摘要:

Alan Date Memorial Award 1991In recognition of the considerable contribution made byDr Alan Date to the field of inductively coupled plasma sourcemass spectrometry, an annual commemorative awardhas been created to encourage talented young scientists tobroaden their overseas scientific experience.Candidates should be working in the field of atmosphericpressure plasma source mass spectrometry (inorganic)and will be required to submit a written resume oftheir work in this field. It is intended thatthis may form the basis of a publication.An amount of up to $1000 will be available to thesuccessful candidate(s)Closing date for submissions will be October 26th 1991.Further details of the award may be obtained from:Dr Robert Hutton, VG Elemental, Ion Path, Road Three,Winsford, Cheshire, CW7 3BX, UKSixth Biennial NationalAtomic Spectroscopy Symposiumwill be held atPolytechnic South West, Plymouth, UK22-24 July I992The symposium will provide a forum where interesting and useful applications of atomic spectros-copy can be reported and discussed. In addition to plenary, invited and submitted lectures, a particu-lar feature of the meeting will be the presentation of posters.There will also be an exhibition and asocial programme for delegates and their guests.Scientific programme will include:Plenary Lecturers-M.W. Blades (Vancouver, BC, Canada)B.V. L’vov (Leningrad, USSR)J. W. McLaren (Ottawa, Ontario, Canada)K. Niemax (Dortmund, Germany)B.L. Sharp (Lmghborough, UK)Invited Lecturers-J. S . Crighton (Sunbury -on-Thames, UK)H. Falk (Kleve, Germany)S.J. Hill (Plymouth, UK)D. Littlejohn (Glasgow, UK)C. McLeod (Shefleld, UK)G. Schlemmer (Uberlingen, Germany)P. Stockwell (Sevenoaks, UK)J.F. Tyson (Amherst, MA, USA)J.G. Williams (Egham, UK)A.M. Ure (Glasgow, UK)This meeting is organized by the Atomic Spectroscopy Group, Analytical Division of The RoyalSociety of Chemistry.Further information can be obtained from the Chairman of the organizing committee:Dr S. J. Hill, Department of Environmental Sciences, Polytechnic South West, Drake Circus,Plymouth, Devon PL4 8AA, UK

ISSN:0003-2654    

DOI:10.1039/AN99116BP031

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

The AnalystThe Analytical Journal of The Royal Society of ChemistryAnalytical Editorial BoardChairman: A. G. Fogg (Loughborough, UK)K. D. Bartle (Leeds, UK)D. Betteridge (Sunbury-on-Thames, UK)J. Egan (Cambridge, UK)H. M. Frey (Reading, UK)D. E. Games (Swansea, UK)S. J. Hill (Plymouth, UK)D. L. Miles (Keyworth, UK)R. M. Miller (Port Sunlight, UK)B. L. Sharp (Loughborough, UK)J. F. Alder (Manchester, UK)A. M. Bond (Australia)R. F. Browner (USA)D. T. Burns (Belfast, UK)J. G. Dorsey USA)J. P. Foley (USA)T. P. Hadjiioannou (Greece)W. R. Heineman (USA)A. Hulanicki (Poland)I. Karube (Japan)E. J. Newman (Poole, UK)T. B. Pierce (Harwell, UME. Pungor (Hungary)Advisory BoardJ. RBiiCka (USA)R. M. Smith (Loughborough, UK)M. Stoeppler (Germany)J.D. R. Thomas (Cardiff, UK)J. M. Thompson (Birmingham, UK)K. C. Thompson (Sheffield, UK)P. C. Uden (USA)A. M. Ure (Aberdeen, UK)P. Vadgama (Manchester, UK)C. M. G. van den Berg (Liverpool, UK)A. Walsh, K.B. (Australia)J. Wang (USA)T. S. West (Aberdeen, UK)Regional Advisory EditorsFor advice and help to authors outside the UKProfessor Dr. U. A. Th. Brinkman, Free University of Amsterdam, 1083 de Boelelaan, 1081 HVProfessor Dr. sc. K. Dittrich, Analytisches Zentrum, Sektion Chemie, Karl-Marx-Universitat,Dr. 0. Osibanjo, Department of Chemistry, University of Ibadan, Ibadan, NIGERIA.Professor K. Saito, Coordination Chemistry Laboratories, Institute for Molecular Science,Professor M . Thompson, Department of Chemistry, University of Toronto, 80 St.GeorgeProfessor Dr. M. Valcarcel, Departamento de Quimica Analitica, Facultad de Ciencias,Professor J. F. van Staden, Department of Chemistry, University of Pretoria, Pretoria 0002,Professor Yu Ru-Qin, Department of Chemistry and Chemical Engineering, Hunan University,Professor Yu. A. Zolotov, Kurnakov Institute of General and Inorganic Chemistry, 31 LeninAmsterdam, THE NETHERLANDS.Talstr. 35, DDR-7010 Leipzig, GERMANY.Myodaiji, Okazaki 444, JAPAN.Street, Toronto, Ontario M5S I A I , CANADA.Universidad de Cordoba, 14005 Cordoba, SPAIN.SOUTH AFRICA.Changsha, PEOPLES REPUBLIC OF CHINA.Avenue, 117907, Moscow V-71, USSR.Editorial Manager, Analytical Journals: Judith EganEditor, The AnalystHarpal S. MinhasThe Royal Society of Chemistry,Thomas Graham House, Science Park,Milton Road, Cambridge CB44WF, UKTelephone 0223 420066.Fax 0223 423623.Telex No. 81 8293 ROYAL.Senior Assistant EditorPaul DelaneyUS Associate Editor, The AnalystDr J. F. TysonDepartment of Chemistry,University of Massachusetts,A m herst M A 01 003, USATelephone 41 3 545 01 95Fax 41 3 545 4490Assis tan t EditorsBrenda Holliday, Paula O’Riordan, Sheryl WhitewoodEditorial Secretary: Claire HarrisAdvertisements: Advertisement Department, The Royal Society of Chemistry, BurlingtonHouse, Piccadilly, London, W I V OBN. Telephone 071-437 8656. Telex No. 268001.Fax 071-437 8883.The Analyst (ISSN 0003-2654) is published monthly by The Royal Society of Chemistry,Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, UK.All orders,accompanied with payment by cheque in sterling, payable o n a UK clearing bank or in USdollars payable on a US clearing bank, should be sent directly to The Royal Society ofChemistry, Turpin Transactions Ltd., Blackhorse Road, Letchworth, Herts SG6 1 HN, UnitedKingdom. Turpin Transactions Ltd., distributors, is wholly owned by the Royal Society ofChemistry. 1991 Annual subscription rate EC f246.00, USA $580, Rest of World f283.00.Purchased with Analytical Abstracts EC €551 .OO, USA $1 299.00, Rest of World €634.00.Purchased with Analytical Abstracts plus Analytical Proceedings EC €648.00, USA $1 527.00,Rest of World €745.00. Purchased with Analytical Proceedings EC €313.00, USA $738.00, Restof World f360.00.Air freight and mailing in the USA by Publications Expediting Inc., 200Meacham Avenue, Elmont, NY 11003.USA Postmaster: Send address changes to: The Analyst, Publications Expediting Inc., 200Meacham Avenue, Elmont, NY 11003. Second class postage paid at Jamaica, NY 11431. Allother despatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Postoutside 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 a l l aspects ofthe theory and practice of analytical chemistry,fundamental and applied, inorganic andorganic, including chemical, physical, biochem-ical, clinical, pharmaceutical, biological,environmental, automatic and computer-basedmethods.Papers on new approaches to existingmethods, new techniques and instrumentation,detectors and sensors, and new areas of appli-cation with due attention to overcoming limita-tions and to underlying principles are all equallywelcome. There is no page charge.The following types of papers will be con-sidered:Full research papers.Communications, 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 at which it is likely to be valuable toworkers faced with similar problems. A fullerpaper may be offered subsequently, if justifiedby later work. Although publication is at thediscretion of the Editor, communications will beexamined by a t least one referee.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 at least two referees, by whoseadvice the Editorial Board of The Analyst will 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 and North America, a Group of RegionalAdvisory Editors exists. Requests for help oradvice on any matter related to the preparationof papers and their submission for publicationin The Analyst can be sent to the nearestmember of the Group. Currently servingRegional Advisory Editors are listed in eachissue of The Analyst.Manuscripts (four copies typed in double spac-ing) should be addressed to:Harpal S. Minhas, Editor, The Analyst,Royal Society of Chemistry,Thomas Graham House,Science Park, Milton Road,CAMBRIDGE CB4 4WF, UK or:Dr.J. F. TysonUS Associate Editor, The AnalystDepartment of ChemistryUniversity of MassachusettsAmherst MA 01003, USAParticular attention should be paid to the use ofstandard methods of literaturecitation, 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 either to the Editor, or AssociateEditor, The Analyst (addresses as above). Mem-bers of the Analytical Editorial Board (who maybe contacted directly or via the Editorial Office)would welcome comments, suggestions andadvice on general policy matters concerningThe Analyst.Fifty reprints are supplied free of charge.@ The Royal Society of Chemistry, 1991. 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 permissionof the publishers

ISSN:0003-2654    

DOI:10.1039/AN99116FX033

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALAO 116(9) 881-974 (1991)The AnalystSeptember 199188189189790190590991 391992392993393794194795195796 1965969973The Analytical Journal of The Royal Society of ChemistryCONTENTSThickness-shear-mode Acoustic Wave Sensors in the Liquid Phase. A Review-Michael Thompson, Arlin L. Kipling,Wendy C. Duncan-Hewitt, Ljubinka V. Rajakovic, Biljana A. Cavic-VlasakOn-line lmmunoaff inity Sample Pre-treatment for Column Liquid Chromatography: Evaluation of DesorptionTechniques and Operating Conditions Using an Anti-estrogen Immuno-precolumn as a Model System-AriaFarjam, Anita E. Brugman, Henk Lingeman, Udo A. Th. BrinkmanSimultaneous Determination of Seven Divalent Metal Cations in Some Anaerobic Sealant Formulations FollowingSolid-phase Extraction and Separation on a Dynamically Coated CI8 High-performance Liquid ChromatographyColumn-Marian Deacon, Malcolm R.Smyth, Raymond G. LeonardThermal Degradation of Some Benzyltrialkylammonium Salts Using Pyrolysis-Gas Chromatography-Mass Spec-trometry-Neville J. Haskins, Robert MitchellHigh-performance Modular Spectrophotometric Flow Cell-Jo3o Carlos de Andrade, Kenneth E. Collins, MBnicaFerreiraGeneralized Treatment of a Stray Radiant Energy Test Method in Absorption Spectrometry-Paddy FlemingDetermination of Iron by Flow Injection Based on the Catalytic Effect of the lron(tti)-Ethylenediaminetetraacetic AcidComplex on the Oxidation of Hydroxylamine by Dissolved Oxygen-Andreu Cladera, Enrique Gomez, Jose ManuelEstela, Victor CerdaSimultaneous Determination of Toxic Metabolites by Linear Combination Derivative Spectrophotometry-Lin Liming,Zhao NaixinPolymer-based Cation-selective Electrodes Modified With Naphthalenesulphonates-Tatsuhiro Okada, HidenoriHayashi, Kazuhisa Hiratani, Hideki Sugihara, Naoto KoshizakiDetermination of Gluthathione at Enzyme-modified and Unmodified Glassy Carbon Electrodes-Chi Hua, Malcolm R.Smyth, Ciaran O’FagainAmperometric Monitoring of Bacteria-induced Milk Acidity Using a Platinum Disc Microelectrode-M.AntoniettaBaldo, Salvatore Daniele, Gian A. Mazzocchin, Marco DonatiStudy of Complexation Equilibria Using Polarized Metallic Electrodes-V. F. Vetere, R. RomagnoliDifferential-pulse Polarographic Behaviour of Selenium in the Presence of Copper, Cadmium and Lead-Hasan Aydin,G.H. TanThermodynamic and Kinetic Implications Involved in the Titration of Polyfunctional Acids by Catalytic ThermometricTitrimetry-Oswaldo E. S. Godinho, Helena S. Nakatani, Ivo M. Raimundo Jr., Luiz M. Aleixo, Gracilliano de OliveiraNet0lodimetric Method for the Determination of Mono- and Disaccharides With Vanadium(v) in Perchloric Acid-AmalenduBanerjee, Banasri Hazra, Anuva Putatunda, Dinabandhu Mandal, Gopal Chandra Banerjee, Sachchidananda DuttApplication of a Microwave Oven for Drying and Nitric Acid Extraction of Mercury and Selenium From FishTissue-Suei Y. LamLeung, Vincent K. W. Cheng, Yuet W. LamSeparation of Niobium From Chloride Media by Solvent Extraction With Dicyclohexyl-l8-crown-&N. V. Deorkar, S. M.Kho pka rSynthesis of a Phosphoramidate Chelating Fibre and Its Adsorption of Trace Amounts of Gallium and Indium-XingyinLuo, Zhixing Su, Xijun Chang, Guangyao Zhan, Xihuan ChaoBOOK REVIEWSCUMULATIVE AUTHOR INDEXTypeset and printed by Black Bear Press Limited, Cambridge, Englan

