. 1. reduction of selenium. For details see text Recorder traces for SeIV and SeV1 standards after pre- overnight treatment. By using this method for the reduction of selenium(V1) to selenium(IV), the surface effects of the glass, described by Sinemus et al. ,2 were eliminated (cf. Fig. 1). Pre-reduction of antimony To an appropriate volume of the final sample solution in an Erlenmeyer flask (the reaction vessel) was added 1 ml of the potassium iodide solution and the determination was per- formed immediately as described under Generalprocedure for the determination of arsenic, antmony, bismuth and selenium.Blanks and standards were run similarly. It is important that the determination be carried out immediately after the addition of potassium iodide, otherwise the formation of iodine in the acidic solution may cause interference in the absorption measurements. Results and Discussion Sample Digestion Hydrochloric acid digestion It is well known that arsenic, antimony and selenium are lost from open vessels during reaction with hydrochloric acid at elevated temperatures. 14 If the reaction with hydrochloric acid is carried out in sealed tubes,ls decomposition bombs or autoclaves, then losses of these elements by evaporation can be avoided.1.16 However, arsenic, antimony, bismuth and selenium can occur in some geological materials, at least in part, as free elements and, consequently, the treatment must then be oxidative.In addition, minerals that are not soluble in hydrochloric acid (e.g., pyrites) must be subjected to oxidative digestion. Therefore, it appears that treatment with hydro- chloric acid alone is not sufficient to extract the four elements quantitatively into a single sample solution. This is confirmed by the results presented in Tables 2 and 3 (see also Table 4). which show that hydrochloric acid extractions give low values for arsenic, antimony and selenium.On the other hand, bismuth is extracted efficiently from stream sediments with hydrochloric acid. There are two reasons for this: firstly, bismuth chloride is not volatile and secondly. bismuth (and also arsenic and antimony) in stream sediments is mainly bound to precipitated iron(II1) oxides for which hydrochloric acid is an excellent solvent. Nitric acid digestion Under normal conditions nitric acid acts as an oxidising agent towards all substances with oxidation potentials more negative than +0.9S Vl7: . . (3) Hence the metallic forms of the elements studied in this work will be decomposed by this acid. Arsenic and antimony will be oxidised to the +S oxidation state, whereas bismuth will be oxidised to the +3 and selenium to the +4 oxidation state. Selenium(V1) will be dissolved as SeVI.However, concentrated nitric acid alone is a poor solvent for sulphides such as pyrites and copper sulphide ores.1 Antimony and its compounds dissolve in concentrated nitric acid to give antimonic acid but this is subsequently pre- cipitated as the sparingly soluble hydrated pentaoxide (Sb20s.I-120).’ The result of an analysis can, therefore, depend on the mass of sample taken for digestion. For example, using a sample mass of about 10 mg, the recovery of antimony from the certified reference material CPB-I Lead Concentrate (obtained from Canadian Certified Reference Materials Project) was satisfactory. I-Iowever, the recovery decreased with increasing sample mass, being only 50”/u if a sample mass of about 20 mg was used (Table 2, concentrated HN03 digestion).This is due to the low solubility of hydrated antimonic acid. NO3- + 4H+ + 3e- S NO + 2H20128 ANALYST. FEBRUARY 1989. VOL,. 114 tj L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . w-- . . . . . . . . . . G 3 3 % a + . . . . . . . . . ' - + c . . . . . . . . . . . . . . : : : 2 2ANALYST, FEBRUARY 1989. VOL. 114 129 Table 3. Determination of arsenic, antimony, bismuth and selenium in certified reference material CPB- 1 Lead Concentrate using different extraction procedures. Certified values: As, 0.056 t 0.003"iO; Sb. 0.36 ?r 0.03%; Bi, 0.023 k 0.002%; and Se. 31 k 3 pg g-l (see also Table 4) Extraction procedure (3 + 1 ) HCI + HNOi Mass of CPB- 1 * per 50 ml of extract/ 1ng 10.7 10.7 20.3 20.3 HNO; + HCI (3 + 1) .. 11.0 11.0 22.1 22.1 HNO; + H2S04 ( 9 + 1) . . HCI (concentrated) HNO, ( con ce n t r a t e d ) H N 0 7 ( 1 + 1 ) HNO? (1 + 1 ) ; rapid method 10.5 10.5 22.4 22.1 13.2 13.2 20.6 20.6 12.5 12.5 20.1 20.1 10.5 10.5 20.6 20.6 12.0 12.0 26.2 26.2 As Sb Bi Se Extract taken per determina- t ion/$ 100 250 S0 100 Mean: Relative error: 100 250 SO 100 Mean: Relative error: I 0 0 250 5 0 100 Mean: Relative error: 100 250 5 0 100 Mean: Relative error: I00 250 50 100 Mean: Relative error : 100 250 50 100 Mean: Re lativc error: 100 250 50 too Mean: Re1 a t i ve error: Found. (Yo 0.058 0.057 0.057 0.058 0.05 8 +3.5 0.057 0.057 0.059 0.056 0.057 + 1.80 0.057 0.059 0.058 0,060 0.059 +5.36 0.017 0.018 0.017 0.019 0.018 -67.9 0.058 0.057 0.058 0.056 0.057 + 1.79 0.058 0.056 0.059 0.059 0.058 +3.57 0.059 0.060 0.057 0.052 0.057 + 1.79 Extract taken per de t e rm i n a- tioniul 10 25 5 I 0 10 25 5 10 10 25 5 1 0 I 0 25 5 10 10 25 5 1 0 10 25 5 10 10 5 5 Found, Yo 0.35 0.37 0.37 0.36 0.36 * 0.00 0.37 0.37 0.37 0.35 0.37 +2.78 0.36 0.37 0.35 0.34 0.36 t 0.00 0.33 0.3 1 0.31 0.33 0.32 -11.11 0.34 0.33 0.19 0.19 0.26 -27.8 0.35 0.36 0.37 0.34 0.36 kO.00 0.35 0.35 0.35 2.5 0.32 0.34 -5.56 Extract taken per determina- tionipl I 0 0 200 50 100 100 200 50 100 100 200 50 100 100 200 50 100 100 200 5 0 100 100 200 50 100 I00 200 50 100 Found, % 0.023 0.024 0.024 0.025 0.024 4.35 0.022 0.024 0.022 0.025 0.023 k 0.00 0,020 0.021 0.019 0.016 0.019 - 17.4 0.(12 1 0.025 0.024 0.023 0.023 t 0.00 0.022 0.024 0.024 0.O24 0.024 +4.35 0.023 0.025 0.024 0.O24 0.024 k4.35 0.024 0.025 0.024 0.02s 0.025 +%.70 Extract taken per determina- tionipl 500 200 500t - 500 200 500t - 500 200 5001- - 500 200 500t - 500 200 500t - 500 200 500t - 250 100 - 250t Found1 pgg ' 30.1 30.4 32.0 30.8 -0.65 - 32.0 30.0 31.3 31.1 t 0 .3 2 - 30.0 30.6 33.1 31.2 +0.65 - 14.7 13.1 14.7 14.2 - -54.2 30.8 30.7 30.9 30.8 -0.65 30.5 31.6 32.1 31.4 + 1.29 - - 29.2 28.8 - 30.5 29.5 -4.84 Lead Concentrate. Canadian Certified Reference Materials Project, Canada Centre for Mineral and Energy Technology. Ottawa. Canada. L- The malyre addition technique was used.130 ANALYST, FEBRUARY 1989, VOL. 114 Digestion with 7 M nitric acid Virtually the same results were obtained with this medium as were obtained for arsenic, bismuth and selenium with concentrated nitric acid.With antimony, the hydrated trioxide (Sb203.H20) is precipitated in dilute nitric acid and is more soluble than the hydrated pentaoxide. 1 Good recovery of antimony was, therefore obtained from the certified reference material CPB-1 Lead Concentrate, partly because of the higher solubility of the trioxide and partly because of the cmall sample masses used (Table 3). However, the results for antimony in stream sediment samples (Table 2) were not satisfactory, even though the values were slightly higher than those obtained with concentrated nitric acid. The large sample mass (1 g) required for the determination resulted in a fairly concentrated final solution which decreased the solubility of the hydrated trioxide. Nitric acid - sulphuric acid digestion Bismuth Lvas not extracted quantitatively from the certified reference material CPB-1 Lead Concentrate using this medium (Table 3).because of coprecipitation with lead sulphate. Antimony was extracted quantitatively from the reference material (provided that small sample masses were used), but not from stream sediments, because large sample masm were used in thi\ instance. Aquii regia and reversed aqua regia digestion These two extraction media bring all four elements (arsenic, antimony, bismuth and selenium) quantitatively into solution (Table 3). The values obtained for antimony in stream sediments using aqua regia digestion showed good agreement with those obtained by reversed aqua regia digestion (Table 2). Further, the recovery of total selenium from samples spiked with selenium(V1) (Table 5 ) was good, regardless of Table 4.Approximate minerdlogic,il compovtion. certified balues m d confidence interv,il\ f o r certified relerence material ('PB- 1 Lead Con ce n t r ;i t t' Amount . ('% 111 I n 1 Mi n rs ral Galena . . . . 72.5 Pyrrhotite . . . . 12 Sphalerite . . . . 7 Pyrite . . . . . . 3 Ironoxides . . . . 1 Aluminosilicates . . 1 Carbonates . . . . 1 Chalcopyrite . . 0.5 Boulangerite . . 0.5 A 05"o confidence level. Cons t i t uc 11 t Ph S FC Zn SiO, Sb AI2O CU As Bi Cd Se Ag Hs Ccrthed \ alue, ('/<> m vz 64.74 T 0.12 17.8 ? 0.2 8.43 * 0.06 4 42 2 0.04 0 74 2 0.04 0.36 k 0.03 0.28 * 0.02 0.254 L 0.004 0.056 t 0.003 0.023 k 0.002 0.0143 i 0.0005 2 2 6 . 0 + 6 ~ g g - ~ 31.0 k 3ugg-l 5.5 k 0.5 pgg-' whether the digestions had been carried out with aqua regia or reversed aqua regia.This demonstrates that these two nitric - hydrochloric acid mixtures convert all the selenium in the sample to the +4 oxidation state provided that the suggested digestion procedures are followed. Speciation of Selenium in Stream Sediments Fig. 2 shows that the extractions with aqua regia and reversed aqua regia give higher values for the selenium concentration in stream sediments than extraction with nitric acid alone. The regression equations are: y = 1.037~ + 5.122, where y corresponds to the aqua regia digestion and x to the nitric acid digestion; and yl = 1.006~ + 18.033, where y , corresponds to the reversed aqua regia digestion and x to the same nitric acid digestion.However, the nitric acid extracts could contain some selenium(VI), which cannot be measured by hydride genera- tion AAS. Therefore, the nitric acid extracts were also analysed after pre-reduction with aqua regia as described under Treutrnent with aqua regia and the results are shown in Fig. 3. In this instance the regression equation isy = 0.960~ + 4.768, where y corresponds to the total selenium concentra- tion (after pre-reduction) and x to the concentration of selenium( IV) (before pre-reduction). Because of the almost identical results obtained both before and after pre-reduction, the presence of selenium(V1) in these stream sediment samples is very unlikely or, if present. the concentration is negligible. The higher values for the selenium concentration obtained with the hydrochloric acid - nitric acid mixtures compared with the values obtained for extraction with nitric Y V J 0 50 100 150 200 250 300 [Se] after H N 0 3 digestioning g-' Fig.2. Comparison of selenium concentrations (ng g 1 ) found i n the stream sediment samples (see Table 2) after digestion with HN03 (x) and aqua regia ( J ) . and also reversed aqua regia ( y ) . 0, Aqua regia digestion; 0. reversed aqua regia digestion; and 0. results of aqua regia and reversed aqua regia digestion coincide. j' = .Y is the line ol equality Table 5 . Results for the determination of total selenium in spiked CPB-1 (certified value, 31 + 3 pg g-I). Spiking was carried out using a relenium(V1) standard solution. The results show that selenium is converted to the +4 oxidation state when aqua regia or reversed aqua regia (HN07 + HCI, 3 + 1) is used as the decomposition medium Total selenium Decomposition CPB-1 Selenium(V1) Mean medium takenimg addeding Calculatedhg Founding recovery, % Aquaregia .. . . 20.8 2000 2645 2625 Aquaregia . . . . 15.5 2000 248 1 2315 Aquaregia . . . . 11.2 2000 2347 2200 95.4 Reversed aqua regia 19.6 2000 2607 2500 Reversed aqua regia 20.9 2000 2648 2550 97.9 Reversed aqua regia 15.0 2000 2465 2500ANALYST. FEBRUARY 1989. VOL. 114 131 300 250 - ul C a, 200 . - c” - 150 m 0 I- +- 100 Fig. 3. Comparison of concentrations (ng g I ) of Se’b’ (x) and total selenium ( y ) in the stream sediment samples digested with nitric acid. The nitric acid digestion procedure gives only SeIV in the final solution.After pre-reduction of the final solution, the total selenium is determined. y = x is the line of equality acid alone (Fig. 2) can be attributed to the higher extracta- bility of selenium by these mixtures. It is known that selenium principally follows sulphur in geological processes, replacing it diadochically in sulphidesis from ivhich it is released during the weathering of bedrocks. Sulphur is then oxidised mainly to sulphate whereas selenium is oxidised mainly to native selenium or selenite.ls.Ic1 Hence in soils and sediments selenium is present in the insoluble elemental form and/or as selenite, predominantly as hydrated iron selenite. When the soils and sediments come into contact u.ith surface waters, only the soluble selenites are oxidised to the highest oxidation state of selenium according to equation (2). However.the oxidation products are carried away by water streams, which explains the absence of selenium(V1) in stream sediments and its preponderance in surface waters. In addition, most selenates are soluble, the only exception being lead selenate which is characterised by its insolubility.20 For the determination of selenium in the certified reference niaterial CPB-1 Lead Concentrate, all the extraction media, with the exception of concentrated hydrochloric acid, gave similar values for selenium without pre-reduction. This indicates that selenium(V1) is not present in this certified reference material and that the minerals which contribute to the selenium content are readily soluble in the extraction media in question.Conclusions The nitric acid - hydrochloric acid mixtures used in this work bring arsenic, antimony, bismuth and selenium quantitatively into solution and are, therefore, suitable for the determination of these elements in a single sample solution. Selenium is converted to the +4 oxidation state on extraction with aqua regia and reversed aqua regia, i.e., selenium(V1) is reduced to selenium(1V) and the lower oxidation states are oxidised to selenium(1V). Hence the extracts can be analysed for selenium without pre-reduction to selenium(IV), which is necessary for hydride generation atomic absorption spectrometry. The content of selenium(V1) in stream sediments appears to be negligible. The author thanks Egil Kvam for valuable analytical help. References 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Bock, R., “A Handbook of Decomposition Methods in Analytical Chemistry,” Blackie, London, 1979, p. 197. Sinemus, H . W., Melcher, M., and Welz, B., At. Spectrosc., 1981, 2, 81. Reichert. I. K., and Gruber. H., Vom Wusser, 1978. 51. 191. Gutter, G. A . , Anal. Chim. Acta, 1978, 98, 59. Schoeller, W. R., and Powell, A. R., “The Analysis of Minerals and Ores of the Rare Elements,” Third Edition, Griffin, London, 1955, p. 231. Lund, W., and Bye, R . , Anal. Chim. Acta, 1979, 110, 279. Bye, R . , Talanta, 1983. 30, 993. Hoover, T. B., and Yager, G. D., Anal. Chern., 1984,56,221. Roden, D. R., and Tallman. D. E., Anal. Chem., 1982, 54, 307. Bye. R . , Talanta, 1986, 33, 705. Kuldvere, A . , Analyst, 1988, 113, 277. Welz, B., and Melcher, M., Anal. Chim. Acta, 1981, 131, 17. Kuldvere, A , , At. Spectrosc., 1980, 1, 138. Hillebrand, W. F., and Lundell, G. E., “Applied Inorganic Analysis,” Second Edition, Wiley, New York and London, 1953, pp. 259, 273 and 328. Pahlavanpour, B., Thompson, M.. and Thorne, L., Analyst, 1980, 105, 756. Dolezal, J . , Povondra, P., and Sulcek, Z., “Decomposition Techniques in Inorganic Analysis.” Iliffe Books. London, 1968. Holleman, A. F., and Wiberg, E.. “Anorganische Chemie,” Walter de Gruyter, Berlin, 1945, p. 269. Koljonen. T.. Ann. Agric. Fenn., 1975, 14, 240. Koljonen. T.. Ambio, 1978, 7, 169. Weast, R. C., Editor-in-Chief, “CRC Handbook of Chemistry and Physics,” 68th Edition, CRC Press, Boca Raton, FL, 1987. Paper 8101 729H Received May 3rd, 1988 Accepted September 23rd, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400125

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989. VOL. 114 133 Importance of Calibration for Accurate Determination of Vanadium in Soil Samples Bharti Patel, Koon Hung Chan, Stephen J. Haswell and Roman Grzeskowiak School of Chemistry, Thames Polytechnic, Woolwich, London SE 18 6PF, UK The increased combustion of oil has led in recent times to an increase in the emission of vanadium into the environment. Vanadium is an essential trace element for normal cell growth at the p.p.b. level, but has a toxic effect when present at higher concentrations. Hence the possible risks associated with vanadium exposure need to be assessed by careful monitoring of this potentially toxic element. It has been shown that the analyte matrix can have a profound influence on the results obtained for the determination of vanadium in soils by flame atomic absorption spectrometry.Certified soil was used to establish the percentage recoveries and optimum calibration graphs were employed for a reliable determination of vanadium in soil samples. Keywords: Vanadium species; calibration; flame atomic absorption spectrometry; soil samples Although it is known that vanadium is an essential trace element, possessing specific physiological functions,’J interest in this element has been directed mainly towards its toxic effects. Occupational exposure to vanadium has been observed in several industrial processes, for example, in iron and steel production, in the manufacture of pigments, printing inks and paints. in the glass industry and in the cleaning and repairing of oil-fired boilers, particularly in electricity power stations.”-” The discovery of high concentrations of vanadium in the fly ash from the combustion of petroleum products and coal has increased further the interest in the toxicity of vanadium compounds, because environmental background levels of vanadium have been slowly rising as a result of the combustion of fossil fuels.” Vanadium is therefore important not only because of its toxicity at high levels, but also because it is an environmental pollutant .7 -9 The most important factors limiting the investigations in occupationally and non-occupationally exposed subjects are undoubtedly the analytical difficulties involved in determining vanadium.5 Land contaminated by past industrial processes is becoming strategically important for development as building or rec- reation land as the availability of “green field” or undeveloped sites becomes scarce, particularly in the inner city areas.Provided that the hazards of using contaminated land are recognised, and the sites are investigated properly, derelict industrial areas can be brought successfully back into public use by taking appropriate precautionary measures. In this work a suitable methodology for the determination of vanadium in soil samples taken from an old gasworks site has been developed using flame atomic absorption spectrometry (AAS). The results demonstrate the importance of instrumen- tal calibration for the accurate determination of vanadium. Experimental Materials and Methods Soil sampling The site studied was an old gasworks in south east London and soil samples were taken from the surface layers (0-6 in deep) of soil using a stainless-steel spade.Preparation of soil samples The soil samples were air dried for 1 week, ground into a fine powder using a pestle and mortar and passed through a 150-pm brass sieve (Endicotts). The fraction thus obtained was mixed and stored in polyethene vials ready for treatment. Extraction of vanadium Approximately 1 g of air-dried soil was weighed accurately into a long Gerhardt digestion tube. All the samples were digested with a mixture of 10 ml of concentrated nitric acid and de-ionised water (1 + 1). Each tube was placed in a Gerhardt Kjeldatherm 40s aluminium digestion block. The temperature was held constant at 80°C for 1 h, then raised to 120°C for 1 h and finally held constant at 140°C for 2 h.When cool, the digest was filtered through acid-washed Whatman 541 filter- paper into a 25-ml calibrated flask and diluted to volume with de-ionised water. The solutions were then transferred into a plastic screw-top container for storage prior to analysis. All the samples, including the blanks, spiked acid blanks and certified reference material Soil SO-1 (containing 139 i 8 pg g-1 of vanadium) obtained from CANMET (Canada Centre for Mineral and Energy Technology), were digested using the above procedure. Vanadium was detected with a l’hermo Electron 357 atomic absorption spectrometer at a resonance wavelength of 318.5 nm. The flame AAS par- a I rants. ameters were optimised for the various c l’b Preparation of 1000 p.p.m.stock solutions A solution of vanadium in hydrochloric acid (V-IICl) was purchased from BDH (SpectrosoL, for atomic spectrometry); it is prepared commercially by dissolving ammonium meta- vanadate in 0.5 M hydrochloric acid (pH 0.66). A solution of vanadium in nitric acid (V-HN03) was purchased from Aldrich (vanadium atomic absorption stan- dard solution); it is prepared commercially by dissolving vanadium metal in 2% nitric acid (pI1 1.12). A solution of vanadium pentaoxide in sodium hydroxide (V205-NaOH) was prepared by dissolving 1.7852 g of vanadium pentaoxide in 0.04 M sodium hydroxide solution and diluting to volume with de-ionised water in a 1-1 calibrated flask (pH 10.61). A solution of vanadyl sulphate i n sulphuric acid (VOS04- H2S04) was prepared by dissolving 4.0682 g of vanadyl sulphate pentahydrate in sulphuric acid (pH 2.44).A duplicate set of the above stock solutions was also prepared containing 1000 p.p.m. of aluminium added in the form of aluminium nitrate nonahydrate (A1(N03)3.9H20]. The standards used for constructing the individual calibra- tion graphs were prepared from the appropriate stock so 1 u t io ns . Results and Discussion The results obtained for the recovery of spiked vanadium using the nitric acid digestion procedure are shown in Table 1. The spiked vanadium was in the form of vanadium pentaoxide134 ANALYST, FEBRUARY 1989, VOL. 114 Table 1. Percentage recovery of vanadium pentaoxide and vanadyl sulphate spiked into the nitric acid digest (in triplicate) using thrce different aqueous calibrations Recovery, Yo Cali brat io n graph Spiked Identical v o s 0 , - v , o i - vanadium sample V-HCI H,S04 NaOH ViOs .. A 96.6 108.7 121.8 B 98.3 112.5 124.3 C 99.0 109.7 124.3 VOSO, . . D 91.2 100.9 111.5 E 94.2 103.0 112.3 F 92.5 103.0 119.2 0.08 - 0 C m $ 0.06 - Ll a 0.04 - 0.02 - 0 2 4 6 8 10 12 14 Concentration, p.p.m. Fig. 1. Graphs of absorbance versus concentration for the aqueous calibration standards. (1) V-HCI; (2) V-HNO,; (3) V-HCI - VOS0,- H,S04: and (4) VOS04-H2S04 powder or vanadyl sulphate pentahydrate crystals. The percentage recoveries of vanadium were determined using three different aqueous calibration solutions: (i) V-1 ICl; (ii) VOS04-H2S0,; and (iii) V205-NaOH. The percentage recoveries for vanadium pentaoxide and vanadyl sulphate in nitric acid were 98 and 93%, respectively, when V-HCl was used as the calibrant.Using VOS04-H2S04 as the calibration standard the percentage recoveries were 110 and 102% for vanadium pentaoxide and vanadyl sulphate analytes, respectively. However, high percentage recoveries of 123% for vanadium pentaoxide and 114% for vanadyl sulphate were obtained when a solution of V205-NaOII was used to calibrate the instrument. The variation in results using the three different calibration solutions prepared similarly by diluting the appropriate stock solution to give the correct vanadium concentrations is thought to be due to the different oxidation states of the vanadium species formed during the nitric acid extraction procedure. It can be seen from the results in Table 1 that the best recoveries were obtained when the V-HCl calibrant was used for the determination of spiked vanadium pentaoxide and the VOS04-H2S04 calibrant was used for the determination of spiked vanadyl sulphate.It is also evident that the pH of the solution has an influence on the analyte response as shown by the high values obtained when spiked vanadium pentaoxide was determined against the V205-NaOH calibrant. This effect is best illustrated by plotting the absorbance (Fig. 1) and the pH of the calibration solution (Fig. 2) against the concentration. It is therefore concluded that it is not only the oxidation state but also the types of vanadium species present in solution that are important in the determination of this element. It has been 6 1 4 C I I I I I 0 2 6 10 14 Concentration, p.p.rn.Fig. 2. Graphs of pH versus concentration for the aqueous calibra- tion standards. (A) V-HCI; (B) V-HNO?; (C) V-HCI - VOS0,- H2S0,; and (D) VOS04-H2S04 Table 2. Effect of aluminium on the aqueous calibration graphs V-HCI - VOSOJ- VOS0,- V-HCI HZSO, HZSO, V-HNOI Calibration graphs in the absence of aluminium- Intercept Slope . . Correlation . . 6.86 x 10 -1 7.90 x lo-, 1.27 x 10-3 1.61 x lo-? . . 7.59 x 10 8.31 x 10-3 8.49 x 10-3 8.92 x 10-3 coefficient . . 0.9997 0.9992 0.9990 0.9997 Calibration gruphs in the presence of aluminium- Intercept Slope . . Correlation . . 4.80 x 10-4 6.69 x 10-4 1.30 x 10-1 7.20 x 10-4 . . 9.41 x 10-3 1.10 x 10-2 1.15 X lo-’ 1.17 X 10-2 coefficient . . 0.9999 0.9999 1.0000 0.9998 - ~ - - Table 3.Percentage recovery of vanadium from certified reference material Soil SO-1 using four different calibrations. The linear regression results are also given Recovery, % Calibration graph Ident- V-HCl- ical VOS0,- VOS0,- Sample sample V-HC1 H,SO, HZSO, V-HNO, - - - - Acid blank Certified reference material Soil SO-1 A 90.8 107.9 100.7 97.1 B 93.5 11 1.5 102.5 97.1 C 93.5 111.5 100.7 97.1 Linear regression purameters- Intercept Slope Correlation . . -2.54 x 10-4-3.45 x 10-3 5.01 x 10- 3.31 x 10-1 , . 8.34 x 10-3 7.70 x 10- 7 7.27 x 10-3 7.88 x lo-’ coefficient 0.9996 0.9948 0.9984 0.9987 suggested that the addition of matrix modifiers such as aluminium can enhance the analyte signals and partially reduce interference effects when matrix-matched standards are not employed.10 In the work described here the influence of the addition of 1000 p.p.m.of aluminium to the aqueous standards prepared as described above was studied. Both calibration procedures, viz., aqueous and aqueous aluminium modified, were used for the determination of vanadium in reference soil samples. The results given in Table 2 indicate that the vanadium response is enhanced typically by a factor of 1.3 for all the calibration graphs studied. However. the relative position of each specific graph remained unchanged. Hence the addition of aluminium produces no change in the observed relative positions of the calibration graphs (Table 2) associated with specific vanadium species. For the certified soil the best results were obtained when a mixture of non-aluminium modified V-HCI - VOS04-H2S04 was used as the calibrant (Table 3). This suggests that vanadium may be present in the soil in two different oxidation states.The resultsANALYST, FEBRUARY 1989, VOL. 114 7 100 I m m 5 ‘. 80 .- $ -0 m $ 6 0 - + 0 C .: 40 F 4- C a, z 2 0 - u 0 - 135 - - - Table 4. Analysis of soil samples Concentration of vanadium/pg g- 1 - Aqueous calibration - c Soil No. 1 2 3 4 5 6 7 8 9 1 0 11 12 13 14 15 V-HCl - V0SO;I- H W 4 106 79 100 105 77 42 13 58 40 58 78 52 91 83 47 Linear regression parameters- Intercept . . 2.55 x 10-3 Slope . . . . 6.76 x 10-3 Correlation coefficient . . 0.9986 V-HC1 94 70 85 86 67 36 13 51 35 51 70 51 84 77 44 3.42 x 10 4 8.19 X 10-3 0.9991 V-HNO3 106 79 95 104 75 42 13 58 40 58 79 52 91 83 46 4.49 x 10-4 7.23 x 10-3 0.9997 stream --+ I r l l 0 20 40 60 Miles N t Fig.3. Map of the gasworks site for the aluminium modified standards suggested that although they enhanced the instrumental sensitivity. they also lowered the expected recoveries from the reference soil by 24.8%. Thic effect is thought to be due to the relative change in the slope of the calibration graph and suggests that the addition of aluminium at the levels found in the soil samples (9.38%) is not to be recommended in this instance. Analysis of Soil Samples The results for the determination of vanadium in various soil samples using three different calibration graphs are shown in Table 4. The numbers assigned to the soil samples analysed refer to their position on the gasworks site as identified in Fig.3. Good agreement for the determination of vanadium was obtained when V-HCI - VOS04-H2S04 or V-HN03 was used as the calibrant; however, differences were observed for caibration using V-HC1. Investigation of the V-HN03 cali- I I 3 6 7 8 9 10 11 12 Sample No. Fig. 4. Histogram showing vanadium levels on the gasworks site brant by electron spin resonance and electronic spectroscopy showed that both vanadium(1V) and vanadium(V) were present in the solution. Therefore, as V-HN03 is available commercially, the preparation of V-HCI - VOS04-H2SOJ is not necessary. With regard to the vanadium distribution obtained for the various soil samples taken, using V-HCl - VOS04-H2S04 as the calibrant (Fig. 4), samples 1 4 had an average vanadium concentration of 98 vg g-1.These samples were taken from an old waste slag-heap on the site and were observed to be brown. Samples 6 1 0 were black. They were obtained further away from the slag-heap and generally had a lower vanadium content, viz., 50 pg g-1 or less. Not surprisingly it appears that the further from the slag heap the sample was taken the less was the contamination. However, it is clear that the topography of the site and two local ‘.hot spots” of higher contamination were identified by the results obtained for samples 13 and 14. At present there is no legislation pertaining to vanadium in soils. Nevertheless, a “threshold trigger concentration ,” the value of which depends on the intended use of the site, has been introduced to assist in determining the significance of the contamination of soils by vanadium.1 1 Conclusion This work has shown that the analyte response for vanadium in soil samples using flame AAS is dependent on the aqueous standard calibration graphs. Although the addition of alumin- ium as a matrix modifier enhanced the signal, it had little effect on the relative positions of the individual standard graphs for vanadium. Care must therefore be taken to ensure that both the oxidation states and the species (pH of solutions) of vanadium present in the samples to be analysed are as similar as possible to those present in the calibrants. Owing to the impracticability of analysing each sample for the total matrix composition it was found that the use of an appropriate reference material to identify the most suitable calibration graphs was sufficient as it reflected most closely the sample matrix.In practice, commercially available V-HN03 was found to be a suitable calibration standard for the determina- tion of vanadium in digested soil sample matrices using non-modified standards. If a suitable reference material is not available, the method of standard additions can be used as an alternative to compensate for matrix effects when using flame AAS. The distribution of vanadium found in the soil from the gasworks suggested that the areas of contamination were associated with a slag-heap and with local “hot spots.” Hence sampling from such a site should be undertaken with some care.136 ANALYST, FEBRUARY 1989, VOL. 114 1. 2. 3. 4. 5 . 6. 7. 8. References Macara, I. G . , Trends Biochem. Sci., 1980, 5, 92. Hopkins, L., and Mohr, H. E . , Fed. Proc. Fed. Am. Soc. Exp. Biol., 1974, 33, 1773. Kiviluoto, M., Pyy. L., and Pakarinen, A., Int. Arch. Occup. Environ. Health. 1981, 48, 251. Gylseth. B., Leira, H., Steinnes, E.. and Thomassen, Y., Scund. J. Work Environ. Health, 1979, 5 , 188. Sabbioni, E., and Maroni, M., “A Study on Vanadium in Workers from Oil Fired Power Plants,” Commission of the European Communities, Publication No. EUR. 9005, Luxem- bourg, 1983. Nelson, W. L., Oil Gas J., 1973, 54. Vouk, V. B., and Piver, W. T., Environ. Health Perspecl., 1983,47, 20. Henry, W., and Knapp. K. T., Environ. Sci. Technol.. 1980, 14, 450. 9. Ahlberg, M.. Berghem, L., Nordberg, G., Perssom, S . A . , Rudling, L., and Steen, B., Environ, Health Persyrct.. 1983. 47.85. Goecke, R . , Tufanru. 1968, 15, 871. “Site Tnvestigation and Material Problems,” (Greater London Council Guidelines f o r Contaminated Soils), Proceedings o f the Conference on Reclamation of Contaminated Land, Eastbourne, October 1970. 12. “Notes on the Rcdevelopmcnt of Gasworks Sites.” Inter- departmental Committee on the Redevelopment of Con- taminated Land, ICRL 18/79, Fifth Edition, April 1986. 10. 11. Puper 8103253J Received Augusr 9th, 1988 Accepted September 28th, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400133

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 114 137 Determination of Iron Species in Wine by Ion-exchange Chromatography - Flame Atomic Absorption Spectrometry Radmila Ajlec and Janez Stupar Joief Stefan institute, E. Kardelj University, Jamova 39, 61000 Ljubljana, Yugoslavia The direct coupling of ion-exchange chromatography to flame atomic absorption spectrometry (AAS) has been achieved by employing a Babington type nebuliser. The system enables all the processes on the column to be followed directly at flow-rates of between 1 and 5 ml min-I. The potential of the system was investigated for the determination of various iron species in synthetic samples containing iron(1l) and iron(ll1) in ionic or chelated form by employing various ion-exchange (Dowex 50-X8, Dowex 1 -X8) and sorptive (Amberlite XAD-2) resins, respectively.In some instances where direct coupling was impossible, owing to the physical properties of the effluent or eluent, conventional analyses of chromatographically separated iron species were performed by flame AAS. The optimum concentration range, limit of detection and reproducibility of measurement were also determined for a particular column capacity. When direct coupling was employed, the detection limit for the separated iron species was 15 ug with a relative standard deviation (RSD) of +3% and, using the conventional method of analysis, 2-5 pg with an RSD of 2 1%. On the basis of these results the system was applied to the determination of the ratio of iron(ll) to iron(ll1) in wines. Keywords: Iron speciation; ion-exchange chromatography - flame atomic absorption spectrometry; Babington nebuliser; wines The determination of metal species is a complex problem that is specific for a particular sample.A possible approach to surmount such difficulties efficiently and rapidly is to couple various chromatographic and optical spectrometry methods. The main obstacle to the efficient direct coupling of ion chromatography and/or liquid chromatography to flame atomic absorption spectrometry (AAS) lies in the difference between the optimum flow-rates through the column and the conventional pneumatic nebuliser. Several possibilities for overcoming this problem have been investigated. Direct coupling by modifying the flow-rates through the chromato- graphic column and the pneumatic nebuliser of a flame AAS system' makes it impossible to obtain the optimum conditions of measurement for either liquid chromatography or flame AAS.By employing optimum flow-rates for liquid chromato- graphy, the noise level becomes troublesome owing to the resulting low flow-rates through the pneumatic nebuliser used in flame AAS.2 The system employed for the atomic absorption injection method has also been used to connect a liquid chromatography column to flame AAS.3.4 In this instance drops of eluent fell into a funnel attached to the nebuliser capillary and were aspirated sequentially into thc flame. The peak height of the absorption signals corresponded to the concentration of the measured metal in a particular drop, and the sum of all the absorption signals to the total amount of a particular metal species.Some workers5.h employed a vented capillary tube to connect a high-perfor- mance liquid chromatography (HPLC) column to the pneu- matic nebuliser of a flame AAS system. Ebdon et al.7 constructed a special interface consisting of platinum wire spirals mounted on a rotating disc to couple HPLC to flame AAS. By rotation of the disc, the drops from the column were captured on the platinum wires, desolvated using a micro- Bunsen burner and introduced sequentially into the flame. The direct coupling of an ion-exchange chromatography column to the nebuliser of a flame AAS system has been described previously.%"' The flow of eluent through the column was controlled by the aspiration rate of the pneumatic nebuliser. A method of coupling gel chromatography directly to the pneumatic nebuliser of a flame AAS system has also been investigated." An additional water reservoir was con- nected to the tube used to couple the column and nebuliser in order to balance the difference in flow-rates. Although the noise level decreased the sample was diluted.The aim of this work was to develop a novel approach for the direct coupling of ion chromatography to flame AAS. A Babington nebuliser was employed and various ion-exchange (Dowex 50-X8, Dowex 1-XS) and/or sorptive (Amberlite XAD-2) resins were investigated for use in iron speciation. The potential of the proposed coupling technique was evaluated and the system was applied to iron speciation in wines. Experimental Apparatus Detection system For direct coupling of ion-exchange chromatography to flame AAS, an atomic absorption spectrometer was constructed, consisting of a Babington type nebuliser.12 a glass spray chamber, Varian hollow-cathode lamps, a 6-cm slot burner, a Carl Zeiss Jena SPM-2 grating monochromator, a Hamamatsu K-213 photomultiplier tube, a Varian AA-6 indicating module (Type IM-6), a Hewlett-Packard 17501 A recorder and an electronic integrator. 13 A Varian AA-5 atomic absorption spectrometer was employed for iron measurement when direct coupling was not possible.The instrument parameters of the two spectrometers are given in Table 1. Table 1. Instrument parameters used for measurement of iron absorbances Parameter Wavelengthlnm . . Spectral band passinm LampcurrentlmA . . Burnerslot/cm .. . . Flame . . . . . . Height above the burner topimm Nebuliser . . . . Flow-rate . . . . Background correction . . . . Spectrometer constructed for direct coupling 248.5 0.2 7 6 Air - acetylene 10 Babington type Variable (1-4.5mlmin-') Deuterium hollow- cathode lamp Varian AA-5 spectrometer 248.5 0.2 7 1 0 Air - acetylene 1 0 Pneumatic type Fixed ( 5 ml min 1 ) Deuterium hollow- cathode lamp138 ANALYST, FEBRUARY 1989, VOL. 114 Separution system The system was developed for the separation of iron species, employing ion-exchange [Dowex 50-X8 (5&100 mesh) and Dowex 1-X8 (50-100 mesh)] and/or sorptive [Amberlite XAD-2 (20-50 mesh)] resins. Glass columns (6 ml) (75 X 10 mm i.d.) were used for all measurements. This volume ensured that the capacity of the columns was not exceeded, with respect to the concentration of cations and anions in the wines.A peristaltic pump (Ismatec MS 4 Reglo), which allowed variable flow-rates of between 1 and 4.5 ml min-1 to be used, was connected to one end of the column. The other end was coupled directly to the Babington nebuliser. The system enabled the peristaltic pump to be joined either to the bottom or the top of the column. The direct coupling of the ion-exchange column to the Babington nebuliser is shown schematically in Fig. 1. Reagents Suprapur (Merck) acids and doubly distilled water were used for the preparation of sample and standard solutions. All other chemicals were of analytical-reagent grade. A standard iron(II1) stock solution (1000 p.p.m.) was prepared by dissolving 7.22 g of Fe(N0&.9H20 in 10 ml of concentrated nitric acid and diluting to 1000 ml with water.The solution was standardised by titration with EDTA. A standard iron(I1) stock solution (1000 p.p.m.) was prepared by dissolving 4.96 g of FeS04.7Hz0 in 5 ml of concentrated sulphuric acid and diluting to 1000 ml with water. The solution was standardised by titration with EDTA. A standard iron(I1) solution (100 p.p.m.) of iron - phenan- throline was prepared by dissolving 2.5 g of 1,lO-phenanthro- line monohydrochloride in water, adding 10 ml of iron(I1) solution (1000 p.p.m.) and diluting to 100 ml with water. Dowex 50-X8 (50-100 mesh) (hydrogen form) and Dowex 1-X8 (5(%100 mesh) (chloride form) ion-exchange resins and Amberlite XAD-2 (20-50 mesh) sorptive resin (Fluka) were used.To \ Th ree-way Drain \ stopcock Babington nebuliser Peristaltic Pump Fig. 1. nebuliser Direct coupling of an ion-exchange column to a Babington 2.25 M HCI 2.0 M HCI n 1.75 M HCI 0.3 1 . . O L I -Time Fig. 2. Elution curves for Fell': resin. Dowex 50-X8.5&10O mesh, in 0 . 1 M HNO,; column, 6 ml; and flow-rate, 4.5 ml min- l . Standard solution. Fe(NO,),.9H,O. 300 pg of Fe Determination of Iron Species In general iron can exist in two oxidation states, i.e., iron(I1) and iron(II1). The possible iron species in aqueous solution14 are hexaaquairon(I1) and hexaaquairon(II1) ions, various negatively and positively charged iron(I1) and iron(II1) complexes and uncharged organic compounds of iron(I1). The predominant iron species depends on the oxidation - reduction conditions, and on the concentration and stability constants of the anions present in the solution.Several methods have been employed for the determination of iron(I1) and iron(II1) in their mixtures, based on spectrophotometry,l5 ion-exchange colorimetry with spectrophotometric detection16 or flow injection analysis," but there are few reports describing the separation of iron(I1) and iron(II1) by ion-exchange chromat- ography. The cation-exchange18 and anion-exchange19 distri- bution coefficients of several cations including iron(T1) in different acidic media and selected data for ion-exchange separations20 have been reported. Iron adsorbed on organic matter in sea water has been concentrated on the sorptive resin Amberlite XAD-2.21 In addition, iron has been deter- mined by forming a l,l0-phenanthroline complex, which was then extracted from aqueous solutions on adsorbent Amber- lite XAD-2.22 This paper describes a system for the determination of iron species employing cation-exchange, anion-exchange and sorp- tive resins.Separation of Ionic Iron(I1) and Ionic Iron(II1) on Dowex 50-X8 Cation-exchange Resin The directly coupled system was used. The column was conditioned and purified by washing it with various concentra- 0.3 I I I -Time 0 ' Fig. 3. flow-rate 3.5 ml min 1 . Standard solution, Fe(N01)q.9H20. 