ISSN:0003-2654    

DOI:10.1039/AN99116BX035

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

88 1 ANALYST, SEPTEMBER 1991, VOL. 116 Thickness-shear-mode Acoustic Wave Sensors in the Liquid Phase A Review Michael Thompson,* Arlin L. Kiplingt and Wendy C. Duncan-HewittS Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S ?A I , Canada Ljubinka V. Rajakovic and Biljana A. Cavic-Vlasak Faculty of Technology and Metallurgy, University of Belgrade, Carnegie Street 4, I 1000 Belgrade, Yugoslavia Summary of Contents Introduction Theoretical Aspects The TSM Sensor in the Gas Phase The TSM Sensor in the Liquid Phase Measurement Methods Applications Detectors for Liquid Chromatography Determination of Inorganic Species Development of Biosensors Properties of Thin Films Conclusions Appendix References Introduction The use of piezoelectric acoustic wave devices as chemical sensors has its origins in the work of Sauerbreyl and King2 who carried out microgravimetric measurements in the gas phase.In their work, they assumed that a thin film applied to a thickness-shear-mode (TSM) device could be treated in sensor measurements as an equivalent mass change of the quartz crystal itself. Accordingly, a shift in the resonance frequency of an oscillating AT-cut crystal could be correlated quantita- tively with a change in mass added to or removed from the surface of the device. This concept has been exploited extensively in the fabrication of chemically selective sensors for the gas phase, where a binding agent is incorporated into a film which is then deposited onto the TSM device. In recent times, there has been an increasing amount of attention paid to the operation and applications of the TSM sensor when it is exposed to the liquid phase.3-29 Studies have been made of: in situ deposition of films on the sensor surface, interfacial chemistry and bulk liquid phase properties such as density, viscosity and conductivity. Where deposition or removal of surface species has been involved, frequency shifts have invariably been interpreted in terms of Sauerbrey-like alteration of acoustic wavelength as postulated for the gas phase.In contrast, Thompson et 01.17 proposed that the possibility of changes of interfacial properties such as free energy and slippage were related to the behaviour of the liquid-phase TSM sensor. In this present paper the theoretical aspects are reviewed and measurement methods and applications are suggested for the TSM device operated in the liquid phase.Particular emphasis is placed on the role of interfacial parameters, a * To whom correspondence should be addressed. 'F Permanent address: Department of Physics, Concordia Univer- sity, 1455 de Maisonneuve Boulevard, Montreal, Quebec H3G 1M8, Canada. t Permanent address: Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Ontario M5S 1A1, Canada. number of which are depicted in Fig. 1, in determining the response of the device. It is recognized that other acoustic wave structures, such as plate-mode and surface acoustic wave sensors, have been employed in the liquid phase. However, in order to be concise the present review deals for the most part only with thickness-shear-mode devices.Hydrophobicity , Deposited mass Thickness (LA \ Slip 1 \\ \ \ Electrodes Oscillating quartz crystal Schematic diagram of interfacial factors that govern the Fig. 1 behaviour of the oscillating TSM sensor in the liquid phase882 Theoretical Aspects The TSM Sensor in the Gas Phase In order to consider carefully the behaviour of the TSM structure in the liquid phase, it is necessary to review briefly the physical aspects of the operation of the sensor in the gas phase. For AT-cut piezoelectric crystals, the resonant condition corresponds to a thickness shear oscillation in which the shear wave propagates through the bulk of the material, perpen- dicular to the faces of the crystal. The atomic displacements corresponding to this shear motion are thus parallel to the crystal surface.If material is deposited on either one or both faces of the crystal, the resonant frequency decreases. The first quantitative investigation of this effect was made by Sauerbreyl who derived the relationship for the change in frequency AF (in Hz) caused by the added mass AM (in g): 2F 2 (1) A F = - L X - AM where Fq is the fundamental resonant frequency of unloaded quartz, pq is the shear modulus of AT-cut quartz (2.947 x 10" g cm-1 s-2), pq is the density of quartz (2.648 g cm-3) and A is the surface area in cm2. Note that AM is the total mass added to both faces of the crystal but A is the area of only one face. Collecting constants and letting Am = AM/A gives AF= -CIAm (2) where C1 = 2.26 x 10-6 Fq2 Hz cm2 8-1 and Am is added mass per unit area in g cm-2.Hence AF is linearly related to Am and this simple relationship is the basis for the application of piezoelectric crystals with a detection limit that has been estimated to be of the order of 10-12 8.2 For a 5 MHz crystal, according to eqn. (2) a decrease in frequency of 1 Hz is caused by an added mass of 18 ng cm-2. For a 9 MHz crystal Am = 5.5 ng cm-2 when A F = - 1 Hz, or A FlAm = -0.18 Hz cm2 ng- 1 . The sensitivity of a particular crystal can be defined as AF/AM and, for example, for a 9 MHz crystal with an electrode ANALYST. SEPTEMBER 1991. VOL. 116 diameter of 0.5 cm ( A = 0.20 cmz), AFIAM = 0.9 Hz ng-1, ignoring the minus sign. Sauerbrey's theory has been verified for the application of 'rigid' overlayers up to a mass load of Amlm = 2%3O where m is mass per unit area of the unloaded quartz.Several attempts have been made to expand Sauerbrey's theory by including a number of other parameters associated with deposited thin films. These are collected together in ,Table 1 for reference. Stockbridge31 applied a perturbation analysis to the loaded crystal. His mathematically rigorous approach was not, however, of immediate practical utility. Miller and Bolef32.33 applied a continuous acoustic wave analysis to a resonator formed by the quartz crystal and a deposited film. Provided that the acoustic losses in the quartz and thin film are small, it can be shown that the frequency is dependent on the shear wave velocity and density of both the film and quartz.Behrndt34 pointed out that the multiple oscillation period measurement technique is superior to the frequency measure- ment. Lu and Lewis35 and Lu36 have shown that for a metal film the acoustic impedance of the film is an important factor in determining the frequency response. Glassford37.38 made the first attempt to analyse the frequency response associated with an imposed liquid film and droplet deposit. By using the Navier-Stokes equation and a Rayleigh perturbation analysis Glassford was able to derive a relationship which included liquid viscosity, droplet size, velocity distribution and mass loading. Mecea and Bucur39 developed an energy transfer model which constitutes the most complex theory describing the functioning of the TSM acoustic wave device in the gas phase.They considered the mechanism of thin film interaction with the elastic properties of the resonating quartz crystal. The analogy of electrical reactance in series for two components has been applied to a coated piezoelectric crystal by Crane and Fisc her. 40 The TSM Sensor in the Liquid Phase In 1984, the view was expressed that liquid phase operation of the TSM device would not be possible because of oscillation Table 1 Gas phase theories of TSM acoustic wave quartz sensors* Equation AuthorsRef Mathematical model Sauerbre y 1 AM AF= -2.26 X 10-6 Fq2 - (1) A Miller and B 01 e f32.33 (2) ( 3 ) B e h ~ - n d t ~ ~ Lu and Lewis35 and Lu36 (4) ( 5 ) GI assf or d37.38 Mecea and Bucu1-3~ (6) tan = - z, Zf tan (2) AF F4 OL tan Ib(1- tan h2kb) + fi tanh kb (1 + tan%) npqVq (1 + tanh2kb tan2lb) - - _ Crane and Fischer40 - (7) * A glossary of the symbols used is given in the Appendix. Comments Only the effect of added mass is considered Propagation of the acoustic wave from the quartz into the deposited film is considered considered The change of period is The acoustic impedance of the film is considered Mass loading by a liquid film is considered Effects of electrode, film and quartz diameters on the sensitivity of a crystal coated with a thin or non-dissipating film are considered Bulk modulus, viscosity, density, and film thickness are consideredANALYST, SEPTEMBER 1991, VOL.116 883 suppression caused by viscous damping effects,41 despite an earlier study which demonstrated that a coated TSM device could be employed as a detector for liquid chromatography.The subsequent successful measurements in the liquid phase that involved new experimental techniques, spawned attempts to provide theories for coupling of the oscillating surface to a liquid medium. In this area attempts have continued to the present time; the various theories are summarized in Table 2. The first argument, based upon an empirical formulation was presented by Nomura and Minemura.5 The change in the resonant frequency that occurs on coupling one face of a piezoelectric crystal to an aqueous solution is ascribed to the density, p ( g cm-3), and specific conductivity, K ( Q - 1 cm-1) of the solution. For organic liquids the resulting change in frequency for total immersion of the device arises from the density (p) and viscosity [q(cP)] of the liquid.7 Although good agreement with experimental results was obtained, the disadvantages of these relationships are obvious; the numerical constants are characteristic of each particular crystal and the relationships are not based on a physical model. In 1985 two simple physical models were developed by Bruckenstein and Shay13 and Kanazawa and Gordon14 which predict the change of frequency of a sensor immersed in a liquid medium.The latter theory treats the quartz as a lossless elastic solid, and the liquid as a purely viscous fluid. The frequency shift arises from coupling the oscillation of the crystal, involving a standing shear wave, with a damped propagation shear wave in the liquid. A simple relationship was derived which expresses the change in resonant frequency of a piezoelectric crystal, due to the total contact of one face of the crystal with liquid, in terms of parameters that are characteristic of the crystal and the liquid phase [eqn.(lo), Table 21. The relationship for the change in frequency, AF, is derived with the assumption that the transverse velocity of the quartz surface is identical with that of the adjacent fluid layer. This simple shear wave model also provides a physical explanation of the fact that the velocity is important for TSM acoustic wave devices operating in the liquid phase. According to this theory the crystal does not drive the entire bulk of the liquid as the transverse displacement decays exponentially in the liquid with a characteristic decay length (Fig.2). This distance (6) varies with (qL)i and is the effective thickness of the liquid when it is treated as a rigid sheet. Accordingly, the added mass of liquid can be derived as 6pL = ( 2 p ~ q ~ h ) ; . Physically this model predicts that only a thin layer of liquid will undergo displacement at the surface of the bulk wave device, and the device response will be a function of the mass of this layer. By using dimensional analysis Bruckenstein and Shayl3 derived a similar although not identical relationship which can be applied to the bulk acoustic wave devices with one or two faces in contact with the liquid. In eqn. (11) (Table 2) AF is the difference between the frequency in air and the frequency in liquid when the liquid is in contact with one or both electrodes of the crystal (AFfor two electrodes in the liquid is twice as large as for one electrode).For a 9 MHz crystal with one electrode ( n = 1) in contact with water (pL = 1 g cm-3, qL = 10-2 g cm-1 s-I), the theory predicts that AF = It appears that both theories introduced new parameters for -6100 Hz. Table 2 Liquid phase theories of TSM acoustic wave quartz sensors* AuthorsRe' Nomura and Minemuras Comments Empirical formulation for aqueous solution. Conductivity and specific gravity are considered Nomura and Okuhara7 Empirical formulation. Viscosity and density are considered Viscosity and density are considered. Interfacial effects are disregarded Kanazawa and Gordon l 4 Bruckenstein and Shayl3 Viscosity and density are considered.Interfacial effects are disregarded Schumacher and co-workers25329 AF= - Surface roughness is considered Schumacher and co-workers25,29 PLE AmL = - 2 Heusler et al. 42 Hagerj4 Surface stress is considered Hydrodynamic coupling analysis. Liquid dielectric constant is considered Resistance of equivalent circuit is considered Muramatsu et a1.22 Similar to Nomura and Okuhara model7 but liquid dielectric constant is considered Yao and Zhou45 Shana et d.46 Piezoelectric effects are considered * A glossary of all the symbols used is given in the Appendix.884 ANALYST, SEPTEMBER 1991, VOL. 116 Fig. 2 sensor into a liquid Propagation of the transverse shear wave from the TSM the mass of a thin boundary film, in this instance a liquid film. Apparently, qL and pL are the parameters which are relevant to the operation and characterization of these devices in the liquid phase.Although experimental verification has been claimed for both theories, these approaches lack considera- tion of the microscopic boundary conditions between the crystal surface and liquid. Specifically, in order to understand the behaviour of acoustic wave sensors operating in the liquid phase, it is important to consider the effect of chemical reactions and processes that change in addition to mass, structure, surface free energy, interfacial viscosity and, possibly, diffusion on the crystal surface. More recent theories have examined some interfacial phenomena. Schumacher and CO-workers25-29 showed that surface roughness can also drastically affect the resonant frequency.They considered a roughened surface made up from hemicylinders with liquid entrapments which can be equivalently represented by a rigidly attached liquid layer. Under these conditions the frequency shift should contain the parameter AmL = pL&/2, where AmL is the mass per unit area of the liquid confined in the cavities of the roughened surface and E is the mean diameter of the hemicylinders. Heusler et al.42 promoted the theory of surface stress influence on the resonant frequency. The internal strains of quartz crystals immersed in a liquid arise as a result of hydrostatic pressure. Parabolic dependence of the resonant frequency, Fq, on the pressure, p , at the oscillator face [eqn. (14), Table 21 could be related to the elastic energy stored in the quartz.Hager and Verge43 and Hagel-44 derived a model using hydrodynamic coupling analysis to determine fluid properties. In their work , viscous energy losses, fluid velocity at the crystal surface and the dielectric effect of the liquid were considered. The frequency shift is described by an empirical equation with constants depending on the crystal equivalent circuit and the working conditions. Considering an electro- mechanical coupling analysis for the computation of equi- valent circuit parameters of piezoelectric devices in, contact with a liquid, Muramatsu et a1.22 developed a linear relation- ship between AF and (vLpL)g for alcohol-water solutions. Shifts from linearity were observed for high-viscosity liquids and when both faces of the crystals were in contact with water.According to Yao and Zhou45 the frequency response of an acoustic wave sensor in the liquid phase depends on the dielectric and conductance effects of the liquid. Based on their experimental measurements an empirical relationship resem- bling the expression of Nomura and Okuhara7 [eqn. (9), Table 21 was derived [eqn. (17), Table 21, As shown by Shana efal.46 a comprehensive analysis including piezoelectric effects could also be used to study a thin film of viscous liquid [eqn. (18), Table 21. When the piezoelectric effect is neglected this theory is similar to that derived by Kanazawa and Gordon.14 Thompson and co-~orkers17,*8 have shown that the response of the sensor can be associated with changes in interfacial surface structure, surface free energy andor interfacial viscosity. From a qualitative analysis of the operation of TSM acoustic wave devices in a liquid, two aspects were introduced.First, chemistry at the interface can lead to a perturbation of acoustic wave transmission caused by alteration of the partial-slip boundary condition at the interface. Secondly, as discussed by many of the workers mentioned above, the continuum viscosity and density of the bulk liquid will be important in determining the frequency response. Using the Navier-Stokes equation for Newtonian fluids and an expression describing transverse damped wave propagation, it was shown that the penetration depth of the wave is about 1 pm. The influence of the interfacial free energy and viscosity within this range of the penetration depth was considered.These effects could be associated with either new material deposited on the interface or by the time required for molecular re-orientation in the interfacial boun- dary layer. As an efficient acoustic bond demands continuity of stress and displacement across the interface, wave propaga- tion could be altered by out-of-equilibrium interfacial effects. For such an example, it was suggested that the frequency of .a crystal exposed to water may increase despite an apparent increase in deposited material. Clearly, in order to accept this argument it is necessary to invoke the controversial slip boundary condition for a solid-liquid interface in which either the solid or the liquid is in motion. Solving the differential equations of motion for the oscilla- tory system, consisting of the TSM device and the liquid with which it is in contact, requires that mathematical continuity be invoked at the interface.In order to accommodate the possibility of interfacial slip at the sensor-liquid interface,47 Duncan-Hewitt and Thompson48 have introduced the concept of additional interfacial regions of finite thickness in the liquid. These layers are endowed with mechanical properties that can be used to explain the experimental finding that coupling with the liquid phase appears to decrease with decreasing surface wettability. The existence of these pro- posed additional fluid layers has been verified by a wealth of experimental and theoretical work.49 In particular, thermodynamic analyses of adsorption isotherms indicate that at vapour saturation the bulk liquid must be at equilibrium with a surface adjacent film.Between this denser and more viscous layer and the bulk is a thin, monomolecular transitional region which is rarified relative to the bulk under incomplete wetting conditions, as in Fig. 3. Under completely non-wetting conditions, it is expected that this region would be indistinguishable from the vapour; acoustic waves would be reflected from this interface and the TSM sensor should behave as though no bulk liquid were present. Using the molecular theory of viscosity proposed by Krausz and Eyring,so Duncan-Hewitt and Thompson48 have provided a link between measurable bulk liquid properties such as density, viscosity, surface tension and contact angle, and the TSM sensor response.The model has provided results that are in agreement with the experimental results obtained to date and obeys the boundary conditions described above.ANALYST, SEPTEMBER 1991, VOL. 116 885 5.95 x 106 FrequencyIHz E c -8 .- - - - -10 C I -i Fig. 3 Four-layer model of the TSM sensor in liquid with predicted impedance-frequency plots for (a) complete wetting and ( b ) incom- plete wetting. Layers depicted from bottom to top are: solid piezoelectric material, interfacial liquid structure, rarified layer and bulk liquid Measurement Methods There are two types of methods used to characterize a quartz crystal sensor electrically, which may be called the active and passive methods. The active method is more commonly known as the oscillator method.In this method the quartz crystal is part of an oscillator circuit. It is connected between the output and input of the oscillator amplifier and provides positive feedback that causes oscillation of the circuit. The resonant frequency of the quartz crystal is measured by an electronic counter. The quartz crystal is active in the sense that it is continuously oscillating at a frequency controlled by the quartz crystal itself. In the passive method the quartz crystal is connected externally to an instrument which applies a sinusoidal voltage at various frequencies across the terminals of the crystal. Voltages are measured and then the electrical characteristics of the crystal, the impedance for example, can be found from the voltages. The crystal does not determine the frequencies at which the measurements are made and in this sense it is passive.(0 C, Fig. 4 (b) impedances of the circuit elements Equivalent circuit of the TSM sensor with (a) parameters and Cadysl was a pioneer in the development of the piezoelec- tric quartz crystal for frequency control in the communications field. His treatment is advanced and difficult to understand, but is one of the milestones in the literature. Bottom52 has presented the fundamentals of the theory of the quartz crystal simply and clearly. This work is recommended for beginners in the subject. Both Bottom52 and Cadys' have derived the equivalent circuit of the quartz crystal (Fig. 4). This circuit is the electrical model of the crystal in a gas, but it is not a good representation of the crystal in a liquid.When the crystal is immersed in a liquid, energy flows out of the quartz into the liquid in the form of acoustic waves and this is dependent on the properties of the sensor-liquid interface which are not included in the circuit model. However, the equivalent circuit does illustrate the main features of the behaviour of the quartz crystal in liquids of low viscosity. Hence the measurement methods will be described with reference to the impedance of the equivalent circuit in Fig. 4. Most quartz crystals are discs of AT-cut quartz which are TSM devices, that is, the atoms of the quartz oscillate in the plane of the disc. The resonant frequency of the sensors is in the range from 2 to 20 MHz. The impedance, 2, of the quartz crystal is complex: 2 = R + j X , where R is the real part of 2, the resistive part, and Xis the imaginary part of 2, the reactance.For the equivalent circuit (Fig. 4) the expression for the admittance, Y , is simpler than the expression for 2. By definition, Y is the reciprocal of 2 (ie., Y = l/Z) therefore if 2 is complex, so is Y : Y = G + j B ; where the conductance, G, is the real part of Y and the susceptance, B , is the imaginary part of Y . For the circuit shown in Fig. 4, (3) and The quantity, O, is the angular frequency (in rad s-1) and is defined by w = 2nfwherefis the frequency in Hz. For brevity the angular frequency will be called simply the frequency.886 ANALYST, SEPTEMBER 1991, VOL. 116 From Z = l / Y , the real and imaginary parts of Z in terms of G and B can be written as ( 5 ) The magnitude of impedance, 121, and the phase of im- pedance, 8,, are defined as Iz( = V F T Z (7) X 8, = tan-1- R The oscillator method measures the lower of the two frequencies of the quartz crystal for which 8, = 0.The explanation for this is as follows. There are two conditions that must be satisfied for oscillation to occur: the phase shift around the loop should be zero and the loop gain should be unity, the loop being the closed path from the input of the amplifier through the amplifier to its output, and back to the input through the feedback circuit element. The amplifiers used in the oscillator method have a zero phase shift and therefore the crystal must also have a phase shift of zero in order to satisfy the first oscillation condition.When the crystal is connected between the input and output of the amplifier, the loop gain is larger at the low frequency of zero phase and, hence, the second condition of oscillation is satisfied at this frequency. From eqn. (8), X/R = 0 when 8, = 0 and from eqns. ( 5 ) and (6), - B/G = X/R. Equation (4) divided by eqn. (3), with BIG set equal to zero, is a quadratic equation in o and its solution gives the two frequencies of zero phase: LO, called the series resonant frequency, the low frequency of zero phase and therefore the frequency measured by the oscillator method; and cop, called the parallel resonant frequency. If R, = 0, (9) where o,0 is o, when R, = 0 and similarly for oP0. However, Peristaltic Syringe injection in a liquid, assuming that the equivalent circuit is an adequate representation of the crystal, R, is of the order of magnitude of 103 8 hence eqn.(9) is not the frequency measured by the oscillator method. The most often used oscillator consists of two transistor- transistor-logic inverters connected in series to give a non- inverting amplifier (zero phase shift between input and output voltage) with the quartz sensor connected from the output to the input of the amplifier.l0,*3,'5,20 Oscillators with single transistors have also been used.7 A marginal oscillator has been used;17,*8 this is an oscillator with additional feedback applied internally which changes the gain of the oscillator amplifier in response to a change of energy dissipation in the quartz crystal, such that the amplitude of the output voltage of the amplifier remains constant.A wideband marginal ampli- fier is shown in Fig. 5. In order to carry out concurrent electrochemical measure- ments, two different experimental arrangements have been employed. The first is based on a stationary ~onfiguration2~27 in which an immersed quartz crystal in a housing is kept at a fixed position in the cell. The second is a recently developed elegant variant in which a rotating disc electrode was used to measure changes in mass.53 This type of device provides hydrodynamic conditions suitable for the suppression of polarization in experiments with dilute solutions. The oscillator method suffers from three kinds of limita- tions: (a) the method only partially characterizes a quartz crystal sensor because only one quantity, the series resonant frequency, is measured; ( b ) the resonant frequency depends on the capacitance in series with the crystal20 and in some instances on the type of oscillator used;29 and ( c ) the crystal does not oscillate when it is in solutions of high viscosity and if the crystal oscillates when the liquid is in contact with one electrode, it will often not oscillate if both electrodes are exposed to the liquid.Regarding limitation (a), more information can be extrac- ted from this method by measuring the output voltage of the oscillator amplifier,54 and the resonant frequency. The feed- back voltage of the marginal oscillator could also be measured. However, the characterization of the crystal is still incomplete.Limitation ( b ) can be understood in terms of the equivalent circuit of the crystal. A capacitor in series with the equivalent circuit (Fig. 4) will change the zero-phase frequen- cies. Different oscillators may have amplifiers with phase shifts that are not exactly zero and the oscillator circuit may have a capacitance which appears in series or in parallel with valve 0-ri'ng I / cry sta I I Voltage controller amplifier amplifier Amplitude Comparator I Reference voltage Fig. 5 Cell design for TSM sensor operation in liquid with wideband marginal oscillator and automatic gain control system. Also shown is the sensor incorporated into a liquid flow-through arrangement (reprinted by kind permission of the Institute of Electrical and Electronics Engineers)ANALYST, SEPTEMBER 1991.VOL. 116 887 the quartz crystal. Finally, the last limitation ( c ) is a fundamental deficiency of the oscillator method due to the fact that at high viscosities both resonant frequencies cease to exist because the phase shift of the crystal is always less than zero.55 Therefore, one of the conditions for oscillation (for an amplifier of zero phase shift) cannot be satisfied and, in principle, the circuit will not oscillate. The network analysis method is a passive method that has been recently developed by Kipling and Thompson55 that completely characterizes a quartz crystal sensor. This method can be used when a liquid of any viscosity is in contact with one or both electrodes of the sensor. Sinusoidal voltages incident on and reflected from the quartz crystal are measured repeatedly for a large number of frequencies in the resonant frequency range, that is, in the range that includes LO, and oP.(For sensors with f, and fp of the order of 107 Hz, f, - fp is of the order 104 Hz, wheref= 03/2n.) The experimental values of magnitude and phase of impedance, for example, can be calculated at each frequency from the measured voltages and then characteristic quantities can be found from the imped- ance-frequency curves. Some characteristic quantities from the phase measurements are: series resonant frequency; parallel resonant frequency; frequency of maximum phase and the value of maximum phase; and the slope of the phase curve at the series resonant frequency. The prominent characteristics from the impedance magnitude measurements are the frequencies at the minimum and maximum magnitudes of impedance and the corresponding values of impedance magnitudes.The theoretical values of 121 and 8, for the equivalent circuit are found by substituting eqns. (3) to (6) into eqns. (7) and (8). Muramatsu and co-workers22,23 used an impedance analyser to partially characterize a quartz crystal. This is a passive method in which, essentially, measurements are made of a voltage applied across the crystal and the current flowing through the crystal. They measured the maximum value of G and the frequency at which G is maximum. From inspection of eqn. ( 3 ) , the maximum value of G, G,,,, is G,,, = 1/R, and the frequency at G,,, is wSO, eqn. (9). They called this frequency the resonant frequency but it is not the resonant frequency measured by the oscillator method, LO,, unless R, = 0 which is certainly not true when the quartz crystal is immersed in a liquid.When R, f 0, LO, is larger than ws0 and cop is smaller than wPO. Both 03, - LO,,, and oPO - cop are proportional to R,2.52 Applications Detectors for Liquid Chromatography Use of the acoustic wave device as a universal mass detector for liquid chromatography was first suggested by Shulz and King.56 In reality these workers did not employ the sensor in an in situ configuration. Liquid samples from the eluent were simply sprayed onto a crystal surface. The first genuine flow-through design was published by Konash and B a ~ t i a a n s . ~ In this work a coated sensor was employed of which only one face was exposed to the liquid eluent.Although stable and mass sensitive detection was achieved, the system exhibited poor reproducibility. Oda and Sawadas7 incorporated a piezoelectric device in a flow cell as a photoacoustic detector in order to monitor chromatographic eluents, but the effects employed were associated with electroacoustic properties rather than piezoelectric operation. Finally, poly- (ethy1enimine)-coated crystals have been used for detecting various species in hydrocarbon solution,58 but in this study it was concluded that response time must be improved in order to cope with the short residence times involved in chromato- graphic detector cells. For the interest of the reader, we should note at this point that acoustic wave devices have been employed successfully as gas chromatographic detectors in a number of areas.5942 Determination of Inorganic Species Over the last several years, a single group, that of Nomura and co-workers,6.~-1~~12,15,63-69 has contributed significantly to the detection of a variety of inorganic species in the liquid phase using a TSM quartz crystal sensor.Generally these moieties were imposed at the sensor-liquid interface by the processes of electrodeposition or adsorption from aqueous solution. A summary of this work is presented in Table 3. Frequency dependence on liquid properties, such as specific conductance of electrolytes in solution70~71 has also been considered. Examples are the determination of micromolar concentrations of Hg" in waste water and microgram amounts of drugs containing iodine in biological material.Finally, Martin et al.72 studied the mass sensitivity of plate mode devices, modified with ethylenediamine ligands, for low concentrations of Cull ions in solution. Development of Biosensors It has been stressed a number of times in the recent literature that direct detection, possibly in an in situ manner, of analytical pairs such as antibody-antigen complexes and DNA hybrids could offer an attractive alternative to existing procedures. Not surprisingly, the appropriate possibilities for acoustic wave sensors in this area have been explored. Roederer and Bastiaans'l were the first to employ an acoustic wave device in an immunoassay procedure. A surface acoustic wave (SAW) sensor was immersed in an aqueous medium for the detection of an antibody by reaction with an antigen immobilized on the surface of the quartz. Although Table 3 Detection of inorganic species in aqueous solution using a quartz crystal sensor, from the work by Nomura and co-workers Coating on Concentration Analyte Au electrode Procedure range/pmol dm-3 Reference CN- CN- Ag' AE+ Ag' I- I- Pb2+ Pb2+ Fe3+ Fe3+, Al3+ Hg2+ cu2+ None Pt Pt Pt Ag on Pt Ag on Pt Pt Copper(I1) oleate Pt Silicone oil on Pt None Poly(viny1pyridine) Dissolution of Ag, 0.1-10 measurement in air Dissolution of Au 100-1 000 Electrodeposition 0.005-0.05 Internal 1-10 Electrodeposition 0.2-30 Electrodeposition 0.3-10 Electrodeposition 0.5-7 Adsorption 3-50 electrode position Absorption 3-40 Adsorption 10-100 Absorption 5-100 Adsorption 5-35 Electrodeposition 2-30 63 64 6 10 12 8 15 65 66 9 67 68 69888 ANALYST.SEPTEMBER 1991. VOL. 116 the specific binding of the complementary antigen was demonstrated, non-specific adsorption had taken place to a significant extent. Thompson et ul.17 were first to study interfacial immunochemistry by means of a TSM device in the liquid phase. An antigenic component was immobilized on an auxiliary thin film of polyacrylamide gel or directly on the crystal surface. However, the response of the device to the antibody was tentatively ascribed to interfacial perturbation of acoustic energy transmission rather than to a classical mass signal. In subsequent work the same group reviewed, on a qualitative basis, the diffusion-to-capture of interfacial immu- nochemical reactions, acoustic wave propagation through the liquid-solid interface and the characterization of piezoelectric crystal operation in water.18 It was postulated that shear wave devices are capable of generating two distinct types of analytical signals.Firstly, thin films characterized by a shear modulus of elasticity would provide a Sauerbrey-type mass measurement. Secondly, capture at a liquid-solid shearing surface could lead to a differential signal associated with the introduction of new material at the interface. The main idea is that acoustic wave devices operating in the liquid phase respond to a change in interfacial conditions, not to the absolute amount of added mass. A piezoelectric immunosensor for the detection of Cundidu ulbicuns microbes was developed by Muramatsu et ~1.21 Anti-Candidu antibody was covalently bonded on the plated platinum electrodes.The frequency of the crystal was recor- ded before and after dipping into a suspension of Candidu. The frequency change was observed and correlated with a concentration of Cundidu in the range 1 X 106-5 x 108 cells cm-3. Muramatsu et ~1.2’ also measured the response of AT-cut 9 MHz piezoelectric crystals to samples of human IgG under various operational conditions.22 Crystals were modi- fied by immobilizing Protein A on the oxidized palladium layer on the electrode surface with (3-aminopropyl)triethoxy- silane. Shifts in frequency were ascribed to the affinity reaction of Protein A and human IgG. Davis and Leary73 claimed to be able to continuously monitor the reaction of immunoglobulins with Protein A at the sensor surface.A frequency change of approximately 1 Hz for each 10 ng of added immunoglobulin was observed. They pointed out that because frequency decreases were demon- strated for added material at the sensor surface, the Sauerbrey mechanism must be correct. This was postulated despite the fact that no experiments confirming added material at the interface were mentioned. An indirect immunoassay involving polymer particles was developed by Kurosawa et u1.74 In their work the antibody-antigen reaction was carried out on latex and the frequency changes were regarded as being due to viscosity or density changes of bulk solution associated with aggregation of the latex particles.In addition to the TSM-device work outlined above, research has continued on the performance of immunoassays using the SAW device,75 in which the quartz surface of the sensor was etched and treated with (3-glycidyloxypropyl)- trimethoxysilane prior to immobilization of an antibody against influenza virus A. Frequency shifts were observed on exposure of the sensor to the virus. Richards and Bach76 have worked with DNA hybridization systems. In an ingenious experiment a DNA probe-type experiment was performed in which amplification of mass was obtained using iron oxide microparticles. These entities, bearing one of the reacting pairs, were attracted to the sensor surface by a magnetic field. Finally in this section, we should note that a number of studies have been performed in which the final bioanalytical signal has been obtained from the sensor in the gas phase, subsequent to reactions in solution.77-82 In view of the central theme of this review details of this work will not be outlined.Properties of Thin Films As has been the situation for the gas phase, a number of studies have been concerned with the in situ liquid behaviour of organic multilayers in the TSM sensor, both with respect to selective adsorption into such films and to their physical properties. Okahata et ~1.83 argued from studies involving adsorption of hydrophobic alcohols or cholesterol into various synthetic multilayer matrices, that the frequency response could not be correlated with density or viscosity changes of the film. The direct detection of the selective interaction between phospholipid or cholesterol multibilayer films cast on a piezoelectric crystal with cyclodextrins has been reported.84 Depending on their cavity sizes, cyclodextrins form soluble molecular-selective inclusion complexes through interaction with lipidic species. The cast films on crystal surfaces were stable and did not peel off the plate, even under harsh conditions in aqueous solution, which were confirmed by frequency observations of the crystal. Calibration showed that a decrease in frequency of 1 Hz corresponded to an increase in mass of 1.27 ng. Thickness-shear-mode acoustic wave devices have been used successfully to follow phase transitions in liquid crystals and lipid multilayers.85 Rajakovic et ~1.86 examined the role of the device-to-water acoustic interaction and, accordingly, the part played by interfacial viscosity in determining the frequency response.Frequency measurements in water for TSM sensors with gold, aluminium and silanized aluminium electrodes were obtained by the oscillator method. In that paper a number of aluminium electrode-based sensors were silanized using aminopropyltri- ethoxysilane and dichlorodimethylsilane and then exposed to water or aqueous solutions. The results showed that physical conditions at the polymeric silane-water interface can radiate the flow of acoustic energy into the surrounding medium as indicated earlier in this review. Conclusions It is evident from the progress reviewed in this article that the frequency response of the TSM device in liquids is governed by a number of factors.(Indeed the measurement of series resonant frequency produces a less than complete picture of the system.) Among these parameters, significant but hitherto unrecognized for the TSM sensor, is the role played by molecular slip and viscosity at the sensor-liquid interface. This observation opens up a number of new possibilities for application of the technique including, for example, the study of the physics of fluids at interfaces, surface structure of polymer films, extrusion phenomena, flow in porous media, lubrication and the development of a new type of signal for biosensor design. There is evidence in the literature that the above is now being recognized, at least with respect to the physical chemistry of liquids.87-90 Finally, we would like to emphasize that several of the interfacial arguments presented in this review have been discussed by other authors with respect to the different types of acoustic wave devices.In particular, readers are advised to consult the work of Ricco and Martin91 on plate devices and Diller and Frederick92 on torsional structures. The authors are grateful to the Natural Sciences and Engineer- ing Research Council of Canada for support of our work. Additionally, we appreciate the Fellowships provided to two of us, Lj. V. R. and B. A. C-V., by the Serbian Research Council of Yugoslavia.ANALYST. SEPTEMBER 1991, VOL. 116 889 APPENDIX Glossarv of svmbols used: , J Area of the quartz plate Numerical constant Numerical constant Numerical constant Liquid film thickness Numerical constant Numerical constant Numerical constant Numerical constant Numerical constant Numerical constant Stiffened elastic constant due to intrinsic viscosity of the quartz crystal Specific gravity of liquid Ratio of velocity amplitude at z and velocity Numerical constant Frequency change due to film Frequency change due to a liquid film Frequency change due to a solid film Resonant frequency of the quartz crystal with the Resonant frequency of the film amplitude at crystal surface, z = 0 film Fqm Resonant frequency of the quartz crystal without the Change of resonant frequency of the quartz crystal Height of liquid layer Real part of the propagation coefficient of the film Numerical constant Electromechanical coupling constant Imaginary part of the propagation coefficient of the film due to p-pm film Film thickness Quartz crystal thickness Mass per unit area of equivalent liquid layer Mass of the quartz crystal Mass of the film Change in mass due to solid film Number of sides of crystal in contact with liquid (n = Frequency constant of the specific crystal cut Pressure difference between the two sides of the quartz crystal Radius of the film Radius of the electrode Radius of the quartz crystal Resistance of equivalent circuit of quartz crystal Phase velocity of a shear wave in quartz One direction in a rectangular coordinate system Acoustic impedance of the film Acoustic impedance of the quartz Real part of the characteristic impedance of the film Imaginary part of the characteristic impedance of the Mean diameter of hemicylinders Dielectric constant of a liquid Dynamic viscosity of a liquid Specific conductivity of a liquid Shear modulus of quartz Density of the film Density of liquid Density of quartz Period change due to a solid film Phase angle by which the acoustic wave at z lags that at the crystal surface, z = 0 1 or 2) film 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References Sauerbrey, Ci., Z.Phys., 1959. 155, 206. King, W. H.. J r . , Anal. Chem.. 1964, 36, 1735. Mieure, J . P., and Jones, J . L . , Talantu, 1969, 16, 149. Konash, P. L., and Bastiaans. G . J . . Anal. Chem., 1980, 52, 1929. Nomura, R., and Minemura, A., Nippon Kagaku Kaishi, 1980, 1980, 1621. Nomura, T., and Iijima, M., Anal. Chim. Acta, 1981, 131, 97. Nomura, R., and Okuhara, M., Anal.Chim. Acta, 1982, 142, 281. Nomura, T., and Mimatsu, T . , Anal. Chim. Acta, 1982, 143, 237. Nomura, T., and Maruyama, M., Anal. Chim. Acta. 1983, 147, 365. Nomura, T., and Nagamune, T., Anal. Chim. Acta, 1983, 155. 231. Roederer, J . E., Bastiaans, G . J . , Anal. Chem., 1983,55,2333. Nomura, T.. and Tsuge, K., Anal. Chim. Acta. 1985, 169,257. Bruckenstein. S . , and Shay, M., Electrochim. Acta, 1985. 30, 1295. Kanazawa. K. K., and Gordon, J . G . , Anal. Chim. Acta. 1985, 175, 99. Nomura, T., Watanabe, M., and West, T . S . , Anal. Chim. Acta, 1985, 175, 107. Muramatsu, H., Kajiwara. K., Tamiya, E.. and Karube, I.. Anal. Chim. Acta, 1986. 188, 257. Thompson. M., Dhaliwal. G. K., and Arthur, C. L., Anal. Chem.. 1986. 58. 1206. Thompson, M ., Dhaliwal. G . K., Arthur, C. I,.. and Calabrese, G. S . , IEEE Trans. Ultranson. Ferroelec. Freq. Contr.. 1987, Calabrese. G. S., Wohltjen, H., and Roy. M. K., Anal. Chem.. 1987, 59, 833. Yao. S - Z . , and Mo. Z-H., Anal. Chim. Acta, 1987. 193. 97. UFFC-34, 128. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Muramatsu, H., Dicks, J . M., Tamiya. E., and Karube. 1.. Anal. Chem., 1987, 59. 2760. Muramatsu, H . , Tamiya. E., and Karube, I . , Anal. Chem., 1988, 60. 2142. Muramatsu. H., Tamiya, E., Suzuki, M., and Karube, I.. Anal. Chim. Acta. 1988, 215, 91. Schumacher, R., Pesek, J . , and Melroy, O., J . Phys. Chem.. 1985, 89. 4338. Schumacher, R.. Borges, G., and Kanazawa. K. K., Surj: Sci.. 1985. 163, L621. Bruckenstein, S . , and Swathirajan, S., Electrochim.Acta, 1985. 30. 851. Stockel, R . , and Schumacher, R., Be .. Bunsenges. Phys. Chem.. 1987, 91. 345. Beck, R., Pittermann, U.. and Weil, K. G . , Ber. Bunsenges. Phys. Chem., 1988. 92. 1363. Schumacher, R., Angew. Chem. Int. Ed. Engl., 1990, 29. 329. Applications of P iezoelectric Quartz Crystal Microbalances, eds. Lu, C-S.. and Czandcrna, A. W.. Elsevier, Amsterdam, 1984. Stockbridge. C. D . , in Vacuum Microbalance Techniques, ed. Behrndt, K. H., Plenum, New York, vol. 5 , 1966, p. 193. Miller, J. G.. and Bolef, D . I., J . Appl. Phys.. 1968, 39. 4589. Miller. J . G . , and Bolef, D. I., J . Appl. PhyJ., 1968. 39, 5815. Behrndt, K. H . . J . Vac. Sci. Technol., 1971, 8, 622. Lu. C - S . , and Lewis. 0.. J . Appl. Phys., 1972. 43, 4385. Lu, C-S., J .Vac. Sci. Tech.. 1975, 12, 578. Glassford, A. P. M., in Progress in Astronautics and Arronaut- ics, ed. Smith, A. M., American Institute of Aeronautics and Astronautics, New York, 1977, vol. 56, p. 175. Glassford, A. P. M., J . Vac. Sci. Technol.. 1978, 15, 1836. Mecca, V.. and Bucur, R. V., Thin Solid Films, 1979, 60, 73. Crane. R . A., and Fischer, G.. J . Phys. D., 1979, 12, 2019. Guilbault, G . G.. in Applicutions of Piezoelectric Quart? Crystal Microbalances, eds. Lu, C-S.. and Czanderna. A . W., Elsevier, New York, 1984, vol. 7, p. 251.890 ANALYST, SEPTEMBER 1991. VOL. 116 42 Heusler, K. E., Grzegorzewski, A., Jackel, L., and Pietrucha, J., Ber. Bunsenges. Phys. Chem., 1988, 92, 1218. 43 Hager, H. E., and Verge, P. D., Sens. Actuators, 1985,7,271.44 Hager, H. E., Chem. Eng. Commun., 1986, 43, 25. 45 Yao, S-Z., and Zhou, T-A., Anal. Chim. Acta, 1988, 212, 61. 46 Shana, Z. A., Radtke, D. E., Kelkar, U. R., Josse, F., and Haworth, D. T., Anal. Chim. Acta, 1990, 231, 317. 47 Blake, T. D., Coll. Surf., 1990, 47, 135. 48 Duncan-Hewitt, W. C., and Thompson, M., Anal. Chem., submitted for publication. 49 Wiggins, P. M., Progr. Polym. Sci., 1988, 13, 1. SO Krausz, A. S., and Eyring, H., Deformation Kinetics, Wiley, New York, 1974. 51 Cady, W. G., Piezoelectricity, Dover, New York, 1964. 52 Bottom, V. E., Introduction to Quartz Crystal Unit Design, Van Nostrand Reinhold, New York, 1982. 53 Grzegorzewski, A., and Heusler, K., J. Electroanal. Chem. Interfacial Electrochem., 1987, 228, 455. 54 Mecea, V., and Bucur, R.V., J. Phys. E., 1974, 7, 348. 55 Kipling, A. L., and Thompson, M., Anal. Chem., 1990, 62, 1514. 56 Schulz, W. W., and King, W. H., J. Chromatog. Sci., 1973,11, 343. 57 Oda, S., and Sawada, T., Anal. Chem., 1981,53, 471. 58 Charlesworth, J. M., Anal. Chem., 1990, 62, 76. 59 Wohltjen, H., and Dessy, R., Anal. Chem., 1979, 51, 1465. 60 Heckl, W. M., Marassi, F. M., Kallury, K. M. R., Stone, D. C., and Thompson, M., Anal. Chem., 1990, 62, 32. 61 Thompson, M., and Stone, D. C., Anal. Chem.. 1990,62,1897. 62 Grate, J. W., Snow, A., Ballantine, D. S. Jr., Wohltjen, H., Abraham, M. H., McGill, R. A., and Sasson, P., Anal. Chem., 1988,60, 869. 63 Nomura, T., and Hattori, O., Anal. Chim. Acta, 1980,115,323. 64 Nomura, T., Anal. Chim. Acta, 1981, 124, 81. 65 Nomura, T., Yamashita, T., and West, T. S., Anal. Chim. Acta, 1982,143,243. 66 Nomura, T., Okuhara, M., and Hasegawa, T., Anal. Chim. Acta, 1986, 182, 261. 67 Nomura, T., and Ando, M., Anal. Chim. Acta, 1985, 172,353. 68 Nomura, T., and Fujisawa, M., Anal. Chim. Acta, 1986, 182, 267. 69 Nomura, T., and Sakai, M., Anal. Chim. Acta, 1986, 183, 301. 70 Shouzhuo, Y., and Zhihong, M., Anal. Chim. Acta, 1987, 193, 97. 71 Shouzhuo, Y., and Lihua, N., Anal. Proc., 1987, 24,336. 72 Martin, S. J., Ricco, A. J., Niemczyk. T. M., and Frye, G. C., Sens. Actuators, 1989, 20, 253. 73 Davis, K. A., and Leary, T. R., Anal. Chem., 1989. 61, 1227. 74 Kurosawa. S., ‘Tawara, E., Kamo, N., Ohta, F., and Hosokawa, T.. Chem. Pharm. Bull., 1990, 38, 1117. 75 Bastiaans, G. J., US Patent 4735906 A, 1988, April 5th; Chem. Abstr.. 1988, 109, 69931r. 76 Richards, J. C., and Bach, D. T., Eur. Pat. Appl., EP 295965 A2. 1988, December 21st; Chem. Abstr., 1988, 111, 150067b. 77 Ebersole, R., and Ward, M. D.,J. Am. Chem. SOC., 1988,110, 8623. 78 Grande, L. H., Green, C. R., and Paul, D. W., Sens. Actuators, 1988, 14, 387. 79 Prusak-Sochaczewski, E., and Luong, J. H. T., Anal. Lett., 1990, 23, 183. 80 Shous, A., Dorman, F., and Najarian, J . , J. Biomed. Muter. Res., 1972, 6. 565. 81 Oliveria, R. J., and Silver, S. F., US Patent, 4242096, 1980. December 2nd. 82 Fawcett, N. C., Evans, J. A., Chien, L-C., and Flowers, N., Anal. Lett., 1988, 21, 1099. 83 Okahata, Y., Ebato, H., and Ye. X., J. Chem. SOC., Chem. Commun.. 1988, 1037. 84 Okahata, Y., and Ye, X., J. Chem. SOC., Chem. Commun., 1988, 1147. 85 Okahata, Y., and Eeato, H., Anal. Chem., 1989,61, 2185. 86 RajakoviC, Lj. V., CaviC-Vlasak, B. A., Ghaemmaghami. V., Kallury, K. M. R., Kipling, A. L., and Thompson, M., Anal. Chem., 1991,63, 615. 87 Orata, D. O., and Buttry, D. A., J. Electroanaf. Chem. Interfacial Chem., 1988, 257, 71. 88 Lee, P. C. Y., Guo, X., and Tang, M. S. H., J. Appl. Phys., 1988, 63, 18.50. 89 Furukawa, S., Nomura, T., and Yasuda, T., J. Phys. D, 1989, 22, 1785. 90 Lasky, S. J . , and Buttry, D. A., J. Am. Chem. SOC., 1988,110, 6258. 91 Ricco, A. J., and Martin, S. J., App. Phys. Lett., 1987,21,1474. 92 Diller, D., and Frederick, N. V., Int. J. Thermophys., 1989,10, 145. Paper 11015566 Received April 3rd, 1991 Accepted May 7th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600881