300 vg of Fe Reproducibility of measurement for Fc"': resin. Dowcx 50-X8, 50-100 mesh, in 0.1 M HNO,; column. 6 ml; eluent. 2.25 M HCl; andANALYST. FEBRUARY 1989, VOL. 114 a, 0.2 e z 0 C m a o l - 139 D - 200 pg Fe 100 pg Fe 50 Fe 25 pg Fc 0.3 1 min - 300 pg Fe -Time 0 ' Fig.4. 4.5 ml min-I. Standard solution, Fe(N03),.9H20 Calibration graph for Fe11': resin. Dowex SO-X8, SO-100 mesh, in 0.1 M HNO,; column, 6 ml; eluent, 2.25 M HCl; and flow-rate, 0.4 I I 0 ' +Time Fig. 5 . Elution curves for: A, 300 pg of Ferr (as FeSO4.7H2O); B, 300 ug of Fe"' [as Fe(N0,),.9H20]; and C, 300 pg of Fell f Fell' (1 : 1). Resin, Dowex S0-X8, 5&100 mesh, in 0.1 M HN03; column, 6 ml: eluent, 2.25 M HCI: and flow-rate 4.5 ml min 1 tions (0-3 M) of HCl and doubly distilled water. On the basis of distribution coefficient data18 the resin was washed with 25 ml of 0.1 M HN03 prior to the analysis. Standard solutions of iron were also prepared in 0.1 M HN03. A &300-pg amount of iron was bound on the resin at a flow-rate of 1 ml min-1.Then, 5 ml of solvent were passed through the column at the same flow-rate followed by a further 15 ml of solvent at a flow-rate of 3.5 ml min-1 to separate the non-sorbed species. The optimum conditions for measurement were obtained when iron was eluted at a counter flow-rate of 4.5 ml min-I. The peak area was integrated with an electronic integrator until the signal returned to the base line. A reagent blank was integrated similarly. Sorption and elution of iron(IZ) and iron(III)from the resin Iron(II1) nitrate and iron(I1) sulphate standard solutions were used. The direct registration of atomic absorption signals showed that quantitative sorption of up to 300 pg of iron(TT1) and/or iron(I1) ions was obtained on the resin column in 0.1 M HNO? and in the presence of 0.1 M chloride ion, 0.1 M citrate ion or 0.01 M oxalate ion.The optimum conditions for elution were investigated by employing various concentrations of HCl. The results obtained for iron(II1) are shown in Fig. 2. Iron(I1) was eluted similarly and it was found that the optimum eluent for desorption of iron(I1) and iron(II1) from the resin was 2.25 M HCl. Reproducibility of measurement, calibration graph and detec- tion limit The reproducibility of measurement for an iron(II1) nitrate standard solution (300 pg) is shown in Fig. 3. It is evident that the proposed method has good reproducibility of measure- ment with a relative standard deviation (RSD) of k3%. The calibration graph for iron(II1) is shown in Fig. 4.A linear relationship between the peak area and the amount of iron(II1) in the standard solution was obtained in the range 0-300 pg of iron. The detection limit, calculated on the 30 basis, was 15 pg of iron(II1) ion. Sensitivity of measurement of iron(ZZI), iron(II) and their mixtures The sensitivity of measurement was compared for 300 pg of iron(II1) ion, 300 pg of iron(I1) ion and 300 pg of both iron(II1) and iron(I1) ions in a mixture (1 + 1). all in 0.1 M HN03. The recorder traces obtained are shown in Fig. 5 and it can be seen that both iron(TI1) and iron(I1) ions are eluted simultaneously under the described experimental conditions. In mixtures containing various concentrations of iron(I1) and iron(TI1) ions in 0.1 M HN03 only the total ionic iron can be determined, whereas the uncharged organic iron passes through the cation-exchange resin column.It was found that ionic iron could be calibrated with a standard solution containing iron(II1) or iron(I1) ions. Separation of Ionic Iron(II1) on Dowex 1-X8 Anion-exchange Resin The same system and conditions of measurement were employed as described for the determination of ionic iron on Dowex 50-X8 resin. The resin was purified by washing it with 0-3 M HC1 and doubly distilled water. Sorption and elution of negatively charged iron complexes from the resin Various negatively charged complexes were studied. It was found that the sorption of iron(II1) ion as FeCL- in strongly acidic (3 M HCl) media was not quantitative. Although iron(II1) ion was quantitatively sorbed on the resin in 0.1 M oxalic acid, clogging of the nebuliser occurred during elution with 0.1 M HC1 owing to crystallisation of the oxalic acid.The use of citric acid as a complexing agent for iron(TT1) was then investigated. It was found that quantitative sorption of iron(II1) ion occurred when both the standard solutions and the resin were prepared in 0.5 M citric acid. Various concentrations of HC1 (0.5-2.25 M) were also investigated for the elution of iron (Fig. 6) and 2.0 M HC1 was found to be optimum. Reproducibility of measurement and detection limit The reproducibility of measurement was tested by performing six successive determinations of 300 pg of iron(II1) [as140 ANALYST, FEBRUARY 1989, VOL. 114 2.25 M HCI 2.0 M HCI 1.5 M HCI 0 ' I - Time Fig.6. Standard solution. Fe(N0,),.9H20. 30C pg of Fc Elutlon curyes for FeIII: resin, Dowex 1-XX. S(k100 mesh, in 0.5 M citric acid; column, 6 ml; and flow-rate, 4.5 ml min-1. 1 min H A 1 A B ) \ - Time Fig. 7. Sorption of 300 ug of Fell (as FeSC>,.7H,O) on Dowex 1-XX resin. 51)-I00 mesh. in 0.5 M citric acid; column, 6 ml. A, Passage of the Fe" solution through the ion-exchange resin at a flow-rate of 1.0 ml min-1; and B. elution from the column with 2.0 M PIC1 at a flow-rate o f 4.5 ml min-' Fe(N03)3.9H20] in 0.5 M citric acid, eluting with 2.0 M HCl. The RSD was found to be +8%. It was also observed that the sensitivity of measurement decreased with increasing number of measurements owing to crystallisation of citric acid and subsequent clogging of the burner slot.Better results and reproduciblity of measurement (i 1%) were obtained when the eluent was collected in a 25-ml calibrated flask and iron was measured by flame AAS employing a conventional method of analysis. Under these conditions the detection limit was 5 pg of iron(II1) ion. Sorption of iron(l1) ion on the resin A standard solution of iron(I1) ion and a solution of the resin were prepared in 0.5 M citric acid, and 300 pg of iron(T1) ion were passed through the ion-exchange resin using 2.0 M HCl as eluent. The eluate was analysed immediately by AAS. The recorder traces are shown in Fig. 7 from which it can be seen that iron(I1) was not sorbed on the resin column. This allows the separation of iron(I1) and iron(II1) ions in their mixture.Determination of iron(l1) und iron(ll1) ions in synthetic mixtures Synthetic mixtures of iron(I1) and iron(II1) ions were pre- pared in 0.5 M citric acid in various concentration ratios. Iron(I1) ion passed through the column and was collected in a 25-ml calibrated flask, and iron(TI1) ion was then eluted with 2.0 M HCl and also collected in a 25-ml calibrated flask. Equivalent standard solutions of iron(I1) and iron(II1) were prepared separately in 25-ml calibrated flasks in the same concentrations as used in the synthetic mixtures and in the same solvent. This enabled the concentration of the separa- ted iron species to be determined in the solutions by flame AAS employing a conventional procedure of analysis. The results are presented in Table 2, from which it can be seen that the proposed technique is satisfactory for the separation of Table 2.Separation of iron(I1) and iron(II1) ions in synthetic mixtures on Dowex 1-X8 (5Crl00 mesh) anion-exchange resin in 0.5 M citric acid. Resin volume, 6 ml; eluent, 2.0 M HCI; flow-rate, 4.5 ml min-l Iron( 11) ion Iron(I1I) ion Added1 Yg 200 100 40 20 200 200 200 Found/ Yg 200 101 42.5 24.5 199 199 199 Recovery, Added/ Foundi Recovery, 100 200 201 100.5 101 200 200 100 106 200 198.5 99.3 122 200 199 99.5 99.5 100 102 102 99.5 40 42.5 106 99.5 20 24.5 122 Yo CCg ug O/O iron(I1) and iron(I11) ions in ratios ranging from 1 : 5 to 5 : 1 (re cover y 99- 1 06 "/" ) . It is also evident that iron can be speciated by measuring either iron(I1) or iron(TII), depending on the total amount of iron and the ratio of iron(I1) to iron(II1) ions.Separation of Uncharged Organic Iron on Amberlite XAD-2 Sorptive Resin The resin was purified by several washings with methanol and distilled water. Iron - phenanthroline standard solutions were used and various organic solvents (methanol, acetone and ethanol) were used as eluents. When direct coupling of the column to the flame AAS system was employed, poor reproducibility of measurement (RSD = i 15-20%) was obtained owing to variation of the nebulisation efficiency and the stoicheiometry of the flame during elution. For this reason sorbed species were eluted from the column into 25-ml calibrated flasks and iron was determined by flame AAS in the conventional way. Influence of p H on iron - phenunthroline sorption and elution from the resin The resin and a standard solution of iron - phenanthroline (100 pg of Fe) were prepared at pH 1-6, employing various concentrations of HN03 (0.005-0.1 M), and sorbed on the resin column, The solvent passing through the resin was collected in a 25-ml calibrated flask and the non-sorbed iron was measured by flame AAS.The efficiency of the sorption is shown in Fig. 8 as a function of pH from which it can be seen that quantitative sorption was obtained at pIi <3.5. In view of these results 100 pg of iron (as iron - phenanthroline) were sorbed on the resin column at pH 2.5 and then eluted into a 25-ml calibrated flask. The iron content was measured by flame AAS. Methanol was found to be an efficient eluent.ANALYST, FEBRUARY 1989. VOL.114 141 Reproducibility of measurement and detection limit The reproducibility of measurement was tested on a standard solution containing 50 pg of iron (as iron - phenanthroline). The RSD for six parallel determinations was found to be k 1% and the detection limit was 2 pg of iron. Sorption of ionic iron(Il) and ionic iron(III) A 100-pg amount of ionic iron(I1) (as sulphate) and ionic iron(II1) (as nitrate and iron citrate complex) was passed through the column resin under conditions that ensured the quantitative retention of uncharged organic iron on the column. The ionic forms of iron(I1) or iron(II1) were found to pass quantitatively through the resin column and this allowed quantitative separation of uncharged organic iron from ionic iron species.Determination of Iron Species in Wine Employing Ion- exchange Sorptive Chromatography and Flame AAS In wines, iron(I1) exists as the positively charged hexaaqua- iron(I1) ion and as uncharged organic iron(I1) (colouring matter), whereas iron(II1) exists as the positively charged hexaaquairon(II1) ion and is also partially complexed with organic acids (such as citric and oxalic acids) to give negatively charged ions. Positively charged iron(TI1) ions tend to form undesirable precipitates with phosphate ions and tannin, which can affect the appearance of the wine (i.e.. its clarity). The ratio of iron(I1) to iron(II1) in wine depends on the oxidation - reduction conditions. During the fermentation process a reducing atmosphere is present and so iron(T1) is the predominant species. However, during decanting and bottling, the wine is in contact with the air and the iron(1I) is then oxidised to iron(II1).To obtain clear wines (i. e. , without a precipitate), commer- cial producers reduce the iron concentration to below 5 mg 1-1 using K4Fe(CN),.23 Owing to the different stoicheiometries of the reactions of iron(I1) and iron(II1) with K,Fe(CN), it is CC ; 75 .- 0 0 1 2 3 4 5 6 PH Fig. 8. Amberlite XAD-2 (2040 mesh) resin as a function of pH Efficiency of iron - phenanthroline sorption (100 pg of Fe) on necessary to know not only the total concentration of iron, but also the ratio of iron(I1) to iron(II1). Commercially, the exact amount of K,Fe(CN), that has to be added is determined experimentally. In this work eight types of red and white wine from different geographical areas were analysed.Some of the wines were bottled by commercial producers, whereas others were local wines produced by small growers. Determination of Total Iron Concentration For the determination of the total concentration of iron in wine 5 ml of the sample were acidified with 0.5 ml of 6 M HCl and the solution was diluted to 25 ml. The iron content was measured (using aqueous standard solutions) by flame AAS. No measurable background was detected. The reproducibility of measurement was tested by performing six parallel analyses on a sample of Malvazija Vipava white wine. The RSD was found to be +0.5%. All other samples were analysed in duplicate. Determination of Iron Species in Wine The proposed method was applied to the determination of iron species in wine, employing various ion-exchange and/or sorptive resins.The results are summarised in Table 3. Determination of ionic iron(IZ) and ionic iron(III) in wine on Dowex 50-X8 cation-exchange resin Direct coupling of the column to the flame AAS system was employed using the optimum experimental conditions deter- mined previously. Samples of wine, standard solutions and column resin were prepared in 0.1 M HN03. An amount equivalent to 10CL150 pg of ionic iron was sorbed on the column (8-50 ml of wine), depending on the concentration of iron in the sample. Although the colouring matter was sorbed irreversibly on the resin, this did not affect the capacity of the ion exchanger (i.e., the sensitivity of measurement for standard solutions was not changed).The sorbed ionic species were eluted with 2.25 M HCl and the absorbances under the elution curve were integrated. The Malvazija Vipava wine sample was subjected to six parallel analyses; the RSD was &3%. All other samples were analysed in duplicate. Determination of uncharged organic iron in wine on Amberlite XAD-2 sorptive resin As an alternative to using Dowex 50-X8 resin. the uncharged organic iron and ionic iron could also be separated on Amberlite XAD-2 resin. The samples and resin column were prepared in 0.02 M HN03 (pH 2.8). Samples of white wine (10 ml) or red wine (1 ml) were then passed through the column. The colouring matter and uncharged organic iron were retained, whereas ionic iron passed through the column and Table 3.Determination of iron species in wine employing various ion-exchange and sorptive resins Resin Species adsorbed Species passing through the column Dowex 50-XS (50-100 mesh) Dowex 1-X8 (5C100 mesh) in 0.1 M HN03 . . . . . . . . . . Ionic iron(I1) and ionic iron(IT1) Uncharged organic iron in 0.5 M citric acid . . . . . . . . Ionic iron(I1I) as the negatively Ionic iron(I1) and uncharged Amberlite XAD-2 (2C-50 mesh) . . . . Uncharged organic iron Ionic iron(I1) and ionic iron(1TI) charged citrate complex organic iron (1) Total iron = ionic iron + uncharged organic iron. (2) Ionic iron = ionic iron(I1) + ionic iron(II1). (3) Uncharged organic iron = total iron - ionic iron. (4) Ionic iron(I1) = total iron - ionic iron(II1) - uncharged organic iron.142 ANALYST, FEBRUARY 1989, VOL.114 Table 4. Determination of total iron and iron species employing various ion-exchange and sorptive resins Total concentration Uncharged Sample of iron/pg ml-* Iron(II), Yo Iron(III), % organic iron, O h Red wine - Teran Kras“ . . . . . . . . RefoSkKoper* . . . . . . BlatinaMostart . . . . . . Renski rizling Ljutomeri . . ZilavkaMostart . . . . . . MalvazijaKoper” . . . . . . MalvazijaVipava* . . . . Kralj evin a * . . . . . . . . White wine - Local wine (small grower). -t Bottled wine (commercial producer). 10.59 20.0 13.68 4.88 3.97 12.06 6.41 2.41 36.2 46.6 5.8 24.2 17.4 40.6 65.5 6.7 63.8 53.4 94.2 75.8 82.6 59.4 32.0 93.3 0 0 0 0 0 0 2.5 0 was collected in a 2.5-ml calibrated flask. Both the ionic iron and the subsequently eluted uncharged organic iron were measured as described previously. The reproducibility of measurement was tested by performing six parallel analyses on the sample of Malvazija Vipava wine and was found to be il% for ionic iron and i2% for uncharged organic iron.All other samples were analysed in duplicate. The results obtained showed good agreement with those obtained for the determination of ionic iron on Dowex 5O-X8 cation-exchange resin. It was found that any iron present in wine was unlikely to be adsorbed on the colouring matter (excluding any precipitate). Determination of ionic iron(lII) in wine on Dowex 1-X8 an ion -exchange resin The wine samples and the resin column were prepared in 0.5 M citric acid. A 3-20-ml volume of wine was passed through the column, eluting with 2.0 M HC1, and the eluate was collected in a 2.5-ml calibrated flask.The iron absorbances were measured as described previously. The colouring matter was again irreversibly sorbed on the resin, but this did not affect the capacity of the ion exchanger. The sample of Malvazija Vipava wine was subjected to six parallel analyses; the RSD was i2%. All other samples were analysed in duplicate. By combining the data obtained for the total concentration of iron, ionic iron and uncharged organic iron, the ratio of iron(I1) to iron(II1) was calculated using the equations given in Table 3. The results are presented in Table 4, from which it can be seen that the ratio of iron(I1) to iron(II1) in wines changes during the ageing process. The ratio of iron(I1) to iron(II1) also depends on the type of wine and the fermenta- tion procedure employed.Conclusion The determination of iron species by means of an ion- exchange chromatography - flame AAS technique has been investigated. A Babington type nebuliser was used for direct coupling. The potential of the system was evaluated for various ion-exchange (Dowex .5O-X8, Dowex 1-X8) and/or sorptive (Amberlite XAD-2) resins. The system enabled all the processes on the column to be followed directly at various flow-rates. Hence the optimum conditions for sorption on the resin and elution from the column could be determined accurately and rapidly. The use of direct coupling was affected by the large differences in the physical properties of the effluent and eluent, which changed the efficiency of nebulisa- tion and the flame stoicheiometry during measurement. The density of the effluent also affected the accuracy of the direct coupling measurements by causing clogging of the nebuliser and/or the burner slot. On the basis of experimental data obtained for synthetic standard solutions a system was developed for the determination of the ratio of iron(I1) to iron(I1I) in wines, employing a combination of various ion-exchange and/or sorptive resins.Ionic iron(I1) and ionic iron(II1) were determined by ion-exchange chromatography on Dowex 50-X8 resin with direct coupling to a flame AAS system, uncharged organic iron on Amberlite XAD-2 resin and ionic iron(II1) on Dowex 1-X8 resin, the last two both via conventional analysis by flame AAS.The authors thank the Boris Kidrii: Foundation for providing financial support and Dr. A. R. Byrne of the Joief Stefan Institute for valuable suggestions and linguistic correction of the manuscript. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Suzuki, K. T., Anal. Biochem., 1980, 102, 31. Botre, C., Cacace, F., and Cozzani, R . , Anal. Lett., 1976, 9, 825. Slavin, W., and Schmidt, G. J . , J . Chromatogr. Sci., 1979, 17, 610. Atwood, J . G., Schmidt, G. J . , and Slavin, W., J . Chromat- ogr., 1979, 171, 109. Ebdon, L., Hill, S . J., and Jones, P., Analyst, 1985, 110, 515. Hill, S., Ebdon. L., and Jones, P., Anal. Proc., 1986, 23, 6. Ebdon, L., Hill, S., and Jones, P., Analyst, 1987, 112, 437. Manahan, S. E . , and Jones, D. R . , Anal. Lett., 1973, 6, 745. van Loon, J. C., Radziuk, B., Kalm, N., Lichwa, J . , Fernandez, F. J., and Kerber, J. D., At. Absorpt. Newsl., 1977, 16, 79. Kahn, N., and van Loon, J. C.. Anal. Lett., 1978, A l l , 991. Yoza, N., and Ohashi, S . , Anal. Lett., 1973, 6, 595. Mohamed, N., and Fry, R. C., Anal. Chem., 1981, 53, 450. Ajlec, R . , and Stupar, J . , Vestn. Slov. Kern. Drus., 1986,3312, 87. Cotton, F. A . , and Wilkinson, G., “Advanced Inorganic Chemistry,” Wiley, New York, 1962, pp. 707-719. Hoshino, H., and Yotsuyangi, T., Tulantu, 1984, 31, 525. Nigo, S . , Yoshimura, K., and Tarutani, T., Talantu, 1981, 28. 699. Lynch, T. P., Kernoghan, N. J., and Wilson, J. N . , Analyst, 1984, 109, 843. Strelow, F. W. E., Rethemeyer, R . , and Bothma, L. J. C., Anal. Chem., 1965. 37, 107. Strelow, F. W. E . , Liebenberg, C. J . , and von Toerien, S . F., Anal. Chim. Acta, 1969, 47, 251. Saito, N., Pure Appl. Chem., 1984, 56, 523. Florence, T. E . , and Batley, G. E . , Crit. Rev. Anal. Chem., 1980, 9, 259. Willis, R . B., and Sangster, D., Anal. Chem., 1976, 48, 59. Troost, G., “Technologie des Weines,” Ulmer. Stuttgart, 1980, pp. 397-403. Paper 8iO3003 K Received July 25th, 1988 Accepted September 27th, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400137

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 114 143 Direct Determination of Iron in Urine and Serum Using Graphite Furnace Atomic Absorption Spectrometry Lian Liang, Patrick C. D‘Haese, Ludwig V. Larnberts and Marc E. De Broe* University of Antwerp, Department of Nephrolog y-H ypertension, University Hospital Antwerp, Wilrijkstraat 10, B-2520 EdegemIAntwerpen, Belgium A simple, rapid and low-cost method for the routine determination of iron in urine and serum using graphite furnace atomic absorption spectrometry is described which may provide an alternative to the more widespread automated spectrophotometric methods. The urine and serum samples were simply diluted with water prior to analysis. Matrix modification was found to be redundant. The standard additions technique or the use of matrix matched standards (addition calibration) was found to be unnecessary and, therefore, the calibration was performed using aqueous standards.For serum analysis the degree of dilution could be reduced by using the less sensitive 302.0-nm resonance line, yielding more precise determinations, and for urine analysis, interferences were eliminated by means of a L’vov platform. The interferences that exist in the presence of nitric acid are also discussed. Finally, the presence of background absorption was investigated by means of Zeeman effect atomic absorption. Keywords: tron determination; graphite furnace atomic absorption spectrometry; human urine and serum; direct standardisation; sample dilution At present several methods are available for the determina- tion of Fe in serum and urine.Among these, flame atomic absorption spectrometry (FAAS) and inductively coupled plasma atomic emission spectrometry (ICP-AES) appear to be used most.’-9 For the majority of these methods, measure- ments are performed after either sample digestion or pre- concentration using a chelating resin. However, these proce- dures are time consuming, laborious and increase the risk of contamination. In addition large sample volumes are required. an important drawback in the analysis of biological samples. Recently, several u ~ r k e r s 7 - ~ have reported the direct determination of Fe and other trace elements in urine and serum using ICP-AES. An analytical method using microlitre amounts of serum has been described by Lewis and O’HaverlO who used simul- taneous multi-element atomic absorption spectrometry with continuum source (SIMAAC) with graphite furnace atomisa- tion.In contrast to FAAS and ICP-AES, methods for the determination of Fe in serum and urine by graphite furnace atomic absorption spectrometry (GFAAS) are scarce. 11.12 Using these methods, serum proteins are precipitated with trichloroacetic acid (TCA) prior to analysis. Glenn and Savory13 have described the use of a graphite rod atomiser for the determination of Fe in serum and McGahan and Fleisher14 have reported a micro-method for the determination of Fe and the total Fe-binding capacity in intra-ocular fluids and plasma using GFAAS. In the latter study TCA was used in combina- tion with Triton X-100 to increase the sensitivity and improve the precision of the method.In order to study the efficiency of desferrioxamine (DFO), a chelator used in the therapy of Fe overload, we have developed a direct method for the determination of Fe in biological fluids without the need for time-consuming and laborious procedures for sample preparation and standardisa- tion. Instrument settings, background correction, the need for matrix modification, interferences, calibration, contamina- tion and choice of wavelength were investigated systematic- ally. Graphite furnace atomic absorption spectrometry was found to be a sensitive, fast and suitable technique for the determination of Fe in biological fluids, permitting the use of very small sample volumes (10 yl). Therefore, GFAAS provides a good alternative to TCP-AES particularly when To whom correspondence should be addressed.multi-element analysis is not required and ICP-AES is not available. Experimental Instrumentation A Model 372 atomic abgorption spectrometer (Perkin-Elmer, Norwalk, CT, USA) equipped with an HGA-500 graphite furnace, an AS-40 autosampler, a 023 recorder and a PRS-10 printer sequencer were used throughout this study. A Zeeman 3030 atomic absorption spectrometer (Perkin-Elmer) equipped with an HGA-600 graphite furnace, an AS-60 autosampler and an Anadex silent scribe printer were used to study background and analyte absorption profiles and to perform internal laboratory quality control experiments. The graphite tubes employed were either uncoated (serum) or equipped with a L’vov platform (urine).The hollow cathode lamp current was 25 mA and measurements were performed at both the 248.3- and 302.0-nm resonance lines using a spectral slit width of 0.2 nm. No background correction was used and the signals were processed in the peak height mode. Materials and Reagents Precuutions tuken to avoid contaminution. All materials coming into contact with the samples were tested as potential sources of Fe. Materials leaching Fe in excess of 0.4 pg 1 - I after being in contact with doubly distilled water for 7 d were discarded. No glassware was used and doubly distilled water was used throughout the study. Samples were injected automatically and the use of acidified solutions was avoided. In order to control the contamination two aliquots of the same sample were always processed simultaneously.Polystyrene tubes ( 5 or 10 ml) (Biolab, Limal, Belgium), 50-ml polypropylene calibrated flasks with tight-fitting poly- propylene caps (Brand, Wertheim, FRG) and polystyrene sample cups were used. The automatic pipettes employed were a 500-yl Assipette (Ignis, Paris, France), a 4CL200 yl Digital Finnpipette (Labsystems, Helginki, Finland) with disposable tips (Labsystems), a 1-5 ml Macro-Transferpettor (Brand) and a 50-yl Lancer pipette (Lancer, Sherwood Medical Industries, St. Louis, MO, USA). Polystyrene bottles (Sterilin, International Medical, Feltham, UK) were used for storing the urine samples.144 ANALYST. FEBRUARY 1989, VOL. 114 Sampling Venipuncrure. For venipuncture. 10-ml sterile syringes (Monovette, Sarstedt, Numbrecht, FRG) equipped with 18-gauge syringe needles (Terumo Europe, Baasrode, Bel- gium) were used.Serum. After coagulation and centrifugation, serum way transferred into 5-ml polystyrene tubes equipped with tight- fitting polyethylene caps (Biolab) by means of an Assipette with disposable tips (Multilab. Zoersel, Belgium). Urine. Each urine specimen collected over a 24-h period was collected in a 2-1 plastic bottle (Sarstedt). After thorough mixing, about 100 ml of each specimen were transferred into a 150-ml plastic recipient for storage at 4 "C. Preparation of Standard Solutions and Sample Dilution Aliquots of a 1 mg ml-1 standard stock solution of Fe (Sigma, St. Louis, MO, USA) were diluted with water to yield working standards containing 0. 20, 40 and 60 pg 1-1 of Fe.The serum samples were diluted either 1 + 99 (high Fe concentration) or 1 + 49 (low Fe concentration) with water. In order to reduce the degree of dilution, Fe in serum was also measured at the less sensitive 302.0-nm resonance line. In this instance the samples were diluted 1 + 19 and working standards containing 0. 100, 200 and 300 pg 1 - 1 of Fe were used. The urine samples were diluted 1 + 4 with water. In order to obtain homogeneous urine samples, a 10-ml specimen of each well mixed urine sample collected over a 24-h period and stored at 4°C was transferred into a 12-ml polystyrene tube equipped with a tight-fitting polyethylene cap. The tube was then shaken vigorously before sample preparation. Aliquots (10 pl) of the standards and the urine or serum sample solutions were dispensed automatically into the graphite furnace.For the serum samples, the graphite tubes were uncoated, whereas for the urine samples, a L'vov platform was inserted. Results and Discussion Instrument Settings The influence of the char and atomisation temperatures on the Fe absorbance signal for urine and water using a L'vov platform is shown in Fig. 1. For water, Fe losses were observed at 1400 "C compared with a temperature of 1300 "C for urine. 0.3 a, (I m + 8 0.2 Q 0.1 a) '-% 1000 * 1250 1500 ' b) I , 2000 2250 2500 2750 Tern peratu rePC Fig. 1. Relationship between (a) char and ( b ) atomisation tempera- tures and the absorbance signal for the determination of Fe in (A) aqueous solutions of serum ( I + 99) (uncoated graphite tube) and (0) urine (1 -t 4) (platform) and in water using (0) a platform and (A) an uncoated graphite tube.The atomisationichar temperatures used were 220011 100 "C €or serum and 240011200 "C for urine When uncoated graphite tubes were used for serum and water analysis a decrease of the absorbance signal was observed first when the ashing temperature was above 1200°C (Fig. 1). Hence char temperatures of 1200 and I100 "C were selected for urine and serum, respectively. At these temperatures, the maximum signals were obtained without significant back- ground absorption. Although water behaved similarly to both serum and urine in that no signal plateau was observed with increasing atomisation temperatures (Fig. 1) ~ atomisation wai performed at 2400°C for urine and at 2200°C for serum.Hence an adequate sensitivity was maintained and the lifetime of the graphite tubes was prolonged. The furnace conditions employed are summarised in Table 1. Contamination Owing to its prevalence, control of contamination is essential in the determination of Fe in biological fluids, particularly when a highly sensitive technique such as GFAAS is used."-'7 In this study a number of precautions were taken to avoid the addition of extraneous Fe.17 The number of manipulations and reagents required was kept to a minimum and the use of acidified solutions was avoided both in sample storage and sample preparation. Interferences, Calibration and Matrix Modification In contrast to the behaviour of serum, it was found that for urine analysis the slopes of the aqueous calibration graphs differed from those obtained using the standard additions technique when uncoated graphite tubes were used.However, when the L'vov platform was used, interferences were absent. Table 1. Furnace conditions for the determination o f Fe i n urine and serum Urine Serum Step 1 2 3 3 5 1 2 3 4 5 TemperaturePC 100 200 1200 2400 2700 100 120 1100 2200 2700 Ramptimeis . . 15 5 30 0 1 10 10 30 0 1 Holdtimek . . 15 5 10 5 3 10 10 15 5 3 Gas flow-rate/ ml min-' . . . . 300 300 300 0 300 300 300 300 0 300 0.20 a, C m +? n" 0.10 4 /4 Fe concentrationipg I-' Fig, 2. Calibration graphs for Fe in ( X ) water, (M) 0.3% ViV HNOi and (A) 0.5% ViV HNO, and in serum diluted (1 + 99) with (0) water, (0) 0.3%, ViVHNO, and ( A ) 0.5% WVHNO,. Atomisation is from the wallANALYST.FEBRUARY 1989. VOL. 114 145 0.20 0 2 0.10 Q 0 20 40 60 Fe concentrationiug I-' Fig. 3. Calibration graphs for Fe in (e) water and (a) 0.3% Vil' HNO, and in urine diluted (1 + 4) with (0) water and (0) 0.3% V/V HNO,. Atomisation is from a L'vov platform Table 2. Comparison of the slopes of the calibration graphs obtained uith aqueous standards and with the method of standard additions Serum analysis Urine analysis slope* slope>k Sample group n McankSD n MeanLSD Aqucousstandards . . 1 0.40 1 0.43 Dilution. 1 + 99-i- . . 17 0.40k 0.02 - - Dilution. 1 +49+ . . 9 0.39 k 0.02 - - - 6 0.43 t 0.01 Dilution. 1 + 4t . . - DFOpresent . . . . 13 0.40 k 0.02 6 0.42 k0.01 Overall . . . . . . 26 0.40 k 0.02 13 0.43 k 0.01 The slopes are reported as 1000 times the absorbance signal Per Pg.*(. Samples were diluted with water. For GFAAS measurement, nitric acid is generally accepted as favouring the complete ashing of biological samples and hence preventing the presence of carbonaceous residues in the graphite furnace. In addition, nitric acid has been widely used to stabilise urine samples during storage. However, in spite of this. we have found the use of nitric acid to be superfluous for se ve r a1 reasons . Firstly, as Fig. 2 shows for the serum analysis. the slopes of the 0.3 and 0.5% VIV nitric acid calibration graphs differed from those obtained with the aqueous and serum matched standards. For the urine analysis. the aqueous, 0.3"/0 VIV nitric acid and urine matched calibration graphs were parallel.However, when urine was diluted 1 + 4 with 0.3% VIV nitric acid the slopes were different (Fig. 3). These data indicate that in the presence of nitric acid severe multiplicative interfer- ences were observed, necessitating the use of the standard additions method. In contrast, no interference was observed when water was used for sample dilution and for the preparation of calibration standards. Secondly, in the absence of nitric acid, there was no significant decrease in the concentration of Fe in urine stored at 4 "C for at least 11 d. Thirdly. we found that the absence of nitric acid had a beneficial effect on the lifetime of the graphite tube, allowing 150 firings without a significant decrease in the sensitivity. Fourthly. and a well known fact, the absence of acidified solutions considerably reduces the risk of contamination, and hence minimises the difficulty in carrying out the assay.This is an important advantage in the determination of trace levels of ubiquitous elements, such as Fe. In order to ascertain whether matrix variations in the serum and urine samples from different subjects interfered with the measurement, the slopes of the calibration graphs of the Table 3. Precision of thc determination of Fe in urine and serum Serum Urine Sample Mean/ SDI C'V, Mean1 SDI CV, Runtype No. pgl-1 ugl-1 "/o mgl-* mgl-1 "A, - 199 5.0 2.5 4.05 0.078 1.9 Within-run Between-run 1 12.3 0.65 5.2 1.01 0.041 4.1 2 169 6.0 3.6 3.71 0.169 4.6 3 454 27 5.9 6.14 0.223 3.6 aqueous standards, 26 different serum matched standards and 12 different urine matched standards were compared.The results are given in Table 2. The Fe concentration in serum ranged from 0.55 to 6.14 mg 1-1 and that in urine from 5 to 560 pg 1-1. The data given in Table 2 demonstrate that, although the matrices of the urine and serum samples varied individu- ally, the serum samples were diluted to a different degree and that when DFO was present, the slopes obtained were identical with those of the aqueous calibration graphs. Therefore, in the present method of analysis, aqueous standard graphs were used and the samples were simply diluted with doubly distilled water. For the determination of Fe by GFAAS.1-?.15 both Mg(N03). and N€ 13N03 have been used for matrix modifica- tion. Although we found that in the presence of Mg(N03)2 the char temperature could be increased by 100"C, this matrix modifier was found not to be essential i n the present method as neither background nor chemical interferences were observed in its absence.As for nitric acid, the sensitivity of the Fe determination decreased when NH4N03 was used as matrix modifier; this was confirmed by Sturgeon et al. 18 who reported a decrease in sensitivity when NH4N03 was used in the analysis of sea water. Background Correction Using the optimised instrumental condition5, the analyses of serum samples diluted 20-, 50- and 100-fold gave identical results when determined both with and without deuterium background co r r ect i o n . The background absorbance profiles of urine were investi- gated using a Zeeman 3030 atomic absorption spectrometer.The results showed that with the proposed methods no background absorption could be observed and this allowed us to perform both serum and urine analyses without the need for background correction. Homogeneity of Urine Samples Homogenisation of the urine samples was found to be critical. We found that the results for the determination of Fe in urine containing a precipitate were not reproducible and that the results obtained before sample homogenisation were much lower than those obtained after homogenisation. For exam- ple, the Fe concentration was found to be 185 and 560 pg I-' before and after homogenisation, respectively, indicating that most of the Fe was probably absorbed by the precipitate. Therefore, the urine samples were homogenised, using a three-step mixing procedure, from the time of sampling until measurements were performed.This resulted in excellent within- and between-run coefficients of variation (CVs) (Table 3). Choice of Spectral Wavelengths Until now, the 248.3-nm resonance line has been used most widely for the determination of Fe by AAS. However, with GFAAS, owing to its high sensitivity (10 pg per 0.0044 A), the146 ANALYST, FEBRUARY 1989, VOL. 114 0 1 2 3 4 5 Uncorrected serum iron concentrationimg 1-1 Fig. 4. Comparison of the determination of Fe using two different instruments. The background corrccted and the uncorrected concen- trations were obtained using the Zeeman 3030 and Model 372 atomic absorption spectrometers. respectively. tz = 32. r = 0.995 and y = I .06.Y -0.04 serum samples must be diluted ( 1 + 49 or 1 + 99). Other norkers8." have reported that the use of large dilutions limits the analytical performance of GFAAS.In order to overcome thi4 drawback the possibility of using the less sensitive Fe absorption lines at 302.0 and 317.9 nm was examined. The respective sensitivities of these lines were found to be 37 and 65 pg per 0.0044 A. The aqueous calibration graphs were linear up to 60 pg 1-l at the 248.3-nm line and up to 300 pg 1 - 1 at the 302.0-nm line; hence the suitability of the 302.0-nm line for serum analysir. However, in order to exclude matrix interferences resulting from the lower degree of dilution (1 + 19) at 302.0 nm, the slopes of different serum matched and aqueous calibration graphs were compared.These were all identical, again indicating the absence of multiplicative interferences. More- over. the reduced degree of dilution had no effect on the background. the measurements performed both with and without deuterium background correction yielding identical results. The precision at both lines was compared after determining Fe in serum at 302.0 nm (dilution, 1 + 19) and 248.3 nm (dilution, 1 + 99). For the serum analysis within-run CVs obtained for specimens containing 0.8, 2.2 and 6.5 mg 1-1 of Fe were 2.0. 1 .5 and 3.0%, respectively. at 302.0 nm and 7.3, 3.6 and 3.0%, respectively, at 248.3 nm. The lower CVs at 302.0 nm indicate that more precise results were obtained uhen the serum samples were diluted only 1 + 19. Ilence, for serum analysis the 302.0-nm line war preferred to the 238.3-nm line, whereas for urine analysi5 the 248.3-nm resonance wavelength was chosen as a dilution of only 1 + 4 was required for patients with Fe overload and for healthy subjects.Analytical Performance Detectiori limit; characteristic amount The detection limit for Fe in urine. expressed as twice the standard deviation (SD), was 8 pg, which corresponds to 4 ltg 1-1. For Fe in serum. the detection limit was 4 pg ( i t . . 40 iig 1-1). The amount of Fe yielding a 0.0044 A signal (the characteristic amount) was 10 pg. Preckiori und accuracy Within-run CVs were evaluated after determining the Fe concentration in ten dilutions of one serum and one urine sample. Between-run CVs u ere calculated after analysing three urine and three serum samples with various Fe concen- trations five times over a period of 11 d.The results are given in Table 3. Table 4. Determination o f Fe in internal quality control serum samples Control No. Found/mg I-' Recommended valueimg I- Precipath U 157554 . . 1.48 1.49 (1.2&-1.70) Precinorm U 157471 . . 1.14 1.11 (0.95-1.27) The mean recoveries, calculated by comparing the slopes of the matrix-matched calibration graphs with those obtained for the aqueous standards, were 101% (range, 95-107%) for serum and 98% (range, 93-105%) €or urine. The low within- and between-run CVs together with the nearly 100% recoveries for both serum and urine indicated that both the precision and accuracy of the proposed method were satisfactory. In an internal laboratory quality control experiment, the results for the determination of Fe in 40 serum samples obtained using the Model 372 atomic absorption spectrometer were compared with those obtained using the Zeeman 3030 atomic absorption spectrometer.The graphite furnace pro- gramme settings of both instruments were identical. The results, shown in Fig. 4, are in good agreement. As no background correction was used with the Model 372 spec- trometer, the results showed that broad band, structured and line background absorption were absent, hence excluding the presence of additive errors caused by background absorp- Table 4 shows the results obtained for Precipath U and Precinorm U quality control serum samples (Boehringer Mannheim, Mannheim, FRG) using the proposed method, together with the recommended values.tion. 1 Y .20 Conclusion Graphite furnace atomic absorption spectrometry has been found to be a reliable and suitable technique for the determination of Fe in the routine analysis of biological fluids. Together with a significant increase in sensitivity compared with FAAS, it has the additional advantage that only very small samples are required. Also, the method is rapid as no time-consuming procedures are required for standardisation or sample preparation. Contamination, which is often a major problem when highly sensitive methods such as GFAAS are used for the determina- tion of ubiquitous elements (e.g., Fe), was avoided by limiting the number of manipulations and reagents employed. We greatly appreciate the secretarial work of Erik Snelders.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Sprague, S . , and Slavin, W., At. Absorpt. Newsl., 1965,4.228. Olson, A . D., and Hamlin, W. B., Clin. Chem., 1969, 15,438. Burguera, M., Burguera, J . I., and Rivasp, C.. A t . Spectrosc., 1986, 7, 79. Uchida, T., and Vallee, B. L., Anal. Scz., 1986, 2, 243. Barnes, R . M., Fodor, P . , Inagaki, K., and Fodor, M., Spectrochim. Acta, Part B , 1983, 38. 245. Barne\, R. M., and Fodor, P., Spectrochirn. Actu, Purt B , 1983. 38. 1191. Mianzhi, Z . , and Barnes, R . M., Appl. Spectrow., 1985, 39. 793. Leflon, P . , and Plaquet, R.. Clin. Chem., 1986, 32, 521. Nixon, D. E., Moyer, T. P., Johnson, P., McCall, J . T., Ness, A. €3.. Fjerstad, W. H., and Wehde, M. B.. Clin. Chem., 1986, 32. 1660. Lewis, S. A., and O'Haver, T. C.,Anul. Chem., 1984.56,1651.ANALYST, FEBRUARY 1989, VOL. 114 147 11. 12. 13. 14. 15. 16. 17. Olsen, E. D., Jatlow, P. L., Fernandez, F. J . , and Kahn, H. L., Clin. Chem., 1973, 19, 326. Yeh Y.-Y., and Zee P., Clin. Chem., 1974, 20, 360. Glenn, M. T., and Savory, J . , Anal. Chem., 1973,45, 203. McGahan, M. C., and Fleisher, L. N.. Anal. Biochem., 1986, 156, 397. Slavin, W., “Graphite Furnace AAS, a Source Book, Part No. 0993-8139.” Perkin-Elmer, Norwalk, CT, USA, 1984, p. 18. Kragten, J., and Reynaert, A. P., Talanta, 1974, 21, 618. D’Haese, P. C., Van de Vyver, F. L., de Wolff, F. A., and De Broe. M. E . , Clin. Chem., 1985, 31, 24. 18. 19. 20. Sturgeon, R. E., Berman, S. S . . Desaulniers, A., and Russell, D. S . , Anal. Chem., 1979, 51, 2364. Slavin, W., Spectrochirn. Actu, Part B. 1987, 42, 933. Welz, B., Fresenius Z. Anal. Chem., 1986, 325, 95. Paper 8/01 865 K Received May 12th, 1988 Accepted October 6th, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400143