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1991, VOL. 116 891 On-line Immunoaffinity Sample Pre-treatment for Column Liquid Chromatography: Evaluation of Desorption Techniques and Operating Conditions Using an Anti-estrogen Immuno-precolumn as a Model System Aria Farjam, Anita E. Brugman, Henk Lingeman" and Udo A. Th. Brinkman Department of Analytical Chemistry, Free University, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands An i m mu noaff i nity precol u m n (i mm u no-precolu m n) containing an immobilized anti body directed against estrogen steroids, was used as a model system for the evaluation of different desorbing techniques, suitable for on-line coupling to column liquid chromatography (LC). Desorption of estrogen analytes from the immuno-precolumn proved to be impossible with the buffers and chaotropic solutions frequently used in affinity desorption. Micellar solutions are effective in obstructing the antibody-antigen reaction, but their use as desorbing solutions was not found to be practical because of the large interferences introduced into the chromatograms.Desorption with aqueous solutions at elevated temperature, created by microwave action or conventional heating, although effective is not practical in this instance, because the agarose used in this study as the stationary phase for the immuno-precolumn is prone to heat decomposition. The most effective and practical approach is desorption with a methanol-water mixture (95 + 5 v/v). On-line dilution of the eluate allows the concentration of the desorbed analytes using a reversed-phase LC system with subsequent separation and ultraviolet detection.The performance of the system with spiked urine and plasma samples, which were introduced directly into the system, was evaluated and the results were compared with i m mu noselective desorption. Keywords : Im m u n o a ffin it y p reco lu m n ; column switching; colu rn n liquid ch-ro m a tog rap h y; estrogen; urine and plasma samples For the determination of trace levels of analytes in complex samples, a highly selective sample pre-treatment is necessary. In this field immunoaffinity chromatography is a powerful technique. Antibodies raised against the target analytes when immobilized on a solid support and packed into a precolumn, can be used for the selective solid-phase extraction of the analytes. If this technique is combined with chromatographic separation and subsequent detection, 'a powerful analytical system is obtained, which combines the positive features of both techniques.The good group selectivity of the antibody is combined with the efficiency of chromatography in order to discriminate and quantify several analytes in a single run. Full automation of such analytical systems can be achieved by coupling the immuno-precolumn to column liquid chroma- tography (LC) with a column switching system.l-10 The critical point in such an immunoaffinity LC system is the on-line desorption of the analytes from these precolumns. As immunosorbents often have a relatively low capacity for the analytes of interest, in most instances large volume immuno- precolumns (up to 1 ml) have to be used for effective clean-up and preconcentration of the samples. Quantitative desorption of the analytes from the large volume immuno-precolumn must be combined with an efficient refocusing of the analytes on the reversed-phase LC system.Moreover, desorption should neither de-activate the immunosorbent irreversibly nor introduce impurities into the system. Also, it would be of great benefit, if the desorbing technique were not just suitable for a single antibody, but applicable to a wide range of different antibody-analyte systems. In previous studies an immunoselective desorbing tech- nique has been employed,7-9 in which the desorbing solution contains one or more cross-reacting compounds. Because of their affinity for the antibody, these compounds are able to displace trapped analytes from the immuno-precolumn.Using this technique almost totally aqueous solutions could be used * To whom correspondence should be addressed. for desorption, the eluate could be used directly for precon- centration of the analytes on a reversed-phase precolumn. Another on-line .desorbing technique was employed with an anti-afl atoxin immuno-precolumn. 10 A methanol-water mix- ture (70 + 30 v/v) was used to desorb the immuno-precolumn. After desorption the eluate was diluted on-line with water in order to allow concentration of the analytes of interest using a reversed-phase LC column. In this study the performance of several desorbing tech- niques for on-line immunoaffinity sample pre-treatment coupled to LC are evaluated.An anti-estrogen immuno- precolumn is used as a model system. Immunoselective desorption,7-9 non-selective desorption with . methanol,10 desorption with so-called chaotropic solutions (solutions containing compounds that disrupt the water structure), with anionic and cationic micellar solutions and at elevated temperature are compared. The practicability of the on-line methanol desorption is finally demonstrated with spiked urine and plasma samples. The possibility of using this technique with other antibody-analyte systems is also discussed. .Experimental Reagents . The anti-estrogen antiserum was a gift from Organon (Oss, The Netherlands). Cyanogen bromide (CNBr) activated Sepharose 4B was obtained from Pharmacia (Woerden, The Netherlands).Liquid chromatography grade acetonitrile and methanol were obtained from Baker (Deventer, The Nether- lands) and Merck (Darmstadt, Germany), respectively. Liquid chromatography grade water was prepared from demineralized water using a Milli-Q (Millipore, Bedford, MA, USA) water purification system, with subsequent filtration over an LC column filled with 40 ym Baker CI8 bonded phase material. The eluents were de-gassed under vacuum in an ultrasonic bath. The steroids estrone (El), fi-estradiol (E2), @-estriol (E3), ethynylestradiol (EE) and892 ANALYST, SEPTEMBER 1991, VOL. 116 17-fi-estradiol-17-acetate (E2Ac) were obtained from Sigma (St. Louis, MO, USA). Stock solutions of the steroids were prepared in ethanol (1 mg ml-1) and stored at 4 "C until they were used.The E3 stock solution was purified by preparative LC. One milligram was dissolved in 1 ml of an acetonitrile-water mixture (40 + 60 v/v) and injected onto a 250 X 10 mm i.d. Baker CI8 bonded phase column for preparative LC (40 pm particles). Tetrahydrofuran-acetonitrile-water (5 + 20 + 75 v/v) was used as the eluent. All of the other chemicals were from various sources, and of analytical-reagent grade. Apparatus The set-up of the LC system used for the non-selective desorption is schematically shown in Fig. 1. It consisted of a Kratos (Ramsey, NJ, USA) Model 400 pump (pump 1); two Kontron (Zurich, Switzerland) Model 410 pumps (pumps 2 and 3) equipped with laboratory-built membrane pulse dampers; a Kontron MCS 670 tracer valve switching unit; and a Kontron Model Anacomp 200 programmer, which con- trolled the high-pressure switching valves (Vl-V3), the solvent selection valve and the flow of pump 1.A Kratos Model 757 Spectroflow ultraviolet (UV) detector set at 280 nm was used together with a Kipp & Zonen (Delft, The Netherlands) BD 40 recorder for peak area measurements. A coiled i6 in stainless-steel capillary with a 0.5 mm i.d., an internal volume of 1 ml, and a helix of 5 cm was used as the mixing coil. The analytical column was a 100 x 3.0 mm i.d. stainless-steel column packed in-house with 5 pm LiChrosorb RP-18 from Merck. This column was protected with a 10 x 2.0 mm i.d. stainless-steel guard column packed with the same material. The 20 x 3.0 mm i.d. CI8 precolumn was packed with 40 pm Baker C18 bonded phase material.The immuno-precolumns were laboratory-built 10 x 10 mm i.d. or 10 X 4.0 mm i.d. stainless-steel columns equipped with 5 pm stainless-steel screens, and with polytetrafluoroethylene (PTFE) rings as column inlets and outlets. The packing procedure of the immuno precolumns is described elsewhere .7 The mobile phase for the LC was an acetonitrile-water mixture (65 + 35 v/v) with 10% v/v tetrahydrofuran at a flow rate of 0.5 ml min-1. Non-selective Desorption With Aqueous Methanol The final procedure developed in this study was comprised of the following steps (see also Fig. 1): 1, flushing the immuno- precolumn with 15 ml of water; 2, loading the sample on the immuno-precolumn; 3, flushing the immuno-precolumn with 15 ml of water; 4, flushing the CIS precolumn with 6 ml of water; 5, desorbing the immuno-precolumn with 3.7 ml of f Precolumn Pl I I Wa'ter cia Separation column' 95% Methanol & Water Water Sample c Pump 3 + Mobile phase UV Detector i 280 nm Fig.1 Set-up of the on-line immunoaffinity liquid chromatographic system used for non-selective desorption. The system is described under Apparatus methanol-water (95 + 5 v/v). (The eluate is diluted on-line via the T-piece with a 16-fold excess of water and the analytes are subsequently concentrated on the CIS precolumn and, then, separated on the C18 separation column); and 6, flushing the immuno-precolumn with 5 ml of methanol-water (95 + 5 v/v). The flushing volumes indicated for steps 1, 3, 5 and 6 were used with a 10 x 10 mm i.d. immuno-precolumn. If immuno-precolumns of a different size were used, the flushing volumes were adapted correspondingly.Immunoselective Desorption For immunoselective desorption the set-up in Fig. 1 was changed by removing pump 2, the T-piece and the mixing coil that connects V2 to V3, such that the C18 precolumn could be reconditioned with water by pump 1 (cf. Fig. 1 in reference 9). The analytical procedure is essentially the same as that with the non-selective methanol desorption, except that desorption (step 5 ) is performed with 60 ml of acetonitrile-water (5 + 95 v/v) containing 260 pg 1-1 of each of E2 and E2Ac as displacers. The immuno-precolumn and the CIS precolumn are switched in series during this step. Flushing of the immuno-precolumn was performed with 20 ml of methanol- water (70 + 30 v/v).The flushing volumes indicated were used with a 10 x 10 mm i.d. immuno-precolumn. For immuno-precolumns of a different size the flushing volumes were adapted correspond- ingly. Microwave Desorption The immunosorbent was packed into a 40 x 5.2 mm i.d. glass column and placed in a 200 ml water-bath inside a microwave oven, Model H 2500 (Bio-Rad, Richmond, CA, USA). The temperature sensor of the microwave oven was put in the water-bath. The immuno-precolumn was connected to the column switching system via two 0.5 mm i.d. PTFE capillaries. The PTFE outlet capillary was coiled over a length of 50 cm and inserted into a 20 "C water-bath, which was outside the microwave oven, in order to cool the eluate to room temperature after desorption.Analysis was performed according to the immunoselective procedure, except that no displacer was present in the desorbing solution. The micro- wave oven was programmed to raise the temperature to the pre-set value at the start of the desorption (step 5). Usually this was achieved within 20 s. After desorption, the immuno- precolumn was immediately placed into the 20 "C water-bath outside the oven. Immunosorbent The preparation of the anti-estrogen immunosorbent on the CNBr-activated Sepharose 4B has been reported in a previous study.3 If not indicated otherwise, all of the quantitative measurements were performed with immuno-precolumns that had already been used for at least ten runs so as to obtain a constant capacity. The capacity of the 10 x 10 mm i.d. immuno-precolumn was determined by overloading it with 30 ml of a standard solution containing 66 pg 1-1 of each of EE, El and E2.The analysis was performed according to the non-selective desorption procedure. Urine and Plasma Samples Urine samples Samples were collected from three healthy male volunteers, they were then pooled, divided into 100 ml fractions and stored at -20 "C. The samples were thawed 1 h before use.ANALYST, SEPTEMBER 1991, VOL. 116 893 50 N 0 ,- 30 X \ 10 0 2 4 6 8 10 12 Diluted plasma/ml Fig. 2 Pressure as a function of the amount of plasma pumped over different types of screens. Unfiltered plasma, which had been diluted 4-fold with water, was pumped over different types of screens that had been inserted into a in Valco female-female connector: flow rate, 2 ml min- I .A, a polyester screen with 20 pm pores; B, a stainless-steel screen with 7 pm pores; C, a nylon screen with 20 pm pores; D, a stainless-steel screen with 13 pm pores; and E. a stainless-steel screen with 36 pm pores were compared Analysis of plasma samples The plasma samples from healthy volunteers were a gift from the State University of Utrecht (Utrecht, The Netherlands). The samples were divided into 10 ml fractions and stored at -20 "C. The samples were thawed 1 h before use. The set-up for the automated analysis contains a stainless- steel screen (pore size, 7 pm), which is incorporated into a Valco female-female union positioned directly after the pump used for loading the samples. This screen has a filter function and is replaced if the pressure in the system becomes too high; usually once a week.The screen also helps to prevent blocking of the switching valves and the expensive immuno-precolumn. However, with plasma analyses, this screen tends to block after the passage of only small volumes of diluted plasma (see Fig. 2). In order to eliminate this problem, various types of screen were tested. As can be seen in Fig. 2, the 36 pm stainless-steel screen gave the best results. Therefore, this screen was used in the systems if plasma samples were to be analysed. Results and Discussion Desorption with Aqueous Solutions In classical immunoaffinity chromatography, buffers of either high or low pH or aqueous solutions containing chaotropic compounds are used for breaking antibody-antigen interac- tions, i.e., for desorption. Desorption of protein antigens from immunoaffinity columns using these techniques has frequently been reported. 1 In principle, such solutions are a good choice for on-line desorption coupled to reversed-phase LC, as concentration of the analytes can be performed efficiently on a reversed-phase column from the purely aqueous eluate. By using the anti-estrogen immuno-precolumn as a test system, a number of conventional desorbing solutions were investi- gated. An aqueous solution containing the desorbing agent to be tested was used to preconcentrate the analytes of interest on to an immuno-precolumn. Desorption and determination of the trapped analytes were performed by means of immunoselec- tive desorption (see Experimental). In order to define a good desorber, the amount of analyte trapped on the immuno- precolumn should be close to zero.The amount of the analytes EE, E l and E2, trapped on the immuno-precolumn, if loaded from different desorber-containing solutions is shown in Fig. 3. After each measurement, as a control, the analytes were also loaded from a purely aqueous solution, in order to monitor a decrease of immuno-precolumn capacity. With Fig. 3 Recovery of the estrogens EE, El and E2. if loaded from solutions containing the indicated desorbing agents. In all instances 4.8 ml of a solution containing 10.5 pg 1-' of each estrogen dissolved in the desorber-containing solutions or in water (control), were loaded on to a 10 X 4 mm i.d. immuno-precolumn. The desorber solutions were 0.1 mol dm-3 glycine-HCI buffer.pH 2 (Glyc); 0.270 ( m h ) trifluoroacetic acid (TFA); 0.2% (m/m) sodium sulphate plus 0.2% ( m h ) TFA, p H . 2 (Na2S04 + TFA); 3.4 mol dm-3 sodium dodecylsulphate (SDS); 3.1 mol dm-3 hexadecyltrimethylammonium chloride (cetrimide); 4 mol dm-3 urea (urea) and 0.5% (m/m) HCI (HCI). Analysis was performed by means of immunoselective desorption most of the desorbing solutions tested the recoveries were rather close to the recoveries of the control experiments, especially with El and E2. The 'best' deactivation of the antibody-estrogen interaction was obtained using hexadecyl- trimethylammonium chloride (cetrimide) with recoveries ( n = 3) for EE, El and E2 of 14, 19 and 27%, respectively. Obviously, none of the solutions tested efficiently prevents the binding of the analytes to the immobilized antibodies or, alternatively, in the proposed LC system, disrupt these bonds.The buffer and chaotropic solutions, which work well for the desorption of proteins in conventional immunoaffinity chromatography, are obviously not generally suitable for the desorption of small molecules (see also references 7 and 12). The most probable explanation is, that in conventional immunoaffinity chromatography these desorbing solutions interact with the sorbed protein, i.e., change its structure, much more than with the immobilized antibody, which, owing to its multipoint attachment to the stationary phase, is more resistant to structural changes than the free protein. As small molecules are mostly not prone to structural changes as a result of the action of the buffer or chaotropic solutions, their desorption from the immobilized antibody is difficult. If, however, the structure of a small molecule can be changed by a buffer or chaotropic agent, desorption will occur.Such a situation is typically encountered when the small analyte possesses an acidic o r basic group. Desorption can then be achieved with a buffer that changes the analyte structure by either protonation or deprotonation. 