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 113 149 Fourier Analysis Method for Temperature Compensation of a Microcomputer-controlled Piezoelectric Crystal Sulphur Dioxide Sensor R. D. Snook Department of Instrumentation and Analytical Science, UMIST, P.O. Box 88, Manchester M60 700, UK P. E. Zaft" Laboratory of the Government Chemist, Cornwall House, London SEI SXY, UK A Fourier analysis method has been developed for correction of the temperature response of a microcomputer-controlled piezoelectric crystal sensor. The method can compensate for changes i n frequency between 15 and 55 "C for a Quadrol [NrNrN',N'-tetrakis(2-hydroxypropyl)ethylenediamine] coated crystal. When the crystal is used as a sensor for SOzr however, the correction method fails to compensate for temperatures above 35 "C.It is thought that the failure is due t o a change i n adsorption isotherms above this temperature. However, in the range 15-35 "C the compensation for temperature response works well. Keywords: Piezoelectric crystal sensor; sulphur dioxide; Fourier compensation for temperature response There have been many attempts to employ chemically-coated piezoelectric crystals as chemical sensors for vapour and gas detection. All of the proposed methods utilise piezoelectric crystals as the frequency controlling element of a tuned oscillator. When a mass M is coupled to the electrode surface of the crystal of area A a proportional decrease in frequency ( - A q occurs according to the Sauerbrey equation The mass sensitivity of a typical crystal oscillating at 10 MHz is about 1-2 ng Hz-1.Selectivity towards a particular target vapour of gas is achieved by coating the crystal with a substance which preferentially adsorbs the species of interest. It is desirable for this coating to be hydrophobic because reaction of these surface coatings towards atmospheric water vapour will also produce a frequency response. Coatings have been reported to facilitate the detection of CO, C 0 2 , H2, HC1, H20, H2S, Hg, NH3 and SO2 as well as of various organic compounds: amines, aromatics, hydrocarbons and organo- phosphorus compounds.l--h There have also been reports of the use of humidity correction algorithms and microcomputer correction procedures to compensate for the residual interfer- ence of water vapour.7.8 The theoretical basis for the use of these devices has been described well"10 and will not be discussed here.What is apparent from these treatments and from experimental observations is that the crystal frequency also depends on its temperature. This effect can be reduced in a number of ways: first, by the use of piezoelectric crystal cuts that have a low temperature coefficient, such as AT cut, and secondly, by closely regulating the temperature of the detector. However, the latter creates difficulties in ambient monitoring applica- tions. Finally, an accurate temperature-corrected frequency can be measured by beating against a matched reference crystal the thermal characteristics of which are the same. In practice it is difficult to find such a matched pair and tedious to determine the thermal characteristics of a large number of possible matched crystals.In this paper we describe a microcomputer-controlled system that measures the fre- quency and temperature and produces an equation for the thermal response as a 21 term Fourier series. The thermal response is stored in the microcomputer for subsequent use in -AF = 2.3 X 10fJPdMIA * Present address: Chelsea Instruments Limited, Unit 9, Avon Business Centre, Avonmore Road, London W14 8TS. correcting for changes in temperatures of the piezoelectric crystal sensor. The temperature-correction procedure has been tested by applying it to the measurement of sulphur dioxide in a flowing stream of nitrogen at different temperatures and correcting the data to frequency values expected at 25 "C. Comparison with the true frequency value at 25 "C shows that the procedure works well up to 35 "C but fails to correct completely the temperature response between 35 and 50 "C.Experimental Instrumental Apparatus The experimental system is shown schematically in Fig. 1. In the detector cell,s the sample gas is split into two streams which impinge on opposite faces of the crystal. For improved sensitivity the two streams are brought by jets very close to the centre of the crystal which is the most mass sensitive part. An additional feature is the temperature sensor which protrudes into the crystal enclosure and is brought close to the crystal. The interface rack holds 16 function boards on double eurocard PCBs. A software controlled method of addressing the individual boards has been developed to enable the user to access the boards without having to know their exact position in the rack.Temperature measurements are accomplished by an RS 590 kH integrated circuit temperature sensor (RS Components, Data Sheet 3992) for which the output current is proportional to the absolute temperature (1 pA K-1). The current is transmitted over a twisted pair line to a variable load resistor in the interface rack. A voltage is input to a multiplexer board on which the offset and gain are used to signal condition the output voltage so that it lies between -10 and 10 V over the temperature range G80"C. Because the offset and gain are digitised we use the variable resistor as a fine adjustment. The signal then passes to a 12-bit ADC board and then the computer reads the conversion value from bytes 0 and 1 with a resolution of 0.02 degrees per bit.Calibration is only necessary at one temperature because of the linear output of the temperature sensor. In the time it takes to obtain one frequency measurement (1 s) we obtain and average 50 temperature measurements and can observe a noise level corresponding to 0.02"C that is the same as the ADC resolution.150 ANALYST, FEBRUARY 1989, VOL. 114 $:.Y.+ Drying Water-bath r-----l Gas i n c FM board interface -1 rack f Water-bath Parallel twin cable 624 cone Water level Piezoelectric / crystal a Temperature sensor Micro- computer ,5j plotter 1 Gas out Fig. 1. ( a ) Microcomputer and experimental system. ( b ) The detector cell assembly Piezoelectric Crystals and Oscillator The piezoelectric crystals (PZX) used for frequency measure- ments are AT cut, operated at the fundamental mode resonant frequency of 10 MHz and have silver electrodes (Senator Crystals, Mitcham).They were obtained in HC6/U holders and the crystals were not sealed inside their cans. The detector crystal is connected to the rest of the oscillator by an HC6/U socket, a 0.5-m parallel twin cable and a miniature BNC plug and socket. The oscillator is a field effect transistor (FET) type and was built to produce a low-noise sinusoidal output at TTL levels. The output buffer prevents the frequency measuring circuitry from degrading the output signal. The frequency is measured by a purpose-built board in the interface rack at a resolution of k1.0 Hz. The conversion time is the reciprocal of the resolution and in this instance is 1 s.Fourier Series The raw frequency versus temperature data are sorted into ascending temperature, rounded to the nearest tenth of a degree, and all frequency measurements at a particular temperature are collected together and averaged. Each new data pair is, therefore, valid over a range of kO.1 "C. It is easy to visualise this, graphically, as a histogram of 400 columns. The frequency data are stored in an array with the index, i, taking values from 0 to 399; i is related to the temperature datum using the expression given in equation (1) and always takes integral values. The minimum and maximum indices correspond to the temperatures 15.1 and 55.0 "C, respectively. i = lOT, - 151 (always an integer) .. (1) If we let v, be the frequency datum for the ith point, the corresponding temperature datum T, is given by equation (2): i + 151 10 Mathematically, the histogram is a step function: where the index, i, is the integral part of x , an algebraic variable that is defined in the range 0 < x < 400. This function is neither continuous nor differentiable; however, it can be represented by a trigonometric series of the type T, = - (i = 0,1,2 .. ,399) . . . . (2) g(x) = v, (i = 0,1,2 ... 399) . . . . (3) 7- where r takes integral values, ao, a, and b, are constants and 1 is a scaling factor. Because equation (4) is unchanged on replacing x by x + 21k, where k is an integer, it is a periodic function in x of period 21. Equating the interval over which x is defined (0 < x < 400) as the period 0 < x < 21 we obtain 1 = 200.Coefficients ao, a, and br can be found so that f(x) tends to g ( x ) as r tends to 03, when g ( x ) is continuous and differentiable and f(x) tends to [g(x - ax) + g(x + Sx)]/2, elsewhere. The coefficients in equation (4) are defined by [f(x) cos xr(x/l)]dX ( r = 0,1,2.. .) . . ( 5 ) 7 'il' [f(x) sin xr(xl1)]dx ( Y = 0,1,2.. .) . . (6) Substituting in equation ( 5 ) using the identity f(x) = g ( x ) and 1 = 200 we obtain 400 a, = [ g ( x ) cos nr(x/l)]dx . . 0 Splitting equation (7) into 400 integrals gives which on integrating for Y = 0 gives 399 1 xr a, = - C vi [sin nr(i + 1 - sin nr(i/l>] i = o This is rearranged to give 399 Similarly 300 1 xr b , = - Z (v;- vj-~)[cosnr(i/I)] ( r = 1,2 ...) i = 1 .. and integrating equation (8) for r = 0 gives a() = 'C vj . 1 By eliminating the high frequency terms, i.e., limiting r to 10, a smoothed-out representation of the data can beANALYST, FEBRUARY 1989, VOL. 114 151 obtained. This smoothing action would have been a disadvan- tage if the whole Fourier range had been used for valid data with the first and last data points scaled to 0 and 399d200, respectively. The discontinuity between the frequency values at 2nn may indeed be large, requiring many high frequency terms to represent well. Therefore under the limit of Y = 10 the data would not be well represented over a large region to either side. Hence the necessity of nesting as described below. Software has been written to analyse the data contained in the array and to calculate the Fourier coefficients up to Y = 10 (Table 1).For the integration programme to operate effici- ently, each array element must contain data. Therefore we first interpolate data for empty array elements. Under the angular scaling 55.0 "C is equivalent to 15.0 "C, so the data can be considered to be periodic and, therefore, it is valid to perform a single interpolation from the last to the first data point using a cubic expression. The valid data is thus nested inside the Fourier range so that the analysis can be performed on any range of data within 15.1-55.O"C. In addition. undesirable periodic effects are avoided. It must be remembered that the resulting Fourier series is valid only within the limits of the original data.Results and Discussion Temperature Correction The PZX was installed in the cell and gas line previously described and the cell immersed in a thermostatically-con- trolled recirculating water-bath. The temperature dependence of the crystal frequency was recorded at temperatures between 15 and 55 "C using the microcomputer and interface system. Fig. 2(a) shows the raw frequency response with respect to temperature for a typical 10-MHz AT cut crystal as stored and Table 1. Fouricr coefficients a, and br for values of r r 0 1 2 3 3 5 6 7 8 9 1 0 ar 12 144.826 -8.955 -5.879 1.503 2.511 -0.143 -0.176 -0.001 -0.134 0.132 -0.056 b, 2.505 -3.975 -4.142 0.122 0.762 0.054 -0.132 -0.226 0.087 0.217 40 35 30 25 N 2o I 0 a l 5 10 5 0 represented using the microcomputer system. This response graph can be obtained reproducibly over several runs, the deviation between runs being no greater than that observed within any one run.The temperature data thus acquired were then averaged and the Fourier series describing the response graph was obtained. This graph is shown in Fig. 2(b). The smoothed and reconstructed response graph can now be used to correct the over-all response to a target gas, in this instance sulphur dioxide (SO,) in a flowing stream of dry nitrogen. This mixture has been chosen as a model to study temperature effects to avoid the complicating interferences due to changes of humidity, which would occur in the real situation of monitoring SO2 in air. Fig. 2(c) shows the corrected response graph over the range 35-55 "C for an uncoated crystal.The correction is to a temperature of 25 "C. Determination of SO2 For the determination of SO2 the situation was somewhat more complicated as the crystal was coated with a hydro- phobic chemical that selectively binds SO2. The coating was N , N , N' ,N'-tetrakis(2-hydroxypropyl)ethylenediamine (Quadrol) which was added, dropwise, in a solvent (1.7 X 10-4 mlV in dichloromethane) to the crystal surface and then allowed to cure. The curing process is important as an SO2 measurement taken during evaporation of solvents would present erroneous results concerning the detection of the target gas. The curing was conveniently followed by monitor- ing the frequency according to the Sauerbrey equation. Fig. 3 illustrates the process of curing over successive temperature cycles.Curve 1 was the temperature response when the same crystal was coated with Quadrol and then heated through 15-55 "C. It can be seen that at temperatures above 35 "C there was progressive loss of mass, which we assumed to be evaporation of the solvent and Quadrol. When the temperat- ure cycle was repeated, the temperature at which mass loss was observed occurred at progressively higher temperatures (curves 2-4) until the coated crystal behaved in the same way as the uncoated crystal. Repeated temperature cycles of the uncoated sensor are shown in Fig. 4. Within experimental error (relative standard deviation of 7% in AF for four consecutive temperature cycles) the response of the sensor with respect to temperature was the same for each cycle and hence the same general form as the cured, coated crystal, although the nominal starting frequency (AF = 0 Hz) was higher owing to the absence of coating.It can be assumed therefore that the coating was cured after four heating cycles and that the device was in its prepared state for use as a sensor. At temperatures above 55 "C rapid mass loss from the cured, coated crystal corresponded to loss of Quadrol itself from the crystal surface. . . . . . . . . . . . . . . . . -.-.- .---..."..-*I - -- ..--...--- - .. . ._ - . ___ - .__.. -__ -____. ..-. __. . . . . . . . . . ...... I I I I I I I I I I I I 20 30 40 50 20 30 40 50 20 30 40 50 Temperatu re:"C Fig. 2. data and (c) temperature corrected data Fourier analysis and temperature correction of a piezoelectric crystal sensor.(a) Uncorrected frequency data, ( b ) reconstructed152 ANALYST, FEBRUARY 1989, VOL. 114 32 "i 24 1. I A2 I 3 1 4 -8 -16 _._- -24 1 -28 ' I I I I I I I I 20 25 30 35 40 45 50 55 Tern perat u re/"C Fig. 3. cycling. Plots 1-4 are defined in the text Frequency response of a coated crystal during temperature 48 I I N 16 -28' I I I I I I I 20 25 30 35 40 45 50 Tern pera t u re/"C Fig. 4. ture cycling. Plots 1-4 as in Fig. 3 Frequency response of an uncoated crystal during tempera- 0 -10 - 20 -30 - 40 I C -50 - 60 - 70 - 80 - 90 -100 N a I I I I 20 25 30 35 40 45 TemperaturePC Fig. 5. v.p.p.m. of SO2 Temperature correction of the SO2 sensor exposed to 17 To study the temperature effect on SO2 determination the detector cell was initially brought to 19 "C with the sensor under an atmosphere of dry nitrogen (< 1 % relative humidity) and the base-line frequency and temperature were recorded.The cell and sensor were then raised to one of five temperatures (20-50 "C) with the sensor under an atmosphere of 17 v.p.p.m. (parts per million by volume) of SO2 in dry nitrogen. Frequency and temperature measurements were taken at each of the five temperatures and the temperature correction procedure was applied. The result of applying the temperature correction pro- cedure is shown in Fig. 5 . The chosen corrected temperature was 25 "C and it can be seen that the correction procedure works well up to about 32-33 "C after which the corrected - AF value gradually increases. Clearly some other process competes with temperature in changing the PZX frequency when SO2 is sorbed on to the coated crystal.There are several possible explanations for the breakdown of the correction procedure which can be postulated after careful analysis of the data obtained. Close inspection of the plots presented in Fig. 5 reveals that the gradient of the tangent to part A of the graph (up to ca. 32 "C) is the same as that for the equivalent part of the cured, coated crystal response with respect to temperature (Fig. 3), at ca. 0.8 Hz "C-1, whereas the gradient of part B of the graph is greater (1.6 Hz "C-1) when compared with the equivalent part of the cured, coated crystal which changes to 0 Hz "C-1. As the temperature correction algorithm is based on Fourier coeffi- cients obtained from the cured crystal plots shown in Fig.3 the method breaks down progressively above 30 "C. The break- down becomes significant at temperatures above 35 "C. This apparently negative result is in fact valuable as it reveals a completely different mechanism responsible for the frequency response above 35 "C. A likely cause of the increase in the rate of change of frequency with temperature is desorption of residual water vapour from the PZX cell and its subsequent deposition on the crystal, therefore increasing the mass adsorbed and hence decreasing the frequency of the PZX. This mechanism, however, would also be manifest in the experiments con- cerned with the frequency response of the cured, coated crystal exposed only to dry nitrogen. There is no evidence of this behaviour and therefore the possible role of residual water vapour can be eliminated.Another conceivable mechanism is a change in the proper- ties of SO2 sorption on to Quadrol with temperature. Intuitively it may seem likely that the rate of desorption of SO2 molecules from the surface would increase when the temperat- ure is increased. However the results presented in Fig. 5 suggest that the mass loading is increased. This apparently anomalous behaviour could be related to the adsorption capacity of the Quadrol coating, assuming no interaction between sorbate molecules at each site. For small values of the fraction of sites occupied ( O ) , the Freundlich isotherm can be written as the equation11 KT E( ) In0= - lnp + constant . . . . (13) where E~ is the energy of adsorption and p is the partial pressure, which in this experiment is effectively constant.Therefore, an increase in temperature may actually increase the fraction of sites available for occupancy. This is especially true if the nature of the sorption process changes, i.e., from physical adsorption to chemisorption where the energy of adsorption for the latter is much higher than for physical adsorption. These explanations are offered only as tentative explana- tions for the failure of the temperature correction method at temperatures above 35 "C. Carefully constructed experiments must be carried out to control vapour concentrations and pressures during the sorption process in order to elucidate the apparently anomalous sorption behaviour of the Quadrol- coated crystal towards SO2 at temperatures between 35 and 55 "C. Experiments of this nature are presently being carried out using a more sophisticated test rig and more sensitive surface acoustic wave sensors as substitutes for piezoelectric crystals. We would like to thank the Laboratory of the Government Chemist for financial support (to P. E. Z.) under the SERCKASE Studentship Scheme at Imperial College. References 1. 2. 3. 4. 5. 6. Hlavay, J . , and Guilbault, G. G . , Anal. Chem, 1977,49, 1890. Webber, L. M., Hlavay, J., and Guilbault, G. G., Mikrochim. Acca, 1978, 1, 351. Guilbault, G. G., Ion Sel. Electrode Rev., 1980, 2, 3. Guilbault, G. G., Anal. Proc., 1982, 19, 68. Alder, J . F., and McCallum, J . J . , Analyst, 1983, 108, 1169. Karmarkar, K. H., and Guilbault, G. G., Anal. Chim. Actu, 1974, 71, 419.ANALYST, FEBRUARY 1989, VOL. 114 153 7. 8. 9. 10. Fielden, P. R . , McCallum, J. J., Stanios, T., and Alder, J. F., Anal. Chim. Acta, 1984, 162, 8.5. Fielden, P. R., McCallum, J. J., Volkan, M., and Alder, 3. F., Anal. Chim. Acta, 1984, 162, 75. Sauerbrey, G. Z . , Z. Phys., 1964, 178, 457. Stockbridge, C. D., Vac. Microbalance Tech., 1963, 3, 5.5. 11. Rose, J . , “Dynamic Physical Chemistry,” Pitman and Sons, London, 1961, p. 300. Paper 8103004 I Received July 25th, 1988 Accepted October 21st, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400149

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 114 155 Photolytic Interface for High-performance Liquid Chromatography - Chemiluminescence Detection of Non-volatile N-Nitroso Compounds James J. Conboy and Joseph H. Hotchkiss" Department of Food Science, Institute of Food Science, Stocking Hall, Cornell University, Ithaca, NY 14853, USA A photolytic interface between high-performance liquid chromatography (HPLC) and a chemiluminescence detector has been developed for the trace detection of non-volatile N-nitroso compounds in biological matrices. A chromatographic effluent containing separated N-nitrosoamino acids and N-nitrosamides is introduced into a glass coil with a purge stream of He and irradiated with ultraviolet light. Nitrogen oxide, cleaved by photolysis, is separated rapidly from the solvent through a series of cold traps and carried by the He into the reaction chamber of a chemiluminescence detector.The method is compatible with most types of HPLC, especially reversed-phase, and yields low-nanogram sensitivity for underivatised N-nitrosoamino acids and N-nitrosamides. The detection of a model N-nitrosamide, trimethylnitrosourea, in spiked porcine gastric fluid (42 pg I - I ) , and of N-nitrosoproline and N-nitroso-I ,3-thiazolidine-4-carboxylic acid, in spiked human urine (7-8 pg I-?), is demonstrated. Keywords : N- Nitroso compounds; high-performance liquid chromatography - thermal energy analysis; N -nitroSamides The majority of over 300 N-nitroso compounds tested for carcinogenicity in laboratory animals have been positive. 1 Humans are exposed to N-nitroso compounds from a variety of sources including food, occupational environments, cos- metics and formation within the body.2J Certain N-nitros- amides are also widely used as therapeutic an ticancer agents.4 For these reasons, there is considerable interest in the analysis of trace levels of these compounds in biological and environmental media.Many N-nitrosamines can be analysed, either directly or after derivatisation, by gas chromatography coupled to a chemiluminescence detector and a commercial instrument, the Thermal Energy Analyser Detector (TEA),S has been introduced. The TEA is a modified chemiluminescence detector which relies on thermal cleavage of the N-N bond to produce a nitrogen oxide (NO) radical. The nitrogen oxide is reacted with ozone to produce excited nitrogen dioxide, which emits a photon on decay.6 The photons are detected and amplified by a photomultiplier tube.There are two limitations to this system: first, the N-nitroso compounds must be volatile enough, or made volatile enough for gas chromatography, and secondly, they must yield nitrogen oxide on thermolysis. N-Nitrosamides and related compounds, unlike N-nitros- amines, typically rearrange on thermolysis to yield molecular nitrogen instead of nitrogen oxide and are only weakly detected by the TEA.7 In addition, several N-nitrosamines that are of interest are not suitable for gas chromatography. High-performance liquid chromatography (HPLC) - TEA methods have been reported but mobile phases containing water give inconsistent results.8 Shuker and Tannenbaumg have described a method in which N-nitrosamides were cleaved photolytically by UV irradiation.The resulting nitrogen oxide was oxidised to nitrite, and reacted post-column with Griess reagent to form a chromophore which was detected spectrophotometrically at To whom correspondence should be addressed. t ,V-Nitroso compound abbreviations used are as follows: HNPRO, 3-hydroxy-N-nitrosoproline; NSAR, N-nitrososarcosine; MNU, 1-methyl- 1-nitrosourea; NPRO, N-nitrosoproline; NTHZCA. N-ni- troso-l,3-thiazolidine-4-carboxylic acid; ENU, 1 -ethyl-1-nitrosourea; MNNG. 1-methyl-3-nitro-1-nitrosoguanidine; MNTHZCA, 2-methyl-N-nitroso- 1,3-thiazolidine-4-carboxylic acid; DMNA, dimethylnitrosoamine; TMNU, trimethylnitrosourea; NDPA, ni- trosodipropylamine ; and NPIP, N-nitrosopiperidine.541 nm. The sensitivity was 6-100 ng as injected depending on the specific N-nitroso compound. Fine et al. 10 have modified the pyrolysis chamber in a standard TEA so that N-nitros- amides release nitric oxide during pyrolysis. Sensitivities [signal to noise (SIN) 3 : I] of less than 1 ng injected were reported for standards. Complete details of the instrument were not provided and the need for further development was noted. Singer et al. 11 used dilute acid to cleave N-nitrosamides and coupled the resulting nitrite to Griess reagent. A sensitivity of 50-1000 ng was reported. Sen and Seaman12 have also cleaved the nitroso group chemically from N-nitroso compounds and have detected the nitrogen oxide by chemi- luminescence using a modified TEA. The large dead volume of the system caused considerable peak broadening.Detec- tion of N-nitrosamides by UV,13 mass spectrometry after derivatisationl4 and denitrosation with subsequent detection of the amidel5 have also been proposed. None of these methods satisfied our need for selectivity, sensitivity and simplicity so that we could screen large numbers of samples for N-nitrosamides and non-volatile N-nitrosamines. We therefore undertook the development of an instrument that was capable of selectively detecting N-nitrosamides and N-nitrosamines following separation by reversed-phase HPLC. The instrument was based on the photolysis of N-nitroso compounds (including N-nitros- amidesl6) and the highly selective chemiluminescence detec- tion of the resulting nitrogen oxide.Experimental Standard Compounds? CAUTION: Nitroso compounds are potent animal car- cinogens and must be handled with appropriate care. The N-nitrosoamino acids were synthesised by the method of Lijinsky et a1.17 except for MNTHZCA which was kindly donated by H. Ohshima. The N-nitrosamides, except for TMNU, were obtained from commercial sources and used without further purification; TMNU was synthesised by nitrosating trimethylurea (Alpha) at pH 2.5 in the presence of acetic acid. Analysis by HPLC (254 nm) gave a single peak. Chromatography Both ion-suppression [ 10 mM trifluoroacetic acid (TFA), pH < 21 and ion-pair (2 mM tetrabutylammonium dihydrogen-156 ANALYST, FEBRUARY 1989, VOL. 114 F t Fig. 1. Diagram of the HPLC - chemiluminescence interface.A , He inlet: B. analyser pressure adjustment (He flow-rate); C , effluent from HPLC column; D. 0.125-in stainless steel tubing from interface to traps and chemiluminescence detector: E, electrical connectors t o ballast circuit: F, box fan: G. 200-W lamp; H, glass reactor coil; I , 1.59-mm bulkhead union: J . 0.64-cm bulkhead unions; K, Teflon support towers phosphate, pH = 7) liquid chromatography were used to separate the N-nitroso compounds. Acetonitrile was used as an organic modifier with gradient elution. The flow-rate was 1 ml min-1. Columns were a Brownlee Polymer RP (10 cm x 4.6 mm o.d., 10-ym particle size) or a self-packed Spherisorb ODS (1.5 cm x 4.6 mm i.d., 5-pm particle size, 12% loading). The Beckman HPLC system consisted of two llOB pumps, a 321A controiler, a 210A injector and a 160 UV detector.Photolysis Apparatus The photolysis apparatus is shown in Fig. 1. The effluent of the chromatographic column was transferred to the photolysis coil (H) via a microbore stainless steel tube (C). The end of the stainless steel transfer line was fed into the photolysis coil approximately 3 cm after passing through a 0.64-cm Swagelok bulkhead union (J) and a laboratory-built 0.64 cm x 1 .59 mm reducer. The 0.64-cm union was modified into a T by silver soldering a 1.59 mm o.d. stainless-steel tube into the side in order to mix the carrier gas (A; He) with the column effluent. The photolysis coil (H) consisted of a 3 m X 0.64 cm o.d. X 1 mm i.d. borosilicate glass tube coiled to a diameter of 6 cm and a width of 12 cm.The carrier gas was adjusted using a fine metering valve (B; Nupro, Model M-2MA) to give a total analyser pressure of 1.7-2.0 Torr (with an ozone pressure of 0.9 Torr). A 200-W mercury vapour lamp (G; Model 654A-0100, Canrad-Hanovia, Newark. NJ, powered by a 34245-101 ballasted circuit) was vertically mounted in the centre of the photolysis coil on Teflon towers (K). The entire photolysis apparatus was mounted inside an aluminium box in which a cooling fan (F) had been mounted in the bottom and several air holes had been drilled in the top. The ballast was powered through a thermal safety circuit (Model HTLC1, Cannon Instrument Company, State College, PA) in order to prevent overheating in case of failure of the cooling fan.The temperature of the air leaving the box was 30-40 "C. An optional power attenuator (PAESAR, Lutron, Cooperstown, PA) was used to vary the wattage of the ballasted circuit between 120 and 200 W. The effluent of the photolysis coil (D) was directed to a series of cold traps by 0.32 cm o.d. stainless steel tubing. The first two traps were 100-ml round-bottomed flasks. The inlet tube was located near the neck of the first trap which was cooled by ice. The inlet in the second trap was located midway between the geometric centre and the neck of the flask. The second trap was cooled by a dry ice - acetone bath. In both cases the outlet tube was above the inlet tube. The third trap was a 0.64 cm o.d. stainless steel U-tube cooled to -159 "C with liquid nitrogen. The first and second traps were sealed with tapered Teflon stoppers fitted with O-rings.The nitrogen oxide produced in the photolysis coil was detected using the chemiluminescence reaction chamber of a Model 543 Thermal Energy Analyser (Thermedics, Woburn, MA) by by-passing the pyrolyser and valving portions of the instrument. In some instances, the chromatographic column effluent was first directed through a UV detector (Beckman, Model 160) in series with and ahead of the photolysis coil. Results and Discussion N-Nitrosamines and N-nitrosamides absorb in the regions of 30@380 and 38&430 nm, respectively. Borosilicate glass is transparent in these regions. On photolysis the N-nitroso group is cleaved and nitrogen oxide liberated. 16 The nitrogen oxide is carried to the reaction chamber of the chemilumine\- cence detector by the helium carrier gas and reacted with ozone, thus producing the chemiluminescence reaction. It was necessary to purge the HPLC effluent with helium a s it passed through the photolysis coil so that the nitrogen oxide would be removed from the solution as it was formed.When the HPLC effluent was photolyged prior to being mixed with the helium, only a very weak response was seen, presumably because the nitrogen oxide was oxidised to non-volatile oxides (e.g., NO?). Instrument Design The power of the 200-W lamp was attenuated with an electronic device (PAESAR) in order to determine the effect of various wattages on the response. A wattage of 120 reduced the response for NPRO by about 40% of that of the 200-W lamp.A wattage of 260 (using an attenuated 450-W lamp) did not significantly reduce the response to N-nitrosamines compared to 450 W, but 450 W slightly reduced the response from N-nitrosamides, possibly due to thermal degradation at the higher operating temperature (70-80 "C). The higher wattage lamp gave an unacceptably high base line when urine or gastric juice samples were analysed, especially when non-volatile amine ion-pairing reagents were used in the mobile phase. We believe that the power of the lamp may be analogous to the pyrolyser temperature of the TEA detector in that the selectivity of the instrument is related to the temperature of the pyrolyser. Higher wattage lamps andlor temperatures may produce nitrogen oxide from compounds in addition to those containing the N-nitroso grodp (e.g..C-nitroso and nitro compounds). We investigated several sizes and types of Tefloh [tetrafluo- roethylene (TFE), perfluoroalkoyl (PFA), and fluorinated ethylene polypropylene (FEP)] microbore tubing for irradia-ANALYST, FEBRUARY 1989, VOL. 114 157 tion coils with wall thicknesses of 0.15-0.41 mni. The Teflon resulted in a loss of response after only 2-3 weeks of use with the higher wattage lamp. The response could not be recovered by washing the tubing and could only be recovered by replacing the tubing with new material. Others have also noted that Teflon can be degraded under intense UV irradiation.18 These problems were overcome by the use of thick-walled glass capillary tubing. Glass tubes of lengths between 0.6 and 3.6 m were tested.Lengths of less than 3 m gave a decreased response while longer tube lengths did not result in an increased response. The inside diameter of the tubing was important for chromatographic resolution. Inside diameters of >2 mm required a solvent flow-rate of at least 2-5 ml min-1 in order to produce acceptable peak shapes. Tubing of 1 mm i.d. gave good peak shapes at a flow-rate of 1 ml min-1. The flow of solvent had to be in the direction of gravity or the resulting solvent bumping within the coil resulted in pressure surges in the analyser and an erratic base line. Fig. 2 compares the chromatograms of 2.4-14 ng each of an eight-component standard that was routed first through a UV I '1 x 2 i 1 I I I I I 0 2 4 6 8 t/min Fig.2. Chromatogram obtained by routing the effluent of the HPLC column first through a 254-nm UV detector, then to the photolysis inlet. Compounds are: 1, HNPRO; 2, NSAR; 3, MNU; 4, NPRO: 5 , NTHZCA; 6. €NU; 7, MNNG; 8. MNTHZCA. (A) Chemilumlnes- cence detection and (B) 254-nm UV detector placed in series. Column. Brownlee Polvmer RP; mobile phase, 10 r n M TFA contain- ing MeCN. Gradient elution: &7% MeCN at 1% min-I, hold 5 min, then 7-22% MeCN at 1.5% rnin-1, hold 3 min; 20-yl injection. 2.4-14 ng component; lamp power. 260 W detector, then the photolysis detector. The UV detector was not sensitive enough to produce acceptable peaks. However, the photolysis - chemiluminescence detector showed all peaks to be eluted within 8 min with peak widths of 3&60 s.The configuration of the glass coil is not an optimum for efficiency; however, the fact that the system was operated at 1-2 Torr and that the ultimate analyte was volatile made up for the lack of efficiency. Comparison of peak widths at half height from the UV and chemiluminescence detectors indicated that the loss of efficiency was less than 1&20%. The maximum sensitivity (SIN > 3 : 1) of the chemilumines- cence detector was less than 1 ng for N-nitrosoamino acids (Fig. 2). N-Nitrosamides were somewhat less responsive, with limits of detection of approximately 10 ng. These limits are higher than those for volatile N-nitrosamines, which typically have thresholds of detection of <0.5 ng injected when analysed by gas chromatography - TEA. The molar response ratios for several N-nitrosoamino acids, N-nitrosoureas and N-nitrosamines are given in Table 1, Standard solutions (ca. 0.1 mmol) in acetonitrile were injected (10 pl) into the photolysis coil after passing through a 20.3 cm X 1.6 mm o.d.stainless-steel wide-bore tube, in order to mix the sample with the mobile phase. The mobile phase composition was varied from 100% aqueous to 100% organic modifier and the lamp was set at the lowest power (120 W). At a solvent concentration of MeCN + 10 mM TFA (90 + lo), N-nitrosoamino acids had the highest molar response, while N-nitrosamides and N-nitrosamines typically had reduced relative molar responses (DMNA or MNU, 9%; TMNU, 20%; MNNG, 40%). Increasing the aqueous portion in the mobile phase to 90% reduced the response of the N-nitro- soamino acids by an order of magnitude, while the response of the other N-nitroso compounds decreased less dramatically. Table 1 shows the relative molar response ratios for all the compounds tested, in mobile phase concentrations that are practical for ion-suppression chromatography.Addition of TFA (10 mM) to the aqueous portion did not have any effect on relative response; however, it improved flow through the coil when 10Oo/0 H 2 0 was used. The sensitivity of the system, coupled to the 10-100-fold increase in injection volume of HPLC compared to GC, results in an over-all sensitivity for HPLC that is greater than that of GC - TEA. The response was linear (6 > 0.99) between 1 and 100 ng of NPRO injected. Above 100 ng the response fell off.Repeated injections of NPRO gave a coefficient of variation of 4%. The Brownlee Polymer RP column gave sufficient resolu- tion for quantifying N-nitrosoamino acids in the ion-suppres- sion mode but did not resolve the syn- and anti-isomers.17 The ODS column in the ion-pair mode separated syn- and anti-isomers but produced unacceptable base-line noise at higher lamp wattage. The ion-pair reagent also resulted in a build-up of a residue on the inner surface of the glass coil. N-Nitrosoamino Acids in Urine N-Nitrosoamino acids are components of mammalian urine at the 2-300 pg 1-1 level, and have been used as indicators of the endogenous formation of N-nitroso compounds. 19\20 The analysis of large numbers of samples for these compounds has Table 1. Relative molar response ratios* N-Nitroso compound Solvent ?/oMeCN'i NPRO HNPRO NSAR NPTP NTHZCA MNTWZCA TMNU MNU ENU MNNG DMNA NDPA - 0 1 .0 1.1 - 1 .o 1 .s - 1 .o 0.8 - 1.6 0.5 1 0 1.4 1.4 1.1 1.1 1.7 1.4 1.2 0.9 0.5 1.7 0.6 0.6 25 1.7 1.7 1.2 1.4 2.5 2.1 1.2 1 .o 0.7 2.1 0.6 0.6 * NPRO (0% MeCN) = 1.'i Balance, 10 mM TFA or water158 vi ANALYST, FEBRUARY 1989, VOL. 114 Y L 3 1 u- 0 4 8 12 0 4 8 12 t min Fig. 3. spiked human urine [7.5 and 8.0 ug 1-1 o f (1) NPRO and (21 NTHZCA respectively]. Column and mobile phase as in Fig. 2; gradient elution: 5% MeCN, hold 2 min, increase to 20% MeCN at 3% min-1: lamp power. 120 W A 2000-pl injection o f (A) unspiked human urine and (B NTHZCA I I I I I 0 4 8 12 16 t min Fig. 4. Chromatogram of 10 ml o f (A) human urine that had been spiked with 7-8 pg 1 - 1 of NPRO and NTHZCA, and (B) a standard solution of 0.15-0.16 pg ml-1 of NPRO and NTHZCA.Column and mobile phase as in Fig. 2; gradient elution: S-20% MeCN at 7.5% min 1; 100-pl injection; lamp powcr, 120 W been undertaken in several biochemical epidemiology studies. The current methodology is to extract and methylate the N-nitrosoamino acids to make them amenable to gas chromat- ography (GC) - TEA. The selectivity of the detector suggested that N-nitrosoamino acids could be analysed directly without concentration. A 2000-p1 injection of spiked (ca. 8 pg 1-1 of NPRO or NTHZCA) urine, which had been passed through a CIS solid-phase extraction (SPE) high-capacity cartridge (Baker). gave a significant response (Fig. 3). Fig. 4 is a chromatogram of an extract from human urine that had been spiked with NPRO and NTHZCA, at 7.5 and 8.0 ug 1-1, respectively.Urine (10 ml) at pH 6-7 was passed through a CI8 SPE high-capacity cartridge that had been 0 4 8 0 4 8 timin Fig. 5. Chromatogram o f a standard solution of TMNU (A: 17 ng injected) and an extract o f (B) porcine gastric fluid that had been spiked with 42 pg 1- 1 of TMNU. Column as in Fig. 2. Isocratic elution with MeCN + 10 mM TFA (15 + 85); 20-pl injection; lamp power. 