13-15 Desorption with Micellar Solutions The experiments described above included the use of an anionic sodium dodecyl sulphate (SDS) and a cationic (cetrimide) surfactant. It is known that the physico-chemical properties of surfactants can differ significantly below and above their critical micellization concentration (c.m.c.) In order to determine the influence of the surfactants on the antibody-antigen interaction as a function of their concentra- tion, the estrogens were loaded onto the immuno-precolumn894 Type of heat action 116 ( i ) Microwave J J J (ii) Water-bath J J 0 10 20 3 0 4 0 5 0 0 1 2 3 4 5 6 SDS/mmol dm-3 Cet ri m ide/mmol dm - Fig.4 Recovery as a function of the amount of surfactant added to the sample solution. Standard solutions (30 ml) containing 2.5 pg 1-1 of each estrogen: A, EE; B, E l ; and C, E2 and various amounts of the surfactants (a) sodium dodecylsulphate (SDS) and (b) hexadecyl- trimethylammonium chloride (cetrimide) were loaded on a 10 x 10 mm i.d. immuno-precolumn. Analysis was performed according to the non-selective desorption technique.Where c.m.c. = critical micellization concentration ANALYST, SEPTEMBER 1991, VOL 1 TemperaturePC I 20 55 55 55 55 biiiivolumefdes&ber/ml 1 60 60 30 30 30 from solutions containing different concentrations of surfac- tant. In contrast with the earlier experiments, methanol-water (95 + 5 v/v) was now employed for desorption (non-selective). This was to prevent secondary effects caused by the cross- reacting steroids present in the immunoselective desorbing solution. As can be seen in Fig. 4, estrogen recoveries (n = 2) close to 100% were achieved, if the surfactants were present in the sample solutions at concentrations below their c.m.c. At their c.m.c. about 8 mmol dm-3 for SDS and 1 mmol dm-3 for cetrimide, a sharp decrease of analyte recovery was observed.At concentrations of about five times the c.m.c. the recovery (n = 2) of all of the steroids was less than 20%. Obviously the presence of either cationic or anionic micellar species prevents binding of the analytes on to the immobilized antibody. The decrease in the recovery was significantly more rapid with SDS than with cetrimide. Micellar solutions, therefore, seem to be good candidates for desorption of analytes from the immuno-precolumn. When the solutions were used for desorption, however, serious problems arose during analyte preconcentration on the CI8 precolumn and subsequent separation. If the analytes were preconcentrated on the CI8 precolumn from micellar cetrimide solutions (containing concentrations five times greater than the c.m.c.of the surfactant) and subsequently separated, large interfering peaks occurring at the retention times of the analytes prevented quantification. If the exper- iments were performed with solutions containing SDS, the recovery of the analytes dropped to zero. Obviously the micellar solutions change the property of the analytical system such that efficient preconcentration and separation of the analytes is not possible. Thermodesorption Antibody-antigen interactions become weaker at higher temperatures. Therefore desorption from the immuno-pre- column at elevated temperature was investigated. The immu- nosorbent was packed into a 5.2 X 40 mm i.d. glass precolumn, which was placed in a microwave oven. Standard samples (50 ml) spiked with 5 pg 1-1 each of EE, El and E2 were loaded onto the immuno-precolumn.The analyses were carried out according to the immunoselective procedure, except that no displacer was added to the desorbing solution. Instead, the microwave oven was programmed to raise the temperature at the start of the desorption and keep it constant during the desorption process. Analyte recovery was found to improve with increasing temperature. The recoveries (n = 2) of EE, El and E2 with room temperature desorption were 53, 12 and 22% , respect- ively (cf. Fig. 5 , column 1) and at 55 "C, the highest temperature tested, were 86, 43 and 69% (cf. Fig. 5 , column 2). In both instances 60 ml of acetonitrile-water (5 + 95 v/v) Fig. 5 Recovery as a function of thermodesorption conditions. A standard sample (15 ml) containing 5 pg 1-1 of each of the three estrogens EE.E l and E2 was loaded onto a 40 x 5.2 mm i.d. immuno-precolumn (column volume, 850 pl). Analysis was per- formed according to the immunoselective desorption technique, but no displacer was present in the desorbing solutions were used for desorption. Some acetonitrile was added to prevent non-specific retention of the analytes on the immuno- precolumn. In an earlier study,9 the use of 12 instead of 5% acetonitrile was found to enhance the desorption at room temperature. When using microwave desorption at 55 "C however, no significant difference in steroid recovery was found for desorbing solutions containing 5 or 12% acetonitrile (cf. Fig. 5 , columns 3 and 4). Obviously the microwave action is the dominant desorbing effect.Reducing the volume of solution used during microwave desorption at 55 "C from 60 ml (Fig. 5 , column 2) to 30 ml (Fig. 5 , column 3) of acetonitrile-water (5 + 95 v/v) gave only a minor decrease in recovery. This indicates that most of the analytes are desorbed in the early part of the process. In order to determine whether the microwave action, or solely the heat generation as such, was responsible for desorption, a control experiment was set up in which the immuno-precolumn was inserted into a 55 "C water-bath during desorption. Under similar conditions no difference in recovery was found between microwave and water-bath heat generation (Fig. 5 , columns 4 and 5 ) . After ten 'temperature runs', the decrease in the capacity of the immuno-precolumn was about 20% , indicating that, with this technique, repeated use of the immuno-precolumn is possible only for a limited number of analyses.The loss of capacity is probably not due to irreversible heat de-activation of the antibodies, but to decomposition of the agarose backbone: it is well known that agarose-based stationary phases decompose at elevated temperature. For this reason temperatures higher than 55 "C have not been tested. As this study was restricted to agarose-based immunosorbents, further use of thermodesorption, which will be more efficient with temperature-resistant phases, was not investigated here. Non-selective Desorption With Aqueous Methanol Another technique that can be employed for on-line desorp- tion is non-selective elution with methanol-water (70 + 30 v/v), an alternative that is also being used in on-going work on anti-aflatoxin immuno-precolumns.10 In this technique, after desorption the high methanol content of the eluate is reduced by on-line dilution with water, in order to allow concentration of the analytes on a reversed-phase column.A 16-fold dilution of the methanol-water (70 + 30 v/v) with water will result in a final methanol percentage of about 4%, a value sufficiently low for efficient trapping of the estrogens on the C18 precolumn. With a 10 x 4.0 mm i.d. immuno-precolumn, 600 p1 of the methanol-water (70 + 30 v/v) were sufficient toANALYST, SEPTEMBER 1991, VOL. 116 895 obtain a good, viz., over SO%, recovery (n = 3) of all three analytes. The calibration graph, recorded by loading the steroids from a 5 ml standard sample, was linear up to individual amounts of steroid of 30 ng (total amount of steroid, 90 ng).Initially, a rather persistent memory effect was observed for El and E2, although, surprisingly, not for EE. A partial solution for this problem was obtained by flushing the immuno-precolumn with methanol-water (70 + 30 v/v) containing various electrolytes or surfactants. Further work showed that the memory effect was completely eliminated upon desorption with 95% instead of 70% methanol. This procedure was used in all further work. A volume of 3.7 ml of methanol-water (95 + 5 v/v) was chosen for desorption of the analytes from the 10 X 10 mm i.d. immuno-precolumn. The eluate was again diluted on-line with a 16-fold excess of water, before concentration on the C18 precolumn.Owing to the higher methanol concentration, flushing of the immuno-precolumn after desorption (step 6 of the final procedure) was performed with 5 instead of 20 ml of solution. The recoveries for EE, E l and E2, if loaded from 30 ml of standard solutions, were 84, 84 and 90% , respectively. The recovery was constant [a relative standard deviation (RSD) of 2%; n = 41 for up to 100 ng of each steroid (total amount, 300 ng). For larger amounts the recovery slowly decreased, owing to the limited capacity of the immuno- precolumn. If the total amount of steroid did not exceed the immuno-precolumn capacity, good recoveries could also be achieved with larger sample volumes. For example, the recoveries for all three steroids were still over 80% if they were preconcentrated from 500 ml of standard solution containing 75 ng of each steroid.Stability of the immuno-precolumn The long-term stability of the 10 x 10 mm i.d. immuno- precolumn was assessed by measuring its capacity at regular intervals. The capacity had dropped by about 50% after 36 d of operation as can be seen in Fig. 6. During this time about 100 analyses were run; 20 of these were on urine samples. It is interesting that most of the capacity drop occurred during the days when only analyses of standards were performed (days 1-10). After that period of time, the capacity for EE, El and E2 reached plateau values of about 120, 300 and 180 ng, respectively. Obviously, it is not the urine analyses that are responsible for this loss of capacity.The initial capacity decrease, which has also been observed in other work,7-*03*5 occurs because, during the early runs, any ‘fragile’ antibody molecules, i.e., those improperly immob- ilized or easily subject to irreversible structural changes, will be lost. The extent of this capacity drop mainly depends on the immobilization chemistry and on the characteristics of the antibody used. With an anti-nortestosterone immuno-pre- 1000 I 1 I 1 0 10 20 30 40 Ti m e/d Fig. 6 Long-term stability of the 10 X 10 mm i.d. immuno- precolumn. expressed as its capacity for the estrogen: A, EE; B, El; and C, E2. Day zero: first day of use column a capacity drop of 70’%0~.~ was found before reaching plateau conditions. With a commercially available anti-afla- toxin immuno-precolumn the capacity drop was over 90% .lo Analysis of Biological Samples Using Non-selective Desorption Spiked urine samples A chromatogram of a male volunteer’s urine sample spiked with 5.1 pg 1-1 of each of EE, E l and E2, and of the corresponding blank urine are shown in Fig.7(a). Fifteen millilitres of sample were diluted with 15 ml of water and loaded on to the 10 x 10 mm i.d. immuno-precolumn. The recoveries for EE, E l and E2 were 70,67 and 77% with RSDs of 4, 3 and 4 (n = 4). These values are somewhat worse than for standard samples, but better than for the plasma samples. Under these conditions the detection limit of the method is 0.2-0.5 pg 1-1 (signal-to-noise ratio = 3 : 1). The detection can be improved by processing larger volumes of urine.It is worth noting that by analysing 75 ml instead of 15 ml of the same sample, as in Fig. 7(a), gave an identical blank chromatogram. That is, detection limits of 0.1 pg 1-1 or below are easily attainable. Capacity measurements performed before and after 20 urine analyses showed that the immobilized antibo- dies were not affected by the urine samples. Spiked plasma samples A chromatogram of a plasma sample spiked with 12.5 pg 1-1 of each of EE, E l and E2, and of the corresponding blank plasma, is shown in Fig. 7(b). Six millilitres of sample were diluted with 24 ml of water and loaded onto the 10 x 10 mm i.d. immuno-precolumn. Unfortunately, a peak at the reten- tion time of E2, corresponding to an amount of about 25 ng, was found in all of the blank plasma runs.As the sample was from a male volunteer, this peak cannot be due to E2, but it must be another unknown compound. The recoveries for EE, E l and E2 were 50,64 and 71% with RSDs of 7,lO and 10 ( n = 4) for peak area measurements. The immuno-precolumn showed a capacity decrease of about 30% after the analysis of ten plasma samples. Obviously, plasma components, e.g., proteolytic enzymes, irreversibly de-activate the immuno- precolumn. Conclusions With the three estrogen steroids as test compounds, various immuno-precolumn desorption techniques to be used in on-line precolumn and LC systems have been evaluated. a, C ID 2 2 a n 0.0005 a.u.f.s. a’ I I 10 5 ”!I 0.0005 a.u.f.s. 0 10 5 0 Time/min Fig. 7 Chromatograms of ( a ) a male urine sample spiked with 5.1 yg I-’ of each of the estrogens: EE, E l and E2 and (b) a male plasma sample spiked with 12.5 pg I-* of each of the three estrogens, and of the corresponding blank samples.Urine: 15 ml of sample were diluted with an equal volume of water and loaded onto the 10 x 10 mm i.d. immuno-precolumn. Plasma: 6 ml of sample were diluted with 24 ml of water and processed in the same way. The analyses were performed according to the non-selective desorption technique896 ANALYST. SEPTEMBER 1991, VOL. 116 Aqueous solutions of either high or low pH and solutions containing chaotropic compounds were ineffective. This finding, confirmed by studies with other types of immuno- precolumns, suggests that these conventional desorbing solu- tions can only be used, if the sorbed compounds (antigen or hapten) possess structural features that are prone to chao- tropic or ionic changes.Cationic and anionic surfactant solutions were found to disrupt the antibody-antigen (or hapten) interaction if used in concentrations above their c.m.c. Concentrations of below the c.m.c. had no significant effect. Unfortunately, the practicability of micellar solutions is limited because it is difficult to concentrate and/or separate the desorbed species using a reversed-phase column. On-line thermodesorption of the immuno-precolumn at 55 "C gave good recoveries, both with conventional water- bath heating and microwave heat generation. The stability of the immuno-precolumn was less than with the non-selective methanol desorption, probably due to the thermolability of the agarose material used.In further work, temperature- resistant stationary phases such as silica, glass or synthetic polymers should be evaluated. A significant advantage of the technique is that almost totally aqueous solutions are employed for desorption; consequently the degree of contami- nation will be low. From among the methods tested, non-selective desorption from the immuno-precolumn with methanol-water (95 + 5 v/v) was found to be most practical. The eluate is diluted on-line with an excess of water and subsequently concentrated and separated in a reversed-phase LC system. While some problems remain to be solved with plasma samples, the results of urine analysis are excellent. The capacity of the immuno- precolumn remained constant during 20 runs each of 15 ml urine samples.The repeatability was satisfactory, and with 15-75 ml samples, detection limits (UV detection at 280 nm) of 0.05-0.5 pg 1-1 of steroid were obtained. Compared with the previously published immunoselective technique,7-9 where an excess of cross-reacting compounds is used as the displacer, non-selective methanol desorption shows several significant advantages and the chromatogram is cleaner because the broad displacer peaks are absent. Besides, a time-consuming search for and evaluation of the characteris- tics of proper displacers is superfluous. Generally speaking, if a suitable immuno-precolumn for a specific application is available, a non-selective desorption procedure can be devel- oped within a relatively short time.In principle, immunoselective desorption should show better selectivity than non-selective desorption, the additional selectivity is created during desorption. A cross-reacting compound will act as a displacer for selected antibodies only, non-specifically adsorbed compounds, or compounds cap- tured by other types of antibodies, will not be desorbed. However, if the chromatograms recorded with both desorp- tion techniques and performed with the same urine samples are compared (cf. reference 9 and this study), no essential difference in selectivity is found. Obviously, the immunoaffin- ity sorption process is so selective that further improvement is not easily achieved during desorption. Besides, any additional selectivity of the immunoselective desorption might be outweighed by the introduction of impurities into the system by the displacers. The main advantage of immunoselective desorption is that it can be performed with simpler instrumen- tation, as dilution of the eluate from the immuno-precolumn is not necessary. This work was supported by the Dutch Foundation for Technical Science, grant No. VCH 46.0616. Organon (Oss, The Netherlands) is thanked for the gift of the anti-estrogen antiserum. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Janis. L. J . , and Reigner, F. E . , J . Chromatogr., 1988, 444. 1. Nilsson, B . , J . Chromatogr., 1983, 276, 413. Reh, R., J. Chromatogr., 1988, 433, 119. Rybacek, L., D'Andrea. M.. and Tarnowski, J . , J . Chroma- togr., 1987, 397, 355. Janis, L. J., Grott, A., Regnier, F. E., and Smith-Gill, S. J . . J . Chromatogr., 1989, 476, 235. Johansson, B . , J . Chromatogr., 1986, 381, 107. Farjam, A . , de Jong, G. J.. Frei, R. W., Brinkman, U. A . Th., Haasnoot, W., Hamers, A. R. M.. Schilt, R . , and Huf, F. A . , J . Chromatogr., 1988,452,419. Haasnoot. W., Schilt, R . , Hamers, A. R. M., Huf, F. A . , Farjam, A . , Frei, R. W., and Brinkman, U . A. Th., J . Chromatogr., 1989,489, 157. Farjam, A . , Brugman, A . E . , Soldaat, A., Timrnerman, P., Lingeman, H . , de Jong, G . J., Frei, R. W., and Brinkman, U. A. Th., Chromatographia, 1991,31, 469. Farjam, A., van de Merbel, N. C., Lingeman, H . , Frei, R. W., and Brinkman, U . A. Th., Znt. J . Environ. Anal. Chem., in the press. Phillips, T. M., Recept. Biochem. Methodol.. 1989, 14. 129. Davis, G. C., Hein, M. B . , Chapman, D. A . , Neely, B. C., Sharp, C. R . , Durley, R. C., Biest, D . K., Heyde, B . R., and Carnes, M. G . , Plant Growth Substances, Springer, Berlin, 1985, pp. 44-51. Fuchs, Y . , and Gertman. E., Plant Cell. Physiol., 1974,lS. 629. Sundberg. B., Sandberg, G . , and Crozier, A., Phytochemistry, 1986, 25, 295. Van de Water, C., Tebbal, D., and Haagsrna, N . , J . Chroma- togr., 1989, 478, 205. Paper 1/008741 Received February 22nd, 1991 Accepted April 14th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600891