120 w pre-wetted with MeOH and H20. The cartridge was washed with two volumes of water and the combined eluent acidified (1% H2NS03NH, in 1.8 M H2S04) to pH 1, extracted with ethyl acetate (6 x 25 ml). washed with saturated NaCl and dried over Na2S04. The ethyl acetate layer was vacuum concentrated, blown to dryness with N2, redissolved in 10 mM TFA (0.50 ml) and filtered through glass-wool.The sample was injected in the HPLC column (loo-@ sample loop) and eluted with MeCN + 10 mM TFA ( 5 + 95), with a gradient (2% min-1) to MeCN + 10 mM TFA (20 + 80). The apparent low recovery of NTHZCA (approximately 60%) was not investigated but was probably due to the known instability of this compound. This procedure allows the analyses to be carried out without the need for derivatisation. N-Nitrosamides in Gastric Fluid The endogenous formation of N-nitrosamides has been suggested as a factor in the etiology of human gastric cancer.” N-Nitrosamides are also used therapeutically as antitumour agents. They are not amenable to TEA detection due to thermal rearrangement to liberate N2.Krull et a1.13 have investigated HPLC - UV as a method for N-nitrosoureas in blood. The limit of detection was 100-200 pg 1-1 and, in blood cells washed with saline, unknown peaks interfered. Fig. 5 shows a chromatogram of an extract of porcine gastric fluid which had been spiked at 42 pg 1-1 with TMNU. This compound was used because it is relatively stable compared to other N-nitrosamides. Centrifuged gastric fluid was acidified (HCl) to pH 3-3.5, centrifuged and passed through a pre-wetted (CH30H, H20, 2 mM TFA) ClS SPE high-capacity cartridge. The cartridge was rinsed with 3 ml of 2 mM TFA, vacuum dried and eluted with 5 ml of ethyl acetate. The ethyl acetate was dried (Na2S0,), reduced to dryness (N2) and the residue suspended in 0.50 ml of 2 mM TFA.Recovery of the spike was 84% with a limit of detection of approximately 2-5 pg 1-1. This could be improved easily by increasing the sample size and/or the 20-1-11 injection volume. We are presently using this detector to investigate the occurrence, formation and stability of non-volatile N-nitroso compounds, particularly N-nitrosamides and related com- pounds in a variety of biological fluids and food samples. This work was supported by the US National Cancer Institute under Grant No. RO1-CA40833-03. A preliminary report of these results was presented at the Advances in the Biology and Chemistry of N-Nitroso Compounds Meeting in Omaha, NE, May 20th, 1988. References 1. Preussmann, R., and Stewart, B. W., in Searle, C. W., Edifor, ‘Chemical Carcinogens,” Second Edition, American Chemical Society Monograph No.182, American Chemical Society, Washington, DC. 1984, Chapter 12.ANALYST, FEBRUARY 1989, VOL. 114 159 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Preussmann, R . , and Eisenbrand, G., in Searle, C. W., Editor, “Chemical Carcinogens,” Second Edition, American Chemical Society Monograph No. 182, American Chemical Society, Washington, DC, 1984, Chapter 13. Hotchkiss, J. H., Adv. Food Res., 1987, 31, 54. Prestayko. A. D., “Nitrosoureas, Current Status and New Developments,” Academic Press, New York, 1981. Hotchkiss, J . H., J. Assoc. Off. Anal. Chem., 1981, 64, 1037. Fine, D . H . , Lieb, D., and Rufeh, F., J . Chromatogr., 1975, 107, 351. Berry, C. N.. Challis. B. C . , Gribble, A. D . , andjones, S.P., in Scanlan. R. A.. and Tannenbaum, S . R . , Editors, ‘-N- Nitroso Compounds.” American Chemical Society Symposium Series No. 174, American Chemical Society, Washington. DC, Sen. N. P . . and Kubacki, S . J.. Food Additives Corztamin., 1987, 4. 357. Shuker. D. E . G., andTannenbaum, S . R . , Anal. Chem., 1983, 55, 2152. Fine, D. H., Goff, E . U . , Johnson, R. A , , and Rounbehler, D. P., in Bartsch, H.. O’Neill, I., and Schulte-Hermann, R., Editorr, “The Relevance of N-Nitroso Compounds to Human Cancer,” International Agency For Research on Cancer, Lyon, 1987, Publication No. 84, p. 216. Singer, G. M., Singer, S . S . , and Schmidt, D. G., J.Chro- rnatogr.. 1977, 133, 59. 1981, p. 101. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Sen, N. P., and Seaman, S. S . , in O’Neill, I. K., Von Borstel, R. C., Miller, C. T., Long, J., and Bartsch, H . , Editors, “N-Nitroso Compounds: Occurrence, Biological Effects and Relevance to Human Cancer,” International Agency for Research on Cancer. Lyon, 1984. Publication No. 57, p. 137. Krull, I. S . , Strauss, J . , Hochberg, F . , and Zervas, N. T., J . Anal. Toxicol.. 1981, 5 , 42. Weinkam, R. J . , Wen, J . H . C., Furst, D. E . , and Levin, V. A . , Clin. Chem., 1978, 24, 45. Mirvish, S. S . , Karlowski. K., Cairnes, D. A , , Sams, J . P., Abraham, R., and Nielsen, J., J . Agric. Food Chern.. 1980,28, 1175. Chow, Y. L., in Anselme, J.-P., Edztor, “N-Nitrosamines,” American Chemical Society Symposium Series No. 101, American Chemical Society, Washington, DC, 1979, p. 18. Lijinsky, W., Keefer, L., and Loo, J . , Tetrahedron, 1970, 26, 5137. Batley, G. E . , Anal. Chem., 1984. 56, 2261. Ohshima, H.. O’Ncill. I. K., Friesen. M., Berezait. J.-C.. and Bartsch. H., J . Cancer Res. Clzn. Oncol.. 1984, 108, 121. Ohshima, H . , and Bartsch, H., Cancer Res., 1981, 41, 3658. Mirvish, S. S . , J . Nut. Cancer Inrt., 1983. 71, 630. Paper 81034721 Received August 30th, I988 Accepted October 24th, I988

ISSN:0003-2654    

DOI:10.1039/AN9891400155

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST. FEBRUARY 1989, VOL. 114 161 Stability of Bis(2,4,6-trichlorophenyl) Qxalate in Hig h-performance Liquid Chromatography for Chemiluminescence Detection Noriko Imaizumi, Kazuichi Hayakawa and Motoichi Miyazaki Faculty of Pharmaceutical Sciences, Kanaza wa University, Takara-machi 73- 7, Kanazawa 920, Japan Kazuhiro lmai Branch Hospital Pharmacy, University of Tokyo, 3-28-6 Mejirodai, Bunkyo-ku, Tokyo 7 72, Japan The effects of various factors (organic solvent, type of reservoir, temperature, water content and concentration of hydrogen peroxide) were investigated on the stability of a chemilumigenic reagent, bis(2,4,6-trichlorophenyl) oxalate (TCPO), for high-performance liquid chromatography. In a mixture with hydrogen peroxide, TCPO was most stable when dissolved in acetonitrile and kept in a polyethylene bottle or a high quality glass (borosilicate glass free from metals) flask.The stability of TCPO was also increased by lowering the temperature, the water content and the concentration of hydrogen peroxide in the chemilumigenic reagent solution. Under these conditions, TCPO can be used in a mixed acetonitrile solution with hydrogen peroxide for HPLC - chemiluminescence detection. Keywords : Bis (2,4,6- trichlo rop h en yl) oxala te; hydrogen peroxide; ch em ilu rninescence detection; high-performance liquid chromatography; post-column reaction system High-performance liquid chromatography (HPLC) with chemilurninescence detection has attracted much attention because it has the facility to determine several compounds more sensitively and more selectively than fluorescence detection.The combination of bis(2,4,6-trichlorophenyl) oxa- late (TCPO) and hydrogen peroxide has been the most popular post-column chemilumigenic reagent. However, aryl oxalates decompose on the addition of hydrogen peroxide and therefore an extra pump is required for each solution in an HPLC - chemiluminescence detection system. If the stability of a mixture of TCPO and hydrogen peroxide was increased then a conventional one-pump reaction system could be used for the chemiluminescence detection. In this paper, several factors affecting the stability of TCPO mixed with hydrogen peroxide have been examined to obtain the conditions suitable for a one-pump chemiluminescence detec- tion system in HPLC. Experimental Chemicals Analytical-reagent grade TCPO was purchased from Tokyo Kasei Industry (Tokyo, Japan) and used as received.Elec- tronic grade hydrogen peroxide (30%) was obtained from Kanto Chemicals (Tokyo, Japan). Analytical-reagent grade acetonitrile and imidazole were obtained from Nakarai Chemicals (Kyoto, Japan). Spectrograde methanol and ethyl acetate were obtained from Nakarai Chemicals and Dojindo Laboratories (Kumamoto, Japan), respectively and spectro- grade acetone from Nakarai Chemicals and Dojindo Labora- tories. Analytical-reagent grade nitric acid was obtained from Wako Pure Chemical (Osaka, Japan). Distilled water was used after de-ionisation and filtration with a Millipore (Bedford. MA, USA) Milli-Q I1 water purification system. Preparation of Test Solution Vessels used in the investigation of the stability of TCPO were Erlenmeyer flasks (100-ml inner volume) of silica and Hario (Shibata Scientific Technology, Tokyo, Japan) glass, reagent bottles (120 ml) of clear and amber borosilicate glass, white and amber polyethylene bottles (100 ml, Nikko, Osaka, Japan) and an amber soda-lime glass bottle (80 ml).In other studies. centrifuge tubes (10 ml) made of clear borosilicate glass with stoppers were used for storing the TCPO solutions. The TCPO was dissolved in various solvents to give a concentration of 0.5 mM. An aliquot (SO ml) of the solution was transferred into each flask or bottle to study the effect of type of vessel and another aliquot ( 5 ml) to each centrifuge tube for other studies. Various volumes of hydrogen peroxide were added to the TCPO solution following the addition of solvents to give the same incubation mixture volume.The incubation was started by the addition of hydrogen peroxide. The residual TCPO in the test solution was determined periodically by HPLC using the peak-height method. HPLC - Chemiluminescence Detection System Chromatography was carried out with two Shimadzu (Kyoto, Japan) LC-6A pumps, a Rheodyne (Cotati, CA, USA) Model 7125 loop injector (20 pl), an Atto (Tokyo, Japan) AC-2220 chemiluminescence monitor, and, after the column, a Shi- madzu C-R1B integrator, a Kyowa (Tokyo, Japan) KZS-1 mixing device and a stainless-steel tube (0.25 mm i.d. X 300 mm) as a reaction coil. The separation column used was a Nagel (Duran, FRG) Nucleosil ODS (4.6 mm i.d.x 250 mm) and the guard column used was a Brownlee (Berkley, CA, USA) RP-18 (4.6 mm i.d. x 30 mm). Acetonitrile - water (9 + 1) containing 1 mM of imidazole - nitrate buffer (pH 7.0) was used as an eluent. The chemilumigenic reagent was prepared by dissolving 0.5 mM TCPO and 1.50 mM hydrogen peroxide in various solvents just before use and kept in a polyethylene bottle in an ice - water bath. The flow-rate of the eluent and chemilumigenic reagent was 1 ml min-1. HPLC and Conditions for the Determination of TCPO The HPLC system consisted of a Jasco (Tokyo, Japan) BIP-I pump, a Jasco UVTDEC-100-111 UV absorbance detector, a Shimadzu C-R1B integrator and a Rheodyne Model 7125 loop injector (20 pl). The TCPO was separated by a Tosoh (Tokyo, Japan) TSK gel ODs-120T column (4.6 mm i.d.X 150 mm) with acetonitrile - water (95 + 5 ) as eluent at a flow-rate of 1 ml min-i. The monitoring wavelength was fixed at 280 nm. Results and Discussion In order to investigate the stability of TCPO in the presence of hydrogen peroxide, the residual TCPO was determined using reversed-phase HPLC with UV detection (280 nm). Fig. 1 shows the typical chromatograms of TCPO, hydrogen perox-162 ANALYST, FEBRUARY 1989, VOL. 114 0 2 4 6 8 0 2 4 6 0 2 4 6 8 Timelmi n Fig. 1. Chromatograms of ( u ) 0.5 mM TCPO in acetonitrile, ( b ) 150 mM hydrogen peroxide in acetonitrile and ( c ) a mixture of 0.5 mM TCPO and 150 mhf hydrogen peroxide in acetonitrile 1 d after mixing. Detection wavelength, 280 nm 1 B 0 i- 0 2 4 6 8 Timelh Fig.2. Effects of organic solvents on the stability of TCPO in the presence of hydrogen peroxide. Final concentrations of TCPO and hydrogen peroxide were 0.49 and 170 mM. respectively. Solvent: A , acetonitrile; 0, acetone - acetonitrile (1 + 1); 0, ethyl acetate - acetonitrile (1 + 1); and A , methanol Table 1. The effects of three organic solvent \ystem\ on the chemiluminescence intensity of benzo[a]pyrene by HPLC" Peak area o f 5 X 1 0 -' M benzo[a]pyrene/pV s Solvent system After 0.5 h After 4 h After 8 h Acetonitrile . . 245 000 (1 .OO)t 235 000 (0.96) 224 000 (0.91) Acetone - ace t o n it ri le Ethyl acetate - ace to n it r i 1 e * Chemilumigenic reagent contained 0.5 mM TCPO and 150 mhi hydrogen peroxide. Other HPLC conditions are as described in the text. t Value relative to the peak area after 0.5 h in each solvent system.$ Too small to be calculated. (1 + 1) . . . . <2500$ <2500 <2500 ( 1 + 1) . . . . 233000(1.00) 236000(0.97) 221000(0.91) ide and a mixture of the two. The TCPO and hydrogen peroxide were eluted at 6.2 and 1.9 min, respectively. In the chromatogram of the mixture, another peak was observed having a retention time of 2.5 min. The peak was tentatively identified as 2,4,6-trichlorophenol (TCP), a hydrolysed pro- duct of TCPO, by the agreement of its retention time with that of authentic TCP. The effects of organic solvents on the stability of TCPO in the presence of hydrogen peroxide and on the chemilumines- cence intensity were examined. Methanol, acetonitrile, ace- 100 80 8 0 6 60 I- m 3 - 40 a, aT: 20 0 I I I I 0 2 4 6 Timeih Fig.3. Effect of temperature on the stability of TCPO in acetonitrile with hydrogen peroxide. Concentrations of TCPO and hydrogen peroxide in acetonitrile were 0.48 and 340 mM, respectively. Tcm- perature: @, 0 ; A , 20; and +. 40 "C Table 2. Effect of type of vessel on the stability of TCPO in the presence of hydrogen peroxide* Residual TCPO, Yu Vessel Time after mixing with hydrogen peroxide/h 1 6 24 Silicaflask . . . . 90.8 69.2 51.5 Harioflask . . . . 91.4 69.5 49.2 Clear borosilicate Amber borosilicate Amber soda-lime White polyethylene Amber polyethylene glassbottle . . 87.8 65.1 39.7 glassbottlc . . 84.9 66.0 43.9 glassbottle . . 54.2 t t bottle . . . . 91.3 72.9 55.4 bottle . . . . 93.2 74.4 54 8 * The test solutions were prepared by mixing 50 ml of an acetonitrile solution of 0.5 mM TCPO and 2 ml of 30% hydrogen peroxide.The final concentrations of TCPO and hydrogen peroxide were 0.48 mbi and 340 mM, respectively. f Not detected. tone and ethyl acetate were tested because the former two are popular a5 mobile phases for HPLC and the latter two are generally used as solvents for TCPO and hydrogen peroxide. The stability of TCPO in four solvent systems (acetonitrile, acetone - acetonitrile, ethyl acetate - acetonitrile and methanol) is shown in Fig. 2. In methanol, the decomposition of TCPO was significant but in the other solvent systems TCPO was much more stable. Therefore those solvent systems, except methanol, were applied to the HPLC - chemi- luminescence detection system. The peak areas of benzo[a]- pyrene thus obtained are summarised in Table 1.The peak areas corresponding to chemiluminescence intensities with both acetonitrile and ethyl acetate - acetonitrile were signifi- cantly greater than that with acetone - acetonitrile. After 8 h, 91% of the intensities obtained after 0.5 h with both solvent systems remained. Even though TCPO remained stable in acetone - acetonitrile longest, it was not suitable for chemilu- minescence detection because it gave a smaller peak and higher background than the other two systems. In spite of the high solubility of TCPO in ethyl acetate, the ethyl acetate - acetonitrile system increased the base-line noise owing to the immiscibility of the solutions when the water content in the eluent was high. Consequently, acetonitrile was used as the most suitable solvent for the chemilumigenic reagent.The effect of temperature on the stability of the TCPO -ANALYST, FEBRUARY 1989, VOL. 114 100 80 8 0 6 0 - I- - m 3 -0 'Z 40- [r 20 163 - - - 0- I I 0 2 4 6 Timeih Fig. 4. Effect of the concentration of hydrogen peroxide on the stability of TCPO in acetonitrile. Concentration of TCPO in the acetonitrile mixture, 0.45 mM. Concentration of hydrogen peroxide: C, 0; A. 170; 0, 340; and +. 680 mM I I 1 0 1 2 Tirneih Fig. 5. Effect of water on the stability of 'TCPO in acetonitrile. Solvent of 0.3 msi TCPO: 0, acetonitrile; and A, acetonitrile - water (9 + I ) hydrogen peroxide mixture is shown in Fig. 3. The TCPO was more stable at the lower temperatures; the residual TCPO in a Hario flask at 40°C was 64% after 6 h, whereas at 0°C the residual was 720%. The effects of the type of vessel on the stability of TCPO - hydrogen peroxide were compared, at a higher concentration of' hydrogen peroxide than that in Table 1 to increase the effects.Table 2 shows that the mixture was more stable in a polyethylene bottle than in any of the glass bottles tested. The residual TCPO in an amber polyethylene bottle was 74% after 6 h at room temperature (ca. 15 "C). No significant difference was observed between white and amber polyethylene bottles. On the contrary, the quality of the glass did affect the stability. The stability of TCPO was lowest in an amber soda-lime glass bottle; the degradation of TCPO in a soda-lime glass bottle has been reported previously in the absence of hydrogen peroxide.3 The TCPO was more stable in a silica or Hario glass flask than in a clear or amber borosilicate glass bottle and the residual TCPO in a silica or Hario glass flask was close to those in the polyethylene bottles.Therefore in the experiments on the chemiluminescence detection in HPLC a polyethylene bottle was used as a reservoir for the chemilumigenic reagent. The chemilurninescence intensity of benzo[a]pyrene in HPLC with TCPO - hydrogen peroxide increased with increasing concentration of hydrogen peroxide. However. the detection limits for analytes, defined as the signal to noise ratio, were not improved because of the increase in back- ground noise level. In addition, the decomposition of TCPO A I I 1 I 0 5 10 15 20 Timeimi n Fig.6. Typical chrornatogram of (A) benzo[a]pyrene (2.5 ng) and (B) benzo[ghi]perylene (12 ng) by chemiluminescence detection. They were injected as the mixed acetonitrile solution of the two. Other conditions are as described in the text 25 20 v) $ 1 5 2 $ 10 Y a 5 0 I I 4 8 Timeih 2 0 m [r ._ + 1 0 Fig. 7. Time courses of peak intensities of benzo[a]pyrene and benzo ghi]perylcne by HPLC - cherniluminesccnce detection. @, benzo[a]pyrene (2.5 ng); A, benzo[ghi]pelyrcne (12 n ); 0. ratio of the peak area of benzo[a]pyrcne to that of benrofghi]perylene. Conditions are the same as in Fig. 6 was accelerated with increasing concentrations of hydrogen peroxide in the mixture, as shown in Fig. 4. The TCPO was also hydrolysed rapidly with water, as shown in Fig.5 , which might result in a decrease in chemiluminescence intensity.8 Hydrogen peroxide is commer- cially available as a 30% water solution, and hence an increase in hydrogen peroxide concentration is accompanied by an increase in water concentration. Therefore, the concentration of hydrogen peroxide was set at < 150 mM, which gave enough chemiluminescence intensity with a large signal to noise ratio for HPLC, and the mixture was prepared just before use to prevent decomposition of TCPO. According to the above data, the post-column chemilu- minescence detection could be performed on a one-pump system by using an acetonitrile mixture of TCPO and hydrogen peroxide under the following conditions: the aceto- nitrile used for the solvent for both TCPO and hydrogen peroxide should not contain water; the concentration of hydrogen peroxide in the reagent mixture should be less than 170 mM vs.0.5 mM TCPO; and the mixture should be prepared just before use and stored in a suitable vessel such as a polyethylene bottle or a silica glass bottle at as low a temperature as possible during the HPLC operation. By chemiluminescence detection under these conditions, HPLC analysis of the two polycyclic aromatic hydrocarbons benzo[a]pyrene and benzo[ghi]perylene was performed. Benzo[a]pyrene and benzo[ghi]perylene were eluted at 13.3164 ANALYST, FEBRUARY 1989, VOL. 114 and 18.3 min, respectively (Fig. 6). The detection limits for benzo[a]pyrene and benzo[ghi]perylene, calculated from their peak heights (signal to noise ratio = 3), were 15 and 95 pg, respectively. These values were significantly lower than those reported using a two-pump system and more concentrated TCPO and hydrogen peroxide dissolved in acetone.’ In an 8-h period following preparation of the chemilumigenic reagent, the peak areas of the two decreased slightly with time, while the ratio of peak area of benzo[a]pyrene to that of benzo- [ghilperylene did not change (Fig.7). Because the decrease in cheniiluminescence intensity was less than 10% over 8 h a one-pump system could be used for post-column chemilu- minescence detection in HPLC. Moreover, the reliability would be increased by the use of an internal standard. References 1. Imai. K.. in DeLuca, M., and McElroy, W. D., Editors, “Methods in Enzymology,” Volume 133, Academic Press, London, 1986, p. 435. 2. Kobayashi, S . , and Imai, K., Anal. Chem.. 1980, 52, 424. 3. Sigvardson, K. W . , and Birks, J . W . , Anal. Chem., 1983, 55, 432. 4. Sigvardson, K. W., and Birks, .I. W., Anal. Chem., 1984, 56. 1096. 5 . Miyaguchi, K., Honda, K., and Imai, K., J . Chromatogr., 1984, 303, 173. 6. Miyaguchi, K., Honda, K., and Imai, K., J . Chromatogr., 1984, 316, 501. 7. Grayesky, M. L., and DeVasto, J . K., Anal. Chem.. 1987.59, 1203. 8. van Zooncn, P . , Kamminga, D. A , , Gooijer. C., Velthorst, N. H.. Frei. R . W., and Gubitz, G . , Anal. Chim. Acta, 1985, 174. 151. Paper 8102878H Received July 18th, 1988 Accepted October 5th, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400161