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

897 ANALYST, SEPTEMBER 1991, VOL. 116 Simultaneous Determination of Seven Divalent Metal Cations in Some Anaerobic Sealant Formulations Following Solid-phase Extraction and Separation on a Dynamically Coated CI8 High-performance Liquid Chromatography Column Marian Deacon and Malcolm R. Smyth* School of Chemical Sciences, Dublin City University, Dublin 9, Ireland Raymond G. Leonard Loctite (Ireland) Ltd., Whitestown Industrial Estate, Tallaght, Co. Dublin, Ireland A method has been developed for the simultaneous determination of eight divalent metal ions based on solid-phase extraction and separation on a dynamically coated CI8 reversed-phase high-performance liquid chromatography column. The method has been applied to the determination of these metal ions in some anaerobic sealant formulations, with limits of detection as low as 30 ppb for certain metal ions.It has been shown that this approach is capable of distinguishing between free and complexed metal ions in such matrices. Keywords: anaerobic sealants; reversed-phase high-performance liquid chromatography; solid-phase extraction; trace metal analysis Anaerobic sealants are used predominantly in mechanical engineering applications. They are also used in a wide range of bonding, locking and retaining applications, and are designed to remain in the liquid form in air, but once confined between closely fitting active metal surfaces, to polymerize into tough, heat and solvent resistant solid materials. A typical formula- tion contains a monomer, free radical initiators, accelerators, metal chelating agents, plasticizers, thickeners and dyestuffs.The determination of trace metal ions in anaerobic sealant formulations is of vital importance, as these metal ions can initiate the polymerization process, resulting in premature setting of products in their packaging material. A metal ion can react with a peroxide initiator to form a free radical. Any oxygen present will quickly react with this radical to render it inactive. In the absence of oxygen, however, this free radical will react with the monomer to form a polymer as follows: R-0-0-H + H+ + M"+ --+ RO' + M(n + I ) + + H20 R-0-0-H + M(n + l ) + + R-0-0' + Mn+ + H+ I I / \ I I +- RO-C-C' etc. \ / RO' + C=C This requires, therefore, that free metal cations must be excluded from a typical product.This is achieved by monitor- ing the raw materials and the finished product for trace metal ion content, and by the addition of chelating agents to the product. To date, analysis of trace metal ion content has been carried out by atomic absorption spectroscopy or direct current plasma spectrometry.' Both of these techniques determine total metal ion concentrations and, as described, only allow for single element determination. There is there- fore a need for a technique which allows for a multi-element determination approach, and one which can distinguish between free and complexed metal ions. In previous papers, high-performance liquid chromato- graphic methods for the determination of Cu" and Fe"' in anaerobic adhesive formulations using 8-hydroxyquinoline as a complexing agent,2 and using a cation-exchange column have been described.3 The first approach had the disadvan- tage that only a limited range of metal ions could be determined simultaneously, whereas the latter method suf- fered from incomplete resolution of the metal ions investi- * To whom correspondence should be addressed.gated. It was therefore decided to investigate the approach reported by Cassidy and Elchuk4 and refined by Krols using a CIS column dynamically coated with sodium octyl sulphonate for the particular industrial application described in this paper. Experimental Reagents All aqueous solutions were prepared with water obtained by passing distilled water through an ELGA-STAT (Elga, High Wycombe, UK) water purification apparatus. 4-(2-Pyridyl- azo)resorcinol (PAR) was obtained from Aldrich.Sodium octyl sulphonate was obtained from Rohm Chemicals. Tar- taric and citric acids were obtained from Riedel-de Haen. Metal standards were prepared from atomic absorption standard metal ion solutions obtained from May and Baker, with the exception of Fe" which was made from ammonium iron(1r) sulphate hexahydrate. Isooctane and diethyl ether were obtained from Riedel-de Haen, whereas chloroform was obtained from Aldrich. Apparatus The high-performance liquid chromatography (HPLC) system consisted of a Waters 501 HPLC pump, a Waters RCSS Guard Pak Module, and a Waters PBondapak CI8 cartridge (4.6 mm x 10 cm) contained within a Waters Z module. The eluted metal ions were detected after post-column derivatization with PAR reagent.The post-column reactant was delivered using a Waters reagent delivery module to a T-piece situated between the end of the column and the reaction coil. The complexed metal ions were detected using a Shimadzu 8PD-6AV ultraviolet detector. The detector output was monitored with a Drew Scientific data capture unit and the chromatograms were processed with a Roseate Chromato- graphy data management package. All sample preparation cartridges were obtained from Waters with the exception of the OnGuard H cartridge which was obtained from Dionex. Plastipak syringes from Becton Dickinson were used for the introduction of samples onto the sample preparation cartridges.898 G ANALYST, SEPTEMBER 1991, VOL. 116 Methods Solid-phase extraction was performed using a variety of sample preparation cartridges.The cartridges were first wetted with 2 ml of ethanol, then conditioned by washing with 5 ml of de-ionized water. Samples (30 ml) were then applied to the top of each cartridge and the metal ions eluted with 3 ml of the mobile phase. The first 1 ml (representing the dead volume of the cartridge) was discarded and the remaining 3 ml used for injection into the HPLC system. The spiking of metal ions into adhesive formulations was achieved using standards prepared in ethanol-water (90 + 10). The mobile phase consisted of 2 mmol dm-3 sodium octyl sulphonate, 20 mmol dm-3 citric acid and 30 mmol dm-3 tartaric acid, pH 3.4, as described by Krol.5 Results and Discussion Optimization of Chromatographic Separation and Detec- tion System The choice of chromatographic conditions was made based on the work of Cassidy and Elchuk4 and Krol.5 Further optimiza- tion was then carried out with respect to the size of the sample loop, the length of the reaction coil and the temperature of the reaction coil.The large capacity of the column and the large volume of eluents (3 ml), emanating from the solid-phase extraction cartridges used for sample preparation, required the use of a large sample loop. Sample loop sizes investigated were 200 pl, 500 pl, 1 ml and 2 ml. It was found that the peak height for each metal ion increased with increasing sample loop size up to 1 ml (Fig. 1). Further increase in the loop size to 2 ml resulted in the capacity of the column being exceeded. A 1 ml sample loop size was therefore used in all further investiga- tions.The length of the reaction coil from the mixing T-piece to the detector determines the reaction time of the PAR reagent with the metal ion. The length of the reaction coil was varied from 25 to 200 cm. For all metal ions investigated, the peak height increased with increasing length of the reaction coil (Fig. 2). However, a reaction coil length of greater than 135 cm produced a high back-pressure, which affected the mixing of the PAR reagent and column eluent in the mixing T-piece. Thus a reaction coil length of 135 cm was used in all further investigations. The temperature of the reaction coil was then varied from room temperature to 60 "C (Fig. 3). For most metal ions the peak height remained relatively constant, but an increase was noted in the peak height of Fellr above 40 "C, whereas a decrease was found for ZnlI between 10 and 40 "C.As the peak heights of most metal ions remained relatively constant, it was decided to maintain the reaction coil at room temperature. 30 > 5 E 6 20 .- Q) L Y 0, tL 10 0 Fig. 1 200 400 600 800 1000 Sample loop size/pl Effect of sample loou size on peak height obtained following separation of eight di\;alent metal catidns. A, Cill; B, Co"; C, Zn"; DY Mn"; E, Fell; F, Nil1; G, Pb"; and H, Cd" The chromatographic separation of a 1 ppm mixture of eight divalent metal ions achieved using this system is shown in Fig. 4. Optimization of Sample Preparation Solid-phase extraction was preferred to liquid-liquid extrac- tion because of the smaller amounts of sample, solvent and glassware needed, and considering also that liquid-liquid extraction is difficult to perform with samples which contain surfactants because of the formation of emulsions. Six commercially available solid-phase extraction cartridges were investigated with regard to their adsorption properties for metal ions from aqueous solutions, and the subsequent release of the metal ions from the adsorption sites upon washing with the mobile phase.These included Silica, Florasil, Alumina, Diol and Accel Plus CM Sep-Paks, and the 2 l2 t 20 40 60 80 100 120 140 Reaction coil lengthkm Fig. 2 Effect of length of reaction coil on peak height obtained following separation of eight divalent metal cations. A, Zn"; B, Co"; C, Cu"; D, Fe"; E, Mn"; F, Ni"; G, Pb"; and H, Cd" i a C \ 4 -1 1 I 1 1 I I L 0 7 14 21 28 35 42 Timelmin Fig.4 Separation of a 1 ppm aqueous mixture of: 1, Cu"; 2, Pb"; 3, Ni"; 4, ZdI; 5 , Co"; 6, Fe"; 7, Cd"; and 8, Mn"ANALYST, SEPTEMBER 1991, VOL. 116 899 OnGuard H sample preparation cartridge. The silica Sep-Pak proved the most promising in terms of recoveries of the majority of the metal ions studied. The recoveries of the eight divalent metal ions from 30 ml of de-ionized water spiked at the 100 ppb level following solid-phase extraction using the silica Sep-Pak ranged from 61% for Mn" to 110% for Zn". It was then decided to investigate the extraction of the metal ions from a typical plasticizer used in anaerobic adhesive formulations, i.e., tetraethylene glycol db(2-ethylhexanoate).The wetting-conditioning procedure was the same as that described above. A 30 ml sample of plasticizer solubilized in an equal volume of iso-octane was passed through the Sep-Pak with the aid of a 10 ml syringe at a rate of 2-3 ml min-1. The Sep-Pak was then washed with 10 ml of iso-octane to remove any remaining plasticizer. The metal ions were eluted from the Sep-Pak with 3 ml of the mobile phase, the first 1 ml being discarded and the remaining 3 ml collected. This fraction was passed through a CIS Sep-Pak in order to remove any remaining organic components. A typical chromatogram, obtained following this proce- dure, is shown in Fig. 5, where it can be seen that there was a good recovery for most of the metal ions except Fell. To investigate the loss of Fell, a liquid-liquid extraction technique was developed involving the extraction of 50 ml of plasticizer solubilized in 50 ml of iso-octane with 5 ml of the mobile phase.Again Fe" was not extracted using this procedure. The loss of Fell can therefore be attributed either to complex formation with the plasticizer or interconversion of Fell to the Fell1 form. Unfortunately, although the analysis technique used is able to distinguish between Fell and Fell1, the sensitivity for Fell1 is so low that it was not possible to use this technique to monitor such an interconversion process. Iso-octane was chosen as a solubilizer as it was the most non-polar solvent in which the plasticizer was soluble. This implied that any residual organic carryover in the elution of the metal ions with the mobile phase would be easily removed using the CIS Sep-Pak.Procedures were then developed for raw materials, inter- mediate and finished products. For formulations which were not soluble in iso-octane, more polar solvents such as chloroform and dichloromethane were used. It was found, however, that when such solvents were employed, splitting of the Cull peak occurred. This was thought to be due to carryover of these solvents because of their miscibility with water and decreased retention on the CIS Sep-Pak. This problem was counteracted by employing a 10 ml wash with iso-octane following the 10 ml wash of the Sep-Pak with the solubilizing solvent (i. e . , chloroform-dichloromethane). As 1 i 7 8 0 7 14 21 28 35 Ti me/m i n Fig.5 Chromatogram obtained following solid-phase extraction of metal ions present at the 100 ppb level in a solution of tetraethylene glycol di-(2-ethylhexanoate). Peak numbers as in Fig. 4 chloroform is miscible with iso-octane, the chloroform was removed, and the traces of the less polar iso-octane were trapped by passage through the CIS Sep-Pak. A typical chromatogram obtained for six of the metal ions following solid-phase extraction from a formulation containing triethyl- ene glycol dimethacrylate, cumene hydroperoxide, saccharin, dodecyl methacrylate and a primary alcohol ethoxylate is shown in Fig. 6. In samples containing aromatic amines such as N,N- dimethyl-o-toluidine and N , N-diethyl-p-toluidine, peak split- ting again occurred. For the removal of these amines, solvents such as ethanol, acetone and diethyl ether were investigated.Diethyl ether proved to be the most suitable. Again, the order in which the washing of the Sep-Pak was carried out proved to be critical to prevent the carryover of washing solvents. Therefore, samples containing amines were solubilized in chloroform and passed through the Sep-Pak, which was then washed with 10 ml aliquots of the chloroform, followed by diethyl ether and iso-octane. With samples containing a typical surfactant such as the linear primary alcohol-based ethoxylate ( CI2-Cl5) named dobanol25-3, it was found that water was a better solubilizing agent than chloroform, and adjustment of the sample water mixture from the initial pH of 4.2 to a pH of 7.0 resulted in improved recoveries of the metal ions. The method involved solubilizing the sample in an equivalent volume of water and adjusting the pH to 7.This was applied to a pre-conditioned silica Sep-Pak. The Sep-Pak was then washed with water and 10 ml of iso-octane, and the metal ions were eluted with the mobile phase and passed through the CIS Sep-Pak. The determination of six divalent metal ions, namely Cu", Nil1, Zn", Co", Cd" and MnlI was then investigated in a typical anaerobic sealant. Linear calibration graphs were constructed for all the metal ions at concentration levels of between 60 and 500 ppb. Limits of detection from 40 to 70 ppb and recoveries of 61-110% were obtained for these metal ions using a sample size of 30 ml (Table 1). 1 I I 1 I I 0 7 14 21 28 35 Fig.6 Chromatogram obtained following solid-phase extraction of a sample of anaerobic sealant formulation containing 100 ppb of Cu", Pb", Zn", Co", Cd" and Mn". Peak numbers as in Fig. 4 Table 1 Recoveries and limits of detection for metal ions using the proposed method Limit of Metal ion Recovery (%) detectiodppb CU" 80 60 Pb" 93 70 Ni" 89 70 Zn" 110 40 CO" 75 40 Fe" ND* 0 Cd" 97 60 Mn" 61 60 * ND = not determined.900 ANALYST, SEPTEMBER 1991, VOL. 116 In order to investigate the ability of the method to differentiate between free metal cations and their ethylene- diaminetetracetic acid (EDTA) complexes, a sample of tetraethylene glycol di-(2-ethylhexanoate) was spiked at the 100 ppb level with the metal ions under investigation and an excess of EDTA was added. These samples were then extracted using the solid-phase extraction procedure des- cribed above. The chromatogram obtained showed no response for any of the metal ions investigated, indicating that the procedure could be used to detect free metal ions in the presence of complexed metal ions. Conclusion A method has been developed for the determination of eight transition metal ions in some anaerobic adhesive formulations following solid-phase extraction and separation on a dynamic- ally coated CI8 HPLC column. The method is easy to perform, offers good detection limits for most of the metal ions studied, provides a multi-element detection capability and permits the free and complexed metal ions to be distinguished in the sealant formulations studied. References Brennan, M., and Svehla, G., Zr. Chem. News, 1990 Autumn edition, 35. Mooney, J. P., Meaney, M., Smyth, M. R., Leonard, R. G., and Wallace, G. G., Analyst, 1987, 112, 1555. O’Dea, P., Deacon, M., Smyth, M. R., and Leonard, R. G., Anal. Proc., 1991, 28,82. Cassidy, R. M., and Elchuk, S., Anal. Chem., 1982, 54, 1558. Krol, J . , Waters Zon Chromatography Notes, Waters Chromato- graphy Division, Millipore, Milford, MA, 1988, vol. 2, p. 1. Paper 1 /00871 D Received February 22nd, 1991 Accepted April 8th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600897

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1991, VOL. 116 901 Thermal Degradation of Some Benzyltrialkylammonium Salts Using Pyrolysis-Gas Chromatography-Mass Spectrometry Neville J. Haskins and Robert Mitchell* SmithKline Beecham Pharmaceuticals, The Frythe, Welwyn, Hertfordshire AL6 9A R, UK A number of benzyl quaternary ammonium salts have been examined by pyrolysis-gas chromatography-mass spectrometry using various temperatures for the pyrolysis. The products show that simple substitution reactions dominate at low temperatures with slight evidence for classical Hofmann degradation. Raising the temperature gave increasing concentrations of Ir2-diphenylethane and stil bene which must be produced by intermolecular reactions. There appeared to be a linear relationship between the amount of Ir2-diphenylethane produced and the temperature of the pyrolysis.Keywords: Benzyltrialkylammonium salts; pyrolysis-gas chromatograph y-mass spectrometry Quaternary ammonium salts such as choline (an endogenous component) are important in medical research, or as drugs (dequalinium bromide or propantheline bromide for ex- ample). However, the stability of such compounds at elevated temperatures is poor. The elimination of alkyl substituents from quaternary salts by either the Hofmann or Saytzeff reaction is well known. It was decided to examine the stability of benzyl quaternary ammonium salts at high temperature. This work describes the products formed by Curie-Point pyrolysis using support wires of various Curie-Point tempera- tures. Experimental All samples were obtained from Aldrich (Gillingham, Dorset, UK) and were used without further purification. Each of the salts was prepared for spectral analysis by dissolving it in methanol (100 mg per 10 ml) and adding 5 pl of the solution to a wire pre-washed in methanol.The wires were loaded into a glass carrier. The carriers were loaded into a Curie-Point pyrolyser (Horizon Instruments, Heathfield, Sussex, UK) and the radiofrequency field was applied for 4 s. Wires giving Curie-Points of 385, 510, 610 and 770 "C were used. Helium was passed over the wire at a flow rate of 1 ml min-1. The pyrolysate was trapped on a fused silica column (15 m x 0.25 mm i.d.) coated with BP-5 (SGE, Milton Keynes, Buckinghamshire, UK) at ambient temperature. The column was heated from ambient temperature to 300 "C at a rate of 10 "C min-1.The helium passing through the pyrolyser was used as the carrier gas at a flow rate of 1 ml min-1. The column was passed directly into the electron ionization source of a VG Analytical 7070F double focusing mass spectrometer linked to a VG 2050 data system (VG Analytical, Wythenshawe, Manchester, UK). The mass spec- trometer was continually scanned from 600 to 20 u at a rate of 1.5 s decade-' giving a total cycle time of about 3 s. Ionization was effected at 70 eV. After acquisition, spectra were transferred to a VG Analytical 11-2505 data system for processing. Results and Discussion The benzyltrialkylammonium salts examined were benzyl- trimethylammonium chloride (I), benzyltriethylammonium chloride (11), benzyldimethyldodecylammonium chloride (111) , benzyldimethyltetradecylammonium chloride (IV), benzyldimethylhexadecylammonium chloride (V) and benzyl- dimethylstearylammonium chloride (VI). * Present address: The Sheffield Polytechnic, Sheffield S1 lWB, UK.Fig. 1 shows a typical reconstructed gas chromatogram obtained after the pyrolysis of benzyldimethylhexadecylam- monium chloride (V). The major components are mainly those that might be expected from simple nucleophilic substitution reactions between the chloride anion and the quaternary cation (Scheme 1). This type of reaction in most instances should give rise to three pairs of compounds, consisting of the halide and the tertiary amine. This is observed in the pyrolysis of compound (V): the observed pairs are benzyl chloride and dimethylhexadecylamine; chloro- methane and benzylhexadecylmethylamine; and l-chloro- hexadecane and benzyldimethylamine.The proportion of each pair varies, with the least abundant being benzyldi- methylamine/chloroalkane. This order of abundance was observed for all the salts examined (Fig. 1). The products from the substitution reactions are observed at all wire temperatures. The most studied reaction of quaternary ammonium salts is the Hofmann degradation1 which should give rise to benzyldimethylamine, hydrogen chloride and the corresponding alkene (Scheme 2). Benzyl- dimethylamine can also arise from a substitution reaction, whereas hydrogen chloride is unlikely to be detected through a gas chromatograph. Accordingly, the evidence for Hofmann elimination depends on the detection of the alkene.Trace amounts of alkene were found but at a low concentration and only at the higher wire temperatures, implying that under the Time --c Fig. 1 Reconstructed gas chromatogram for the pyrolysate from the pyrolysis of benzyldimethylhexadecylammonium chloride at 385 "C. A, Benzyl chloride [relative molecular mass (M,) = 1261 and chloromethane (M, = 50); B, dimethylbenzylamine (M, = 135); C, 1,2-diphenylethane (M, = 182); D, 1-chlorhexadecane (M, = 260); E, dimethylhexadecylamine (M, = 269); F, hexadecanoic acid (M, = 256); G, dimethylheptadecylamine (M, = 283); H, benzylhexadecyl- methylamine (M, = 345); I, unknown, but containing a CI6 alkyl group?; J, unknown, but containing a Ci6 alkyl group?; and *, expected elution position of stilbene (observed with higher Curie- Point temperatures)902 ANALYST, SEPTEMBER 1991, VOL.116 f t H3C\ ,CH3 N I H &r3 'CH3 + Scheme 1 anion and the quaternary cation Nucleophilic substitution reactions between the chloride r C I - Scheme 2 nium salt 100 90 80 s < 70 E 60 A c .- c .- a 50 .- 40 30 20 10 0 - a CK CH=CHR + + HCI Classical Hofmann degradation of a quaternary ammo- m/z Fig. 2 of benzyltrialkylammonium salts Mass spectrum of 1,2-diphenyIethane obtained after pyrolysis conditions used for the pyrolysis, Hofmann elimination is not a major degradation pathway. One product observed (peak F, Fig. 1) was identified as hexadecanoic acid. This probably arises from oxidation of the hexadecyl moiety. Homologous acids were observed in the other long-chain alkyl quaternary salts.The amount was variable, suggesting that oxidation was caused by residual air after loading the pyrolyser. 100 90 80 70 '5 60 .E 50 .z 40 30 20 10 0 1 % c a 4- a - 20 1 51 89 152 I I 1 I, II 60 100 140 m/z Fig. 3 Mass spectrum of stilbene obtained after pyrolysis of benzyltrialkylammonium salts 0.36 $ 0.32 m 0.28 0.24 Y al al g 0.20 z 0.12 $ 0.16 - 5 0.08 0.04 ". - 0 340 420 500 580 660 740 Wire temperaturePC Fig. 4 Plot of p (%) versus T ("C), where p (%) = proportion of 1,2-diphenylethane formed on pyrolysis. A, Benzyltrimethylammo- nium chloride: y = 0.00103~ -0.340; B, benzyltriethylammonium chloride: y = 0.00049~ -0.174; C, benzyldimethyltetradecylammo- nium chloride: y = 0.00014~ -0.043; and D, benzyldimethylstearyl- ammonium chloride: y = 0.00010~ -0.028 -1.0 -2.0 -3.0 - 4.0 s -- - -5.0 I -7.0 -8.0 1 -9.0 \ -10.0 I I I I I I '.0.95 1.05 1.15 1.25 1.35 1.45 1.55 1/T x K-' Fig. 5 Arrhenius plot of log [p (%)I versus 1/T (K), wherep (%) = proportion of 1,2-diphenylethane formed on pyrolysis. A, Benzyl- trimethylammonium chloride; B, benzyltriethylammonium chloride; C, benzyldimethyldodecylammonium chloride; and D, benzyl- dimethyltetradecylammonium chloride At the highest wire temperatures additional products were observed. Examination of the spectra (Fig. 2) indicated that these were 172-diphenylethane and smaller amounts of stil- bene (Fig. 3). Such products would require some type of reaction between two molecules of the quaternary cation o r perhaps between benzyl chloride and the quaternary salt.ANALYST, SEPTEMBER 1991, VOL.116 903 N CH( ‘CH, R 1 O C H : + HCI Scheme 4 on pyrolysis Possible formation of stilbene via a carbene intermediate Scheme 3 explain the formation of 1,2-diphenyIethane Possible modification of the Stevens’ rearrangement to Further examination of the proportion of 1,2-diphenylethane formed showed that it was linearly dependent on the wire temperature and appeared to be less abundant as the size of the alkyl substituents increased (Fig. 4). The classical Arrhe- nius plot showed a similar trend (Fig. 5 ) . The formation of 1,2-diphenylethane might arise from a combination of benzyl free radicals themselves, or generated from an ylide formed by abstraction of the a-proton (Scheme 3). Such reactions have been observed in the study of the Stevens’ rearrangement carried out by Ollis et al.1 However, the ylide in the Stevens’ rearrangement is normally stabilized by the presence of a fl-keto group. It must be stressed that in the present situation considerably higher temperatures are used and the concentrations are not equivalent. Hence the formation of 1,2-diphenylethane in the hot melt after pyrolysis is occurring under far more rigorous conditions than the normal Stevens’ rearrangement. The presence of stilbene could arise from a combination of carbenes formed by simple disproportionation of the inter- mediate ylide formed by elimination of a benzyl methylene proton (Scheme 4). The presence of such reactive species might account for trace amounts of higher relative molecular mass material seen after some reactions, but too weak to identify conclusively. Conclusions The analysis of quaternary salts has often generated problems owing to their non-volatility as intact species, and the formation of decomposition products if heat is used. This paper shows that the decomposition reactions are more complex than might be expected. Pyrolysis of benzyl quaternary ammonium salts proceeds by several competing mechanisms. The products formed depend on the temperature of pyrolysis and the bulkiness of the alkyl substituents about the quaternary ammonium centre. At lower temperatures simple displacement reactions appear to predominate. At higher temperatures more complex multi- centre reactions, probably involving free radicals, take place giving rise to 1,2-diphenylethane and trace amounts of stilbene. Reference 1 Ollis, W. D., Rey, M., and Sutherland, I . O., J. Chem. SOC., Perkin Trans. I, 1983, 1009. Paper 1 I01 2486 Received March 15th, 1991 Accepted April 30th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600901