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1989, VOL. 114 165 Ion Chromatographic Determination of Aluminium With Ultraviolet Spect rop hot0 met ric Detection John R. Dean* Department of Chemical and Physical Sciences, The Polytechnic, Queensgate, Huddersfield HD 1 3DH, UK A method for the determination of aluminium by high-performance ion-exchange chromatography followed by post-column derivatisation with 4,5-dihydroxy-I ,3-benzenedisulphonic acid, disodium salt and ultraviolet detection at 310 nm is described. The calibration graph was linear from 10 to 5000 pg I-' of Al with a detection limit (30) of 7 pg 1-1. Typical relative standard deviations in the range 1.3-3.9% were obtained for a 100-pg 1-1 Al standard. Freedom from interference from six cations investigated is reported, the exception being titanium.The method was applied successfully to the determination of aluminium in certified reference materials after conventional wet digestion (NBS SRM 1572 and NIES No. 7). The incomplete digestion of biological materials in a microwave oven is reported. The problems of using a non-destructive analytical technique for the determination of aluminium in a natural riverine water reference material are discussed. Keywords: High-performance ion chromatography; aluminium determination; ultraviolet detection; biolog- ical reference materials; water analysis The recent re-emergence of interest in aluminium in the water supply and foodstuffs'-6 and its connection with organic mental impairment in the elderly7.8 has demonstrated the need to obtain precise and accurate information on aluminium rapidly and cost effectively.One technique that has the potential to fulfil this need is ion chromatography.9.10 Various systems for ion chromatographic separation and post-column derivatisation with fluorescence11 and ultravioletl2.13 detec- tion have been discussed. This paper describes the use of a high-performance ion-exchange column followed by post- column derivatisation using 4,5-dihydroxy-l,3-benzenedisul- phonic acid, disodium salt (Tiron)'4.15 and ultraviolet detec- tion to determine aluminium in the sub-pg ml-1 range. Experimental conditions and analytical figures of merit including linear dynamic range, peak dispersion and interfer- ences are discussed. Results for the determination of alu- minium in biological and environmental certified reference materials are also presented.Experimental Apparatus The instrument used was a Dionex Model 201 Oi ion chromato- graphy system (Dionex, Camberley, UK) equipped with an ultraviolet detector (Pye Unicam, Cambridge, UK), moni- tored at 310 nm. The ultraviolet absorption data were obtained using a Beckman DU-50 spectrophotometer (Beck- man-RIIC. High Wycombe, UK). Operating Conditions Mobile phase. 0.2 M (NH4)2S04 adjusted to pH 2.8 with concentrated sulphuric acid. Post-column reagent. 3 X 10-4 M Tiron in 3 M ammonium acetate adjusted to pH 6.2 with concentrated sulphuric acid. Guard column. HPIC-CG2. Analytical column. HPIC-CS2. Flow-rate. Mobile phase 1.0 ml min-1, post-column reagent (56 Ib in-2) 0.5 ml min-1. Combined flow-rate 1.5 ml min-1.Injection loop. 100 p1. Present address: Department of Chemical and Life Sciences, Newcastle upon Tyne Polytechnic, Ellison Terrace, Ellison Place, Newcastle upon Tyne NE1 8ST, UK. Reagents All samples and the mobile phase were prepared from analytical-reagent grade materials. De-ionised, distilled water was obtained from a Water-1 system (Jencons Scientific, Leighton Buzzard, UK). Aluminium wire (AnalaR, BDH, Poole, UK) was dissolved in hydrochloric acid (AristaR) to provide a stock solution containing 1000 1.18 ml-1 of Al. Working standards were prepared by serial dilution daily and nitric acid (AristaR) was added so that the pH of the final solutions was 1.6. All standards and samples were prepared in polypropylene flasks and beakers that had previously been soaked for at least 48 h in 20% nitric acid and rinsed four times with de-ionised, distilled water. For the interference studies the following salts were used: potassium titanium oxalate, indium oxide, manganese chloride, molybdenum trioxide, iron(I1) sulphate, ammonium iron( 111) sulphate and zinc sulphate.The National Bureau of Standards (NBS) SRM 1572 (Citrus Leaves) was obtained from the NBS, Washington, DC, USA; Tea Leaves National Institute for Environmental Studies (NIES) No. 7 from NIES, Tsukuba Ibaraki, Japan; and Riverine Water reference material, SLRS-1, from the National Research Council of Canada (NRCC), Ottawa, Canada. Digestion of Certified Reference Materials Wet digestion Samples (0.2 g) of NBS SRM 1572 and NIES No. 7 were weighed accurately into pre-cleaned tubes (approximate volume 250 ml) and heated with concentrated nitric and sulphuric acids on a DG-1 digestion block using a PTC-2 programmable temperature controller (Techne Cambridge, Cambridge, UK).Duplicate samples, spiked (10 pg of A1 for SRM 1572 and 100 pg of A1 for NIES No. 7) samples and sample blanks were prepared. The final solutions contained 4% sulphuric acid. Calibration standards were prepared to contain the same acid content. Microwave digestion Samples (0.2 g) of NBS SRM 1572 were weighed accurately into pre-cleaned, sealed PTFE digestion bombs, volume 100 ml (Cowie Scientific, Middlesbrough, UK) and 3 ml of concentrated nitric acid and 1 ml of concentrated H2S04 added. The bombs were heated singularly in a commercial166 ANALYST, FEBRUARY 1989, VOL.114 microwave oven (Toshiba electronic oven, frequency 2450 MHz, power 1.17 kW and single phase) for 5 min and left sealed overnight. The clear solutions produced contained 6% nitric acid and 2% sulphuric acid. Duplicate samples, spiked (10 pg of Al) samples and sample blanks were prepared. Calibration standards were prepared to contain the same acid content. Analysis of Riverine Water Reference Material The reference material had been filtered (0.2 pm) previously and acidified with nitric acid to pH 1.6 in its preparation. A calibration graph was prepared from 0 to 50 pg 1-1 in acidified solution (pH 1.6) and the reference material analysed by concentration calibration. Results and Discussion Complexing Reagent Tiron forms complexes with a variety of metal ions'" and has been shown to be suitable for aluminium determination.17 It is known that Al3+ forms six-coordinate compounds.5 There- fore, it is suggested that Al3+ combines with the two hydroxy sites of three Tiron molecules. Detection Conditions Studies on the ultraviolet absorption of Tiron in 3 M ammonium acetate were carried out at several pH with and without the addition of aluminium. It was observed (Fig. 1) that the addition of 1 and 100 pg ml-1 of A1 produced bathochromic shifts of 7 and 14 nm, respectively, at pH 6.2. Similar results were observed with aluminium at solution pH of 7.6, 7.0. 5 . 5 and 5.0. The wavelength range that produced the maximum absorbance upon the addition of aluminium was 305-309 nm. Therefore, 310 nm was confirmed as the optimum wavelength for aluminium determination.17 The post-column derivatising reagent pH was found not to be detrimental to aluminium absorption within the range 5.5-7.6. Separation Conditions Mobile phase Initial studies were carried out using the conditions described under Experimental. The effect of mobile phase flow-rate on the aluminium response while keeping the post-column reactor pressure constant at 56 lb in-2, corresponding to a flow-rate of 0.5 ml min-1, is shown in Fig. 2. It was observed that the maximum response was obtained with a total combined flow-rate of 1.55-1.63 ml min- 1 (experimentally determined after UV detection). The effect of varying the post-column reactor pressure on the analytical response for 100 pg 1-1 of Al while keeping the mobile phase flow-rate constant at 1.0 ml min-1 is shown in Fig.3. It can be seen that provided a minimum combined total flow-rate of 1.25 ml min-1 (experimentally determined after UV detection) was present. no difference in the analytical response was observed. Therefore, we can conclude that the post-column reactor Table 1. Compensating nature of pressurised post-column reactor. Post-column reactor pressure, 56 Ib in-2 Mobile phase Measured total flow- flow-ratel rate after UV ml min I detectioniml min-I 0.4 0.8 1 .0 1.2 1.6 1.30 1.47 1.55 1 .h2 1.75 pressure is not critical whereas the mobile phase flow-rate in the separation column is. During the course of these experiments it was observed that there was a discrepancy between the set mobile phase flow-rate on the peristaltic pump and the measured total flow-rate after UV detection.This discrepancy can be explained by the compensating nature of the pressurised post-column reactor, see Table 1 where a set of results is presented. When samples containing high acid strengths (5% nitric acid) were injected, the analytical signal was masked com- pletely owing to the overlapping of the solvent front peak. Therefore, the concentration of the mobile phase was altered and its effect on the aluminium retention time observed. Using 0.02 M (NH4)2S04 solutions at pH 1.5 and 5.5 resulted in retention of the aluminium for periods of longer than 20 min. Compromise conditions were obtained with freedom from nitric acid interference with 0.15 M (NH4)2S04 at pH 2.8.This provided an aluminium retention time of approximately 7.5 min. This ability to increase the retention time was applied to the analysis of acid-digested reference materials. Linearity and Detection Limit Good linear calibrations were obtained over 2.5 orders of magnitude from 10 to 5000 pg 1 - 1 of Al. The detection limit, calculated using the definition of three times the standard deviation (30) of ten blank peak height determinations, was 7 ug 1-1 of A1 using a mobile phase of 0.2 M (NH4)2S04 at pH 2.8, and 37 pg 1-1 using a mobile phase of 0.15 M (NH4)?S04 at pH 2.8. The former was calculated using the blank and 10-pg 1- 1 250 270 290 310 330 350 Wavelength, n m Fig. 1. Ultraviolet absorption spectra for Tiron. A , 3 hi ammonium acetate at pH 6.2; B. 3 x 10-1 M Tiron in 3 M ammonium acetate at pH 6.2; C, 1 pg ml 1 of A1 + B; and D, 100 pg ml- 1 of Al + B.Scan speed. 500 nm min-1 I 1.3 1.4 1.5 1.6 1.7 Measured total flow-rateiml min-l Fig. 2. for 100 pg I - I of Al; post-column reactor pressure, 56 Ib in-' Effect of mobile phase flow-rate on the analytical responseANALYST. FEBRUARY 1989, VOL. 114 167 0 1 .0 1.2 1.4 1.6 Measured total flow-rateiml min-' 32 40 52 61 70 Post-column reactor pressureilb in 2 1 1 1 I Fig. 3. response for 100 pg I 1 of Al Effect of post-column reactor pressure on the analytical ( a ) 0.004 A 1 ib) 8 6 4 2 0 6 4 2 0 6 4 2 0 Ti meim i n Fig. 4. Chromatographic responses for aluminium at low concentra- tions. ( u ) 100 i ~ g 1-1 A1 standard, ( h ) 10 ug 1 - 1 Al standard. and (c.) aci d i fie d de - i o n i se d .di s t i 1 I e d water A1 responses, the latter using the blank and 100-yg 1 - 1 A1 responses. Typical responses using the mobile phase 0.2 M (NH,)2S04 at pH 2.8 for the blank, 10 pg 1 - 1 of A1 and 100 pg 1 - 1 of A1 are shown in Fig. 4. The lack of any signal response for aluminium (retention time 4.5 min) in the acid blank shows the low contamination in the analytical methodol- ogy, chromatographic and detection systems. Repeatability The repeatability for ten replicate injections of a 1OO-pg 1-1 standard on seven separate occasions is shown in Table 2. Both peak height and retention time data are consistent with a stable, reproducible system. Table 2 also compares the response obtained for a 100-pg 1-1 standard with a mobile phase of 0.15 M (NH3)?S04 at pH 2.8.The lower peak height and longer retention time lead to less reproducible data; the former accounts for the poorer detection limit observed with this mobile phase. The use of this mobile phase is necessary, however, when solutions of high acidic strength (5% nitric acid) are employed, e . g . , acid digests, as explained previously. Peak Dispersion A comparison of the effect of peak dispersion, measured as the column efficiency, N , on the response for 100 pg 1-1 of A1 in the mobile phases [0.20 and 0.15 M (NH4)?S04 at pH 2.81 is shown in Table 3. The column efficiency was calculated using the equation Table 2. Repeatability o f 100-~~1 injections of a 100-ug I - ' Al standard ( n = 10) Pcak height/ RSD. (Yo Retention RSD, '%, mm ( 1 2 = 10) timeimin ( n = 10) 24.05.88 .. , . 87.8 01.06.88 . . . . 74.5 02.06.88 . . . . 79.2 07.06.88 . . . . 83.3 22.06.88 . . . . 76.5 01.08.88 . . . . 82.6 02.08.88 . . . . 78.8 Mean(n = 7) . . . . 80.4 3.72 4.34 1.14 I .32 4.25 0.33 2.61 4.28 0.73 2.24 4.25 0.70 3.47 4.23 0.61 3.93 4.37 1.23 3.70 4.37 0.80 4.29 95% confidence 02.08.88* . . . . 45.1 5.91 7.32 2.03 interval . . . . k4.2 k0.05 * Eluent 0.15 M (NH4)?S04 at pH 2.8. Table 3. Comparison of peak dispersion for 100 ug 1 I of A1 Mobile phase 0.20 M 0.15 M (N I&)~SO,~ (NHJ)?S04 at pH 2.8 at pH 2.8 99% confidence interval . . k 97 k 166 Mean peak dispersion ( n = 10) 1185 955 1 .o 10 100 0.1 [Cation interferentllvg ml ' Fig. 5 . Effect of cation interferents on the response for 100 p 1 ~ I of Al.Interferent: (0) Ti, ( A ) In. (0) Mn, (0) Mo, (0) Fc", ?X) Zn and (0) Fell1 where t = retention time of peak and Wo.5 = peak width at half height. It was observed that there was no statistical difference, at the 99% confidence interval, for 100 pg 1-1 of A1 using the two mobile phases, thereby indicating that the aluminium peaks have similar column efficiencies irrespective of the mobile phase concentration employed. Interferences Tiron is not a specific spectrophotometric reagent for alu- minium. Therefore, interferences are possible if other species elute at similar retention times to aluminium. This was investigated by analysing a lOO-pg 1-1 Al solution with 0,O. 1, 1 and 10 pg ml-1 of added interferents. The interferents investigated were Ti, In, Mn, Mo, Fe", Fe"' and Zn.Fig. 5 shows the percentage suppression of the aluminium response. It was observed that suppression of the aluminium response was minimal except for Ti. Titanium, present as the oxalate, produced 85% suppression of the aluminium response at the 10-yg ml-1 level. This finding is perhaps not so surprising when one considers the possibility of an aluminium - oxalate complex, [A1(C204)3]3-.1X Three cations had retention times168 ANALYST, FEBRUARY 1989, VOL. 114 Table 4. Analysis of certified reference materials: wet versus microwave digestion. Mobile phase, 0.15 M (NH4)2S04 at pH 2.8 Recovery of added spike. YO A l i ~ g g-I NBS SRM 1572- Certifiedvalue . . 92 5 15 Wet digestion . . 90” (108,84,79) 94 Microwave digestion 18* (18, 18) 87,93 NIES NO.7- Certifiedvalue . . 775 t- 20 Wet digestion . .755* (723,718,767,813) 86,95 * Mean (individual determinations). Table 5. Analysis of certified reference material Alipg 1-1 Certifiedvalue . . . . 23.5 ?c 1.2 Direct analysis” . . . . 11.8t (10.7,12.1, 12.5) 10.6-t (12.1,9.2,10.4) GFAASi: . . . . . . 18k 1 SLRS- 1- ’ Mobile phase 0.20 M (NH&S04 at pH 2.8. t Mean (individual determinations). i: GFAAS = graphite furnace atomic absorption spectrometry separate from that of the solvent peak at 2.2 min. These were Fell, Fell’ and In with retention times of 3 min. This capacity for retention by the analytical column is not surprising for the last two as both exist as the 3+ ion, cf., Al3+. Determination of Aluminium in Certified Reference Materials The results for the determination of aluminium in the certified reference materials after wet and microwave digestion using a mobile phase of 0.15 M (NH4)2S04 at pH 2.8, for reasons previously outlined, are shown in Table 4.The results indicate that wet digestion provides a method for accurate and relatively precise determination of aluminium in biological materials. However, unsatisfactory results were obtained using microwave digestion. Microwave digestion has typically been employed with analytical techniques that provide further heating on aqueous samples, i. e. atomic absorption spec- trometry with flamel9-21 and graphite furnace atomisation21 or inductively coupled plasma atomic emission spec- trometry.21-24 The conclusion, therefore, is that a non-des- tructive technique such as ion chromatography is not suitable for the analysis of microwave samples and that the term microwave dissolution2’ is to be preferred to microwave digestion.This recommendation is due to the fact that the microwave energy employed in this work was not sufficient to dissociate organic matter and to liberate Al”, or more correctly [AI(H2O)#+, in acidic (pH <5) solution.5 The results for the determination of aluminium in the riverine water reference material, SLRS-1, using the mobile phase 0.20 M (NI-T4)2S04 at pH 2.8 are shown in Table 5 . It was observed that there was a discrepancy between the certified value and that obtained using the proposed method. This could be due to the employment of a non-destructive technique and its inability to “see” combined aluminium1.2 naturally present in water supplies such as the St.Lawrence River in Canada from where the reference material was sampled. Analysis by graphite furnace atomic absorption spcctrometry produced a result in close agreement with the certified value. Future studies will investigate the aluminium species associated with environmental samples. The lack of interferences indicates the benefits of the chromatographic system for rapid determinations. The method was applied to the analysis of biological and environ- mental materials by either acid digestion (conventional wet ashing or microwave dissolution) or direct analysis. Dis- crepancies with microwave dissolution were believed to be due to incomplete digestion of the biological reference material.The complex nature of aluminium in natural waters prevented direct determination by this method. However, the possibility of using a similar system to identify the naturally occurring species of aluminium in environmental samples is under investigation. I gratefully acknowledge the Analytical Chemistry Research Fund for the purchase of some of the experimental equipment. The assistance of Mrs. M. O’Hara (Department of Chemistry, University of Manchester, UK) for the graphite furnace atomic absorption spectrometry analysis and Mr. M. Kelly (Huddersfield Polytechnic) is acknowledged. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. References Driseoll, C. T., Environ. Health Perspect., 1985, 63, 93. Bull, K. R . , and Hall, J .R., Environ. Pollut. (Series B ) , 1986. 12, 165. Lawrence, G. B . , Driscoll, C. T., and Fuller, R. D., Water Resour. Res., 1988, 24, 659. Goenaga, X.. Bryant, R.. and Williams, D. J . A., Anal. Chem., 1987. 59, 2673. Martin, R. B . , Clin. Chem., 1986, 32, 1797. Tennakone, K., Wickramanayake, S . , and Fernando, C. A. N., Environ. Pollut., 1988, 49, 133. Perl, D. P., Environ. Health Perspect., 1985, 63, 149. Wills, M. R., and Savory. J . , Environ. Health Perspect., 1985, 63, 141. Fritz, J. S . , Anal. Chem., 1987, 59, 335A. Weib. J., Fressmius Z. Anal. Chem., 1987, 327, 451. Jones, P., Ebdon, L., and Williams, T., Analysr, 1988, 113, 641. Yan, D.-R., and Schwedt, G., FreseniuJ 2. Anal. Chem., 1985, 320, 252. Bertsch, P. M., and Anderson, M. A., Soil Sci. Soc. Am. J . , 1988, 52, 540. Schwarzenbach, G., and Willi, A., Helv. Chim. Acta, 1950,33. 947. Lajunen. L. H. J., Ruotsalainen. H., Raisanen, K., and Parhi, S . . Finn. Chem. Lett., 1980, 5 , 142. Aldrich Chemical Company Ltd.. Gillingham, UK. Catalogue Dionex (UK) Ltd., Camberlcy, UK, Application Note 42, January 1986. Cotton, F. A., Wilkinson, G., and Gaus, P. L., “Basic Inorganic Chemistry,” Second Edition, Wiley, Chichester, 1987, p. 143. Aysola, P., Anderson, P., and Langford, C. H., Anal. Chem., 1987, 59, 1582. Demura, R., Tsukada, S . , and Yamamoto, I . , Eisei Kaguku, 1985, 31, 405. Nakashima, S . , Sturgeon, R. E., Willie, S. N., and Berman, S. S . , Analyst, 1988, 113, 159. Nadkarni, R. A., Anal. Chem., 1984, 56, 2233. Nikdel, S . , and Temelli, C. M., Microchem. J., 1987, 36, 240. White, R. T.. Jr., and Douthit, G. E., J. Assoc. Off. Anal. Chem., 1985, 68, 766. Borman, S. A , , Anal. Chem., 1988, 60, 715A. NO. 17, 255-3, 1988-89. Conclusions Ion chromatography with post-column derivatisation allows the determination of aluminium in the range 7-5000 pg I-’. Paper 8l03226B Received August 8th, 1988 Accepted October 20th, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400165

出版商:RSC    

年代:1989

数据来源: RSC