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1991, VOL. 116 905 High-performance Modular Spectrophotometric Flow Cell Joiio Carlos de Andrade and Kenneth E. Collins Universidade Estadual de Campinas, lnstituto de Quimica, C.P. 6154, 13081 Campinas, Sao Paulo, Brazil Monica Ferreira lnstituto Agron6mico de Campinas, C. P. 28, 1300 I Campinas, Sao Paulo, Brazil A high-performance modular flow cell is described, which can be used in photometric or spectrophotometric detector systems for analytical and preparative scale low-pressure liquid chromatography, flow injection and related techniques. The basic design is that of an inner absorption cell unit sandwiched between two rugged supports. The novel aspects of this sandwiched cell are the wide range of interchangeable flow cell units of different dimensions that can be used, and the way in which the fluid flow occurs, essentially eliminating problems with gas bubbles and giving rapid cell clearance.The cell is compact and its versatility is enhanced by using optical fibre bundles t o transmit the light beam t o the optical path of the cell and then from there t o the detector. Keywords: Modular flow cell; spectrophotometric detector; liquid chromatography; flow injection Flow cells are critical parts of the flow-through detectors used in chromatographic systems’ and also in the detectors used in other types of systems, e . g . , for continuous-flow analysis, either segmented24 or non-segmented.5 It is the measurement of some property of the fluid as it passes through the flow cell that gives the necessary analytical signal.Optical detector flow cells are usually purchased as an integral part of any commercially available detector system o r constructed on-site by the user. Most flow cells are made in one of four basic configurations: ‘Z’,h37 ‘U’,6.*--’” ‘H’,6 or conical.6-11,12 These cells differ mainly in the way the fluid stream enters and leaves the illuminated absorption chamber. The details of the liquid flow path affect the laminar and turbulent aspects of fluid flow, which in turn determines the over-all dispersion and bubble retention characteristics of the flow cell. Many commercially available and laboratory-built flow cells have serious problems with gas bubbles and suffer from long peak clearance times. We have designed a modular flow cell in which the components can be constructed in a typical machine shop at low cost.The basic unit consists of an inner absorption cell piece sandwiched between two rugged support pieces. A series of different flow cell volumes, pathlengths and fluid flow patterns can be obtained simply by replacing one inner piece with another, somewhat in the manner of changing the rotor of a high-performance liquid chromatography (HPLC) injec- tion valve. Incorporated into the inner cell design is the concept of introduction of the flowing fluid into, and its removal from, the inner light-path piece in a symmetrical way, by means of two or more entrance and exit channels, which avoids the non-symmetrical flow pattern inherent in conventional one-jet designs. Experimental Cell Design The components of the proposed flow cell are shown in Fig.1. Details of the light-source system (a 6 V, 10 W halogen lamp coupled to a compact collimating device13 incorporating a 542 nm interference filter), the sensing system (a GaAsP Hama- matsu photodiode, Model G1126-02) and the corresponding electronic circuits are not included in Fig. 1 nor are they discussed below. However, all of the measurements were made under identical optical conditions, except for the tests with the commercial U-type cell. The central piece of the flow cell (Fig. l ) , which defines the fluid flow characteristics of the detection system, can be made of any hard material that is sufficiently inert to withstand the solvents and solutes to be used. Graphite-filled poly- (tetrafluoroethylene) (PTFE) is a good choice as it is soft enough to cut the channels conveniently and it does not allow light scatter through the cell walls, as do translucent materials, such as white PTFE and Kel-F.The flow path is of a modified Z-type, with coaxial entrance and exit tubes. This set-up permits the use of cells with a shorter light path, which may be of interest in micro-scale analytical work, such as HPLC. The inlet flow stream is directed into a circular channel around the centre hole, from which it is then directed through radial channels (three in the example shown) into the centre hole. Thus the flowing fluid essentially jets in a symmetrical way from the channels towards the centre of the light path, along the inner faces of the entrance and exit windows. In the proposed cell the entrance and exit tubes consist of PTFE tubing (A in 0.d.) inserted into the appropriate holes in the cell body.The assembled cell easily supports a hydrostatic pressure of about 405 kPa without leakage. Conventional HPLC fittings could be used for applications of even higher pressure. The three radial channel design shown in Fig. 1 can be iced by that shown in Fig. 2. 0 0 B In 1 I I c out B 1 cm C Fig. 1 Cross-section and front view of flow cell components: C, flow cell body; B, support pieces for cell and detection transducer; 0, O-rings; W, windows of glass or fused silica; G. polytetrafluoroethylene sealing gasket; and D, well for detection transducer906 ANALYST, SEPTEMBER 1991, VOL. 116 The cell support pieces can be made of stainless steel, brass or hard organic polymer material and, if desired, can be made more compact than shown in Fig.1. Results and Discussion Cell Performance As the detector response should be a function of the amount of analyte present in the cell volume and independent of the existence of other components of the analytical system, such as a chromatographic column, the performance of the proposed flow cells was tested by using a single-line flow injection (FI) manifold.5 Thus the dynamic tests were carried out by injecting, in triplicate, a desired volume (e.g., 75 pl) of KMn04 solutions, having concentrations ranging from 4.0 x 10-5 to 6.0 x 10-4 mol dm-3, into the carrier stream (water), prior to its entry into the detector cell. The transient peaks obtained in this manner give an FI-type calibration graph and simulate the chromatographic peaks, permitting the actual dynamic characteristics of the detector configuration being tested to be observed.We have extensively tested cells having centre pieces of 1.5 mm i.d. and optical pathlengths of 5 mm (volume, 9 PI) and 10 mm (volume, 18 pl). The recorded peaks (Fig. 3) show a stable baseline and an excellent precision of response (see Table l), comparable to those of high quality, but expensive, commercially available constant-volume flow cells. Results obtained with a Hellma U-type, 18 p1, 10 mm optical pathlength cell (with a Zeiss PM2A spectrophotometer operating at 542 nm) are also presented. As the tests of the flow cells were carried out using a single-line FI manifold, comparisons of performance with respect to sensitivity (analytical response) were obtained a A @ B C D Fig.2 Front view of channel configurations tested: A. Z-type flow cell; B, C and D, configurations with two, three and four channels, respectively. The inlet and outlet tube geometries and the flow cell dimensions are the same for all channel configurations (see Fig. 1) through the dispersion coefficients values, D, defined as the ratio of the absorbance for the transient FI peak ( A ) to that for the steady-state signal (Ao). As all of the variables such as carrier flow rate, reagent concentration, injection volume and F -Time Fig. 3 Flow injection transient peaks obtained with a 9 yl flow cell. Sample, KMnO,; injected volume, 75 yl; and carrier flow rate, 1.6 ml min-1.Concentrations of the injected samples: A, 4.0 x 10-5; B, 8.0 x C, 1.0 x lo-,; D, 2.0 x 10-4; E, 4.0 x 10-4; and F, 6.0 x lo-, mol dm-3. The other peaks were obtained at increased recorder chart speeds 75 I 0 3.0 6.0 9.0 12.0 Flow rate/ml min-I Fig. 4 Influence of the carrier flow rate on cell clearance times. Curves A, B, C and D correspond to the flow cells of Fig. 2. Curve E corresponds to a 10 mm optical pathlength flow cell (18 pl), with channel configuration C of Fig. 2. Curve F corresponds to a 10 mm optical pathlength U-type flow cell (18 yl). Sample, 6.0 x 10-4 mol dm-3 KMnO,; injected volume, 75 yl Table 1 Average relative standard deviation (RSD) and dispersion coefficient (D) values for various flow cells. The values were obtained using a carrier flow rate of 1.6 ml min-1 and a 6.0 x 10-4 mol dm-3 KMn04 solution.The RSD values were calculated for injection volumes of 10,25, SO, 75 and 100 PI (n = 10). As the RSD values found were almost constant over this volume range, the values shown are the average Optical pathlength/ Average Flow cell Volume/yl mm RSD (%) D value* U-type configuration (Hellma) 18 10 1 .o 0.583 Three channel, 2 type (Fig. 2, C) 18 10 1.2 0.847 Three channel, Z type (Fig. 2, C) 9 5 0.6 0.790 2-type configuration (Fig. 2, A) 9 5 0.7 0.702 * Results from 75 yl injections.ANALYST. SEPTEMBER 1991, VOL. 116 907 J H G / J F E 0 C B A - Time Fig. 5 Peak shape as a function of the carrier flow rate. Sample, 6.0 x mol dm-3 KMnO,; injected volume, 75 p1; and flow cell volume, 9 PI.Flow rates: A, 1.0; B, 1.6; C, 1.9; D, 2.2; E, 2.7; F, 3.0; G, 3.4; and H, 3.6 ml min-1. The peaks were recorded at chart speeds of 6 and 63 mm min-1 the FI manifold were kept constant for this set of data, it is inferred that values of D directly reflect the performance of the cells. These results are also shown in Table 1. It can be seen that the average relative standard deviation (RSD) for the commercial U-type cell is similar to that of the proposed three-channel cell, but the value of D is markedly lower. The use of channels in both the entrance and exit sides of the proposed flow cell significantly improves the clearance time when compared with the 2-type cell having similar cell dimensions. Centre pieces with more than two radial channels all give similar results, as shown in Fig.4; hence, there is little advantage in having more than three channels. The efficiency of solute clearance for the 9 1.11 three-channel flow cell is characterized by the peak profiles shown in Fig. 5. The preliminary tests show that when radial channels are present only on the exit side of the cell, the results are similar to those obtained when radial channels are present on both the entrance and exit sides. If physical solute dispersion from the channel volume is not of concern, it is recommended that radial channels be placed on both sides of the cell, as a matter of convenience, when mounting the cell. In any configuration gas bubble problems are virtually non-existent . Applications The proposed cell can be incorporated into a dedicated detector or coupled with most conventional photometers and spectrophotometers by means of optical fibre bundles, with- out modification of the cell compartment. It functions particularly well in flow analysis systems where rapid cell clearance is desirable.This characteristic should be useful for determinations that involve frequent sampling in addition to on-line kinetic studies. 1 2 3 4 5 6 7 8 9 10 11 12 13 References Stevenson, R. L., in Liquid Chromatography Detectors, ed. Vickrey, T. M., Marcel Dekker, New York, 1983, vol. 23, Furman, W. B., and Walker, W. H. C., Continuous Flow Analysis, Theory and Practice, Marcel Dekker, New York, 1976. Pasquini, C., and de Oliveira, W. A., Anal. Chem., 1985, 57, 2575. de Andrade. J . C., Ferreira, M., Baccan, N., and Bataglia, 0. C., Analyst, 1988, 113, 289. R6iiEka, J . , and Hansen, E. H., Flow-Injection Analysis, Wiley, New York, 2nd edn., 1988. White, P. C., Analyst, 1984, 109, 677. Kirkland, J. J., Anal. Chem., 1968, 40, 391. Betteridge, D., Dagless, E. L., Fields, B., and Graves, N. F., Analyst, 1978, 103, 897. Weber, J . R., and Purdy, W. C., Clin. Chem., 1980,26, 1010. McClintock, S. A., Weber, J. R., and Purdy, W. C., J. Chem. Educ., 1985, 62, 65. Stewart, J. E., Appl. Opt., 1979, 18, 5. Stewart, J. E., Appl. Opt., 1981, 20, 654. Moore, J . H., Davis, C. C., and Coplan, M. A., Building Scientific Apparatus. A Practical Guide to Design and Construction, Addison-Wesley, New York, 2nd edn., 1989, p. 145. pp. 23-86. Paper 0104932H Received November Ist, 1990 Accepted May 16th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600905

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1991, VOL. 116 909 Generalized Treatment of a Stray Radiant Energy Test Method in Absorption Spectrometry Paddy Fleming Sligo Regional Technical College, Ballinode, Sligo, Ireland The test method of Mielenz et a/. used t o determine the relative stray radiant energy (SRE) level in spectrophotometers is generalized for all sample cell t o reference cell thickness ratios greater than unity. It is extended further t o include situations where the required ‘cut-off solution is not transparent t o the SRE. The experimental SRE value which ensued by applying the above method t o a Shimadzu 260 ultraviolet/visible spectrophotometer is reconciled with the corresponding experimental values arising from two other SRE test methods. Keywords: Spectrophotometer; stray radiant energy; cell pathlength ratio In this paper a method for determining the relative stray radiant energy (SRE) in ultraviolet/visible (UVNIS) spectro- photometers’ is generalized to include all sample cell to reference cell thickness ratios greater than unity and instances where the transmittance of the ‘cut-off‘ solution to SRE is not necessarily unity.The modified test method was used to determine the relative SRE level in a Shimadzu 260 spectro- photometer at a Mielenz peak1 wavelength of 654 nm and with its spectral slit-width set at 1 nm. The exercise was repeated for the same instrumental conditions by using Fleming’s transmittance ratio spectrometry method’ and Fleming and O’Dea’s direct transmittance method.3 Fleming’s SRE test method’ is based on transmittance ratio spectrometry and it may be regarded as the opposite side of the same coin as the test method of Mielenz et al.1 Both test methods measure the differential absorbance of a solution placed in the sample beam relative to an identical solution placed in the reference beam while the sample cell is thicker than the reference cell.The Mielenz method gradually increases differential absorbance by wavelength scanning through the leading or trailing absorbance edge of a cut-off solution which has been placed in both the sample and reference beams, whereas the transmittance ratio method is carried out at a fixed wavelength while differential absorbance is gradually increased by advancing the concentration of the solution held in the sample and reference beams. The differential absorbance for both methods will not increase indefinitely but will peak at a value determined by the relative SRE level of the spectrophotometer, the cell pathlength ratio employed and the SRE transmittance of the test solution.The direct transmittance SRE test method involves determining the actual absorbances ( A ’ ) of a series of concentrated solutions the monochromatic transmittances of which are at least a factor of 50 lower than the relative SRE level to be determined. Although the Mielenz test method is presented here in a generalized form in order to accommodate the attenuation of the SRE by the sample, that attenuation must be determined independently. Fleming and O’Dea’s direct transmittance SRE test method3 yields accurate estimates of both the SRE and the attenuation of the SRE by the sample, and if the attenuation of the SRE by the sample is known then it may be used in the Mielenz method to give a further accurate estimate of the SRE, provided the same sample type is used under identical instrumental conditions in both the test methods.The terms relative transmittance and differential absor- bance are used synonymously in the text. Formulation of Experimental Quantities If the same cut-off solution is placed in the beams of a double beam, ratio recording spectrophotometer but the sample cell is a-times as thick as the reference cell (a = b,/b,, where the subscripts s and r refer to the sample and reference beams, respectively, and b refers to the pathlength of the cells employed) then the transmittance of the sample beam solution relative to the reference beam solution, t’, in the presence of a relative SRE level of s is given by (1) tf + svff t’=- th + sv where th is the monochromatic transmittance of the reference beam solution and v, which may be weakly dependent on A, is its transmittance to SRE.When scanning through the absorbance edge of the cut-off solution a relative transmittance minimum, i.e., t’ (min) = t”, will be encountered1 at th = T . If the derivati used, i.e., setting dt’/dth equal to zero at = T , gives Tm + svm = ( T + S V ) ~ T ~ - ~ Solving for s in eqn. (2) gives (a - 1)Tm V“ - ( ~ v T f f - 1 S = However, the spectrophotometric observable ii e method is hen eqn. ( 1 ) ( 3 ) this experi- ment is not Tbut t”.Ifthe expression for s given by eqn. (3) is substituted into eqn. (l), simplified and rearranged, then the following ensues: T = (~“/a)1/(&-1) (4) Eqn. (4) may now be substituted into eqn. (3) to give (a - l)(T”/a)m’(ff-’) S = vff - vt” Eqn. (5) is an exact general expression which relates the relative SRE level (s) to the Mielenz relative transmittance minimum (t“), transmittance of the reference beam solution to SRE ( v ) , and sample cell to reference cell thickness ratio (a). The SRE transmittance was assumed to have remained constant over the spectral range covered by scanning through a Mielenz peak. The transmittance ratio ( r ) at wavelength h of a sample beam solution the cell pathlength of which is a-times greater than an identical reference beam solution is given by2 where p is the SRE transmittance of the reference sample the monochromatic transmittance (th) of which is 0.1 and Ah = -log th.This transmittance ratio function has a minimum, r(min) = p, at a certain monochromatic reference910 ANALYST, SEPTEMBER 1991, VOL. 116 transmittance, TL = t , which is determined by s, p and a. If p is taken to be unity, then, by using the derivative method, an expression may be derived2 which relates the relative SRE level to p and a (7) (a - 1) (p/a)a’(-l) 1 - P S = However, p is less than unity and its experimental value must be determined independently before eqn. (6) may be numeri- cally modelled for a selected value of a and trial relative SRE values. The trial relative SRE value which, when inserted into eqn.(6) for zh values in the range 1 d rk d O.lt, gives the best match with experimental transmittance ratio measurements may be taken as the best estimate for the said relative SRE level. The direct transmittance SRE test method3 allows for the direct determination of p and s through observing the actual transmittances (T’) of an arithmetic series of concentrated solutions the monochromatic transmittances (T~) of which are less than 0.02s. The relationship between T’, s, p and ‘ch is then given by (where Ah = -loglOth) (8) (9) T’ = spAk Taking the loglo of eqn. (8) gives log1oT’ = Iogl@ + Ah loglop At these very high monochromatic absorbances eqn. (8) is a linear relationship between logloT’ and Ah with a slope of loglop and an ordinate intercept of logl$. Experimental All the spectrophotometric measurements reported here were made with a Shimadzu 260 double-beam spectrophotometer at a spectral slit-width setting of 1 nm.Matched pairs of 1,2,5, 10, 20, 50 and 100 mm quartz-glass cells were at hand. Therefore, various nominal sample to reference cell pathlength ratios were possible and the following nominal values were used: 2 (= 10/5 and 20/10), 2.5 (= 50/20), 4 (= 20/5) and 5 (= 50/10). The working solutions were obtained from a 50 g 1-1 Orleans Blue food dye (E123) stock aqueous solution. The UVNIS absorption spectrum of 10 g 1-1 of the same solution in a 1 mm cell was given previously.2 An arithmetic concentration series of the parent solution was prepared. The most dilute and concentrated members had monochromatic absorbances of 0.025 and 0.5, respectively, in a cell of pathlength 1 mm at 654 nm and the arithmetic series had an absorbance increment of 0.025.This yielded a set of 20 solutions the monochromatic absorbances of which, at 654 nm, ranged from Amin to A,,, [incremented in steps (AA)] in various cell pathlengths (b) as follows: Amin + (AA x 19 steps) = A,,, in a b mm pathlength cell 0.025 + (0.025 x 19 steps) = 0.5 in a 1 mm pathlength cell 0.050 + (0.050 x 19 steps) = 1.0 in a 2 mm pathlength cell 0.125 + (0.125 x 19 steps) = 2.5 in a 5 mm pathlength cell 0.250 + (0.250 x 19 steps) = 5.0 in a 10 mm pathlength cell 0.500 + (0.500 x 19 steps) = 10.0 in a 20 mm pathlength cell 1.250 + (1.250 X 19 steps) = 25.0 in a 50 mm pathlength cell The experimental cell pathlength ratios were determined by measuring the absorbance at 630 nm of a dilute Orleans Blue food dye (E123) solution in all the available cells and this yielded the following relative pathlengths: 1.00 & 0.008; 2.00 k 0.013; 5.00 -t 0.017; 10.00 k 0.013; 20.01 k 0.013; and 50.00 k 0.039.The Mielenz test method was applied repeatedly to the Shimadzu 260 spectrophotometer by scanning slowly in the range 750 3 h(nm) 3 625, the spectral slit-width having been 3 2 1 0 625 650 675 700 725 ( b) 2 Sample concentration increasing 4 625 645 665 685 705 725 625 645 665 685 705 725 Wavelengthlnm Fig. l(a) Four Mielenz differential absorbance spectra in the wavelength range 725 3 h(nm) 3 625 for nominal cell pathlength ratios ((Yb) of A, 2; B. 2.5; C, 4; and D, 5. The Orleans Blue food dye concentration IS gradually increased with decreasing (Y values so as to maintain the Mielenz peak at 654 nm.( b ) Six Mielenz differential absorbance spectra in the range 725 3 h(nm) 3 625 for a nominal cell pathlength ratio ((Yb) of 2. (c) Four Mielenz differential absorbance spectra in the wavelength range 725 3 h(nm) 3 625 for nominal cell pathlength ratios ((Yb) of A , 2; B, 2.5; C, 4; and D, 5 , and for a fixed concentration of Orleans Blue food dye set at 1 nm. A food dye test solution had been placed in a pair of matched quartz cuvettes the nominal pathlength ratio of which was 2 (a = 10 mm : 5 mm). The ensuing differential absorbance spectra displayed the expected SRE Mielenz peaks at 654 nm. The monochromatic absorbance of the test solution at 654 nm in the 5 mm reference cuvette was 2.0 and the Mielenz peaks had an average absorbance of A” = 1.680 k 0.005.Eqn. (4) predicts that if the absorbance of the reference sample in the Mielenz SRE test method (-log T ) is 2 and a is 2, then the absorbance of the Ivlielenz peak should be 1.699 (-log t”). If a more concentrated member of the prepared Orleans Blue food dye test solutions had been used in the above experiment then the Mielenz peak would have occurred at a longer wavelength. The Mielenz analysis, s = 0.25 x 10-2A”, which is only applicable for a = 2, yields a relative SRE level of 0.000113 for the above experiment while eqn. (5); for v = 1 and a = 2, yields s = 0.000116. If the MielenzANALYST. SEPTEMBER 1991. VOL. 116 91 1 peak is to occur at 654 nm for all a values, then a priori knowledge of the absorbance of the test solution in the reference cell at 654 nm (-log 7) is necessary for each a value. A value for T" for a given a may be calculated using trial values for T" in eqn.(5) and assuming s = 0.000116 and v = 1. Eqn. (4) may then be employed to predict the appropriate approximate absorbance of the test solution which, when placed in the sample and reference cells, will give a Mielenz peak at 654 nm, e.g., if s = 0.000116 and a = 2.5, then t" = 0.0085 satisfies eqn. ( 5 ) for v = 1, and eqn. (4) yields T = 0.0226 or -log T = 1.646. This calculation procedure was executed in turn for a = 2.5, 4 and 5 and was facilitated by having prior knowledge of the absorbance of the Mielenz peak (A") at 654 nm which ensued from scanning the differential absorption of a selected food dye sample for CY = 2.The Mielenz test method was replicated for a = 2.5, 4 and 5 by using the food dye sample of appropriate concentration for each a and then eqn. (5) (with v = 1) was used to calculate the relative SRE level. The resulting Mielenz differential absorbance spectra are given in Fig. l(a). The Mielenz peaks occur at approximately the same wavelength (654 k 0.5 nm) for all cell pathlength ratios used but increase in amplitude as Q/ increases. Fig. 1 (b) displays six Mielenz differential absorbance spectra which were obtained for a constant nominal cell pathlength ratio of 2 (= 10/5) and by changing the sample concentration in the cuvettes for each scan. Note the red shift of the Mielenz peaks which occurs with increasing sample concentration.Fig. 1 ( c ) displays four Mielenz differential absorbance spectra which were scanned for constant sample concentration 0 1 .o 2.0 3.0 Monochromatic reference absorbance Fig. 2 Four plots of the differential absorbance versus the monochromatic reference absorbance of an arithmetic concentration series of Orleans Blue food dye (E123) solutions placed in pairs of cells the pathlength ratios (ab) of which were: A. 2.001; B, 2.499; C, 4.00; and D, 5.00. The measurements were made at 654 nm and with a spectral slit-width of 1 nm in the cuvettes and by changing the cell pathlength ratio between the following nominal values: 2 (= 10/5), 2.5 (= 50/20), 4 (= 20/5) and 5 (= 50/10). Note the red shift of the Mielenz peaks which occurs with increasing a values.Fleming's transmittance ratio2 and Fleming and O'Dea's direct transmittance3 SRE test methods were also applied to the Shimadzu 260 spectrophotometer set at 654 nm and a spectral slit-width of 1 nm. The ensuing experimental determinations are plotted in Figs. 2 and 3. The above mentioned food-dye concentration series was appropriately employed in both tests. Fig. 2 displays differential absorbance versus monochromatic reference absorbance plots for four cell pathlength ratios, a = b,/b,. The experimental cell pathlength ratios are as follows: 20.01/10.00 = 2.001 f. 0.005; 50.00/20.01 = 2.499 k 0.004; 20.01/5.00 = 4.00 k 0.016; and 50.00/10.00 = 5.00 f. 0.01. Fig. 3 is a plot on semi-log axes of the observed average transmittance (in 20 mm pathlength cells) of solutions ( T ' ) at 654 nm versus the respective monochromatic absorbance ( A ) in the range 0 d A d 10.0.The exponential regression equation of fit to the linear part of the plot in the upper absorbance range is given by T' = 0.000140 x lO-0.029A. Results Eqns. ( 5 ) and (6) cannot be applied to the differential absorbance maxima in Fig. l(a) and (b), respectively, without a priori knowledge of the transmittances of the samples to SRE at the wavelength of interest. The quantity 'v' in eqn. ( 5 ) is given by PA, where A (= -loglo7) is the monochromatic 100 L - a, 10-1 0 c m c .- E 10-2 2 10-3 2 8 10-4 10-5 m t 4- Q, m 0 2 4 6 8 10 Monochromatic absorbance (A) Fig. 3 Plot on semi-log axes of the observed transmittance (T) versus the monochromatic absorbance ( A ) at 654 nm and with a spectral slit-width of 1 nm.The measurements were made in a pair of matched 20 mm quartz cells with a Shimadzu 260 spectrophotometer. The analyte samples were dilutions of a 50 g I- Orleans Blue food dye in distilled water. The monochromatic absorbance range, 7 d A d 10, yielded s = 0.000140 and p = 0.935 Table 1 Relative SRE level in a Shimadzu 260 spectrophotometer set at 654 nm and a spectral slit-width of 1 nm using a revised Mielenz er af.' and the transmittance ratio spectrometry2 test methods Cell pathlength ratio. a(= b,lb,) 2.001 f 0.005 2.499 -t 0.004 4.00 k 0.016 5.00 k 0.01 Mielenz's peak absorbance at 654 nm (- log,oT") 1.68 2.08 2.73 2.94 Monochromatic reference absorbance at 654 nm (-logloT) from eqn.(4) 2.00 1.65 1.10 0.90 Relative SRE level [s( x ? 0.101 from eqn. ( 5 ) ( a ) for v = 1 1.12 1.12 I .08 1.14 ( h ) for v = 0.9354 1.53 1.51 1.46 1.55 from eqn. (7) and Fig. 2 maxima 1.04 1.01 1.14 1.13 from eqn. (6) for p = 0.935 in Fig. 2 1.35 1.25 1.45 1.40 Relative SRE level [s ( x 10-4) k 0.101 Relative SRE level [s ( x 10-4) f 0.101912 ANALYST, SEPTEMBER 1991, VOL. 116 transmittance of the samples employed was assumed to be unity. The direct transmittance SRE test method of Fleming and O’Dea,3 which allows for the non-transparency of the samples towards SRE, yielded an SRE value of 0.000140 -t- 0.000015 for the Shimadzu 260 spectrophotometer at 654 nm and a spectral slit-width of 1 nm. This was significantly greater than the average SRE values of 0.000112 -t 0.000010 and 0.000108 Ifr 0.000010 which were obtained by the two other test methods mentioned, based on calculations using eqns.( 5 ) and (7), respectively, for v = p = 1 and using four distinct cell pathlength ratios. These results may be reconciled with the direct transmittance SRE test method result if the SRE transmittance value yielded by the last test method, i.e., p = 0.935, is employed in eqns. (5) and (6). Eqn. (5) with p = 0.935 gave s = 0.000151 k 0.000010 and eqn. (6) with p = 0.935 gave s = 0.000136 k 0.000010 from the curves of best fit to the experimental points in Fig. 2. The three test methods gave relative SRE levels which agree within the experimental error, provided allowance is made for the absorption of the SRE by the test solution being used.However, only the direct transmittance SRE test method is self-contained in that it yields all the information required to specify the true relative SRE level in a spectrophotometer without having recourse to any other test method. absorbance at 654 nm of the reference beam sample used in the method of Mielenz et al. and p has been defined in eqn. (6). However, eqn. (8), applied to the linear portion of the upper absorbance range of Fig. 3, yields the following experimental values for p and s: p = 0.935 k 0.015 and s = 0.000140 k 0.000015. If the monochromatic absorbances at 654 nm of the reference beam samples are known, then eqn. (5) can be applied to the maxima in Fig. l ( a ) to yield the relative SRE levels recorded in Table 1.Eqn. (6) can then be used to generate four sets of data points, i.e., a matching set of data points for each set of experimental differential absorbance points in Fig. 2, through using trial s values for the relative SRE level and a setting p = 0.935. The ensuing simulated curves are traced in Fig. 2 and the optimum trial relative SRE values are listed in the bottom row of Table 1. Conclusion The original purpose of this paper was to generalize the theoretical basis of the SRE test method developed by Mielenz et al.1 so as to embrace all cell pathlength ratios greater than unity. The generalized theory was tested experimentally in this paper for four disparate cell pathlength ratios to yield relative SRE levels for a Shimadzu 260 spectrophotometer which agreed within the limits allowed by the experimental errors involved [see Table 1 for eqn. (5) and v = 11. However, the test method, being sample based, would underestimate the relative SRE levels in spectrophotometers if the test solutions absorbed the SRE. This postulate was tested in this paper by comparing the relative SRE levels in a Shimadzu 260 spectrophotometer determined in three semi-independent ways under identical instrumental conditions. The SRE test methods of Mielenz et al.1 and Fleming2 yielded compatible SRE results if the SRE References 1 Mielenz, K. D., Weidner, V. R., and Burke, R. W.,Appl. Opt., 1982, 21, 3354. 2 Fleming, P., Analyst, 1990, 115, 375. 3 Fleming, P., and O’Dea, J., Analyst, 1991, 116, 195. Paper IlOO489A Received February 4th, 1991 Accepted April 29th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600909

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1991, VOL. 116 913 Determination of Iron by Flow Injection Based on the Catalytic Effect of the lron(iii)-Ethylenediaminetetraacetic Acid Complex on the Oxidation of Hydroxylamine by Dissolved Oxygen Andreu Cladera, Enrique Gomez, Jose Manuel Estela and Victor Cerda* Department of Chemistry, Faculty of Sciences, University of the Balearic Islands, E-07071 Palma de Mallorca, Spain A f l o w injection (FI) spectrophotometric method for the determination of Fe based on the catalytic effect of the Fe1I1-ethylenediaminetetraacetic acid-complex on the oxidation of hydroxylamine by dissolved oxygen in a basic medium is described. The FI manifold used was interfaced to a personal computer provided with appropriate software for automatic commanding of sample injection and data acquisition and processing.The absorbance response provided by the spectrophotometer was linear over the range 3.5-150 ng ml-1 of Fe, and the detection limit achieved was 2 ng ml-I, with a relative standard deviation of 2.4% for 50 ng ml-1 of Fell1 ( n = 11) and 6.7% for 5 ng ml-1 of Fell' ( n = 11). The only serious interferences with the determination were those of Colt, CrtI1 and Cull, but these could be reduced such that at 40,40 and 10 times the Fe concentration they had no effect. The results provided by a normal and a reversed FI configuration were compared and the method was satisfactorily applied to the determination of Fe in natural waters and white wine at a sampling rate of 60 samples h-I. Keywords: Flow injection; spectrophotometry; catalytic effect; iron(///)-ethylenediaminetetraacetic acid complex; iron determination Since flow injection (FI) was introduced by RGiiCka and Hansen in 1975,' continuous advances in this technique and its adaptation to different analytical methodologies, particularly spectrometric, electrochemical and separative, have testified to its enormous potential in various areas of interest such as environmental, pharmaceutical and clinical analysis.Throughout the intervening 15 years, FI has attracted a growing number of workers, who have found it to be a reproducible, rapid and inexpensive alternative to existing methodologies, in addition to a most flexible tool for adaptation to particular problems, as demonstrated by the increasing number of papers published on FI, most of which have been compiled in several books and reviews.'-5 The spectrophotometric determination of Fe in an FI system has been addressed by several workers.Mortatti et a1.6 used two different measurement wavelengths to avoid inter- ferences and an automated system for the determination of 0.1-30 pprn of Fe at a high sampling frequency (180 h-1). Other workers'-9 accomplished the speciation of Fe" and Fe"' by different procedures but over similar determination ranges. Rios et al. 10 reported the individual and simultaneous determi- nation of Fe, Cu and Al over the range 0.1-3.0 pprn (Fe) in synthetic waste water samples; however, the determination was subject to severe interferences from carbonate and phosphate ions. Finally, Leach et al. 1 1 succeeded in determin- ing Fe at concentrations between 3.6 ppb and 0.71 pprn by using standard solutions in experiments mainly aimed at evaluating sample injection and differential detection. All of the above methods rely on the use of l,l0-phenanthroline or one of its derivatives as the complexing agent.In previous work'? a thermometric method was developed for the determination of Fe based on the catalytic effect of the Fe"'-ethyIenediaminetetraacetic acid (EDTA) complex (Fey-) on the oxidation of hydroxylamine by dissolved oxygen in a basic medium, which takes place according to the following reaction: * To whom correspondence should be addressed. As the reaction yields nitrite ions, it was envisaged that it could also be applied to the spectrophotometric determination of Fe by coupling it to the well-known Griess reaction.In this work, and in a continuation of our research on the automation of analytical methods,13-16 a semi-automated FI system was designed for the spectrophotometric determina- tion of trace amounts of Fe on the basis of the above reaction, which, as predicted, provided increased reproducibility, sensitivity and sample throughput. The system was controlled by a personal computer and appropriate software which permitted: ( a ) commanding the automatic injection of sam- ples; (b) acquisition of the spectra obtained by the diode array detector; ( c ) continuous recording of the spectra in addition to the FI recordings; and ( d ) processing of the FI peaks for obtaining the sample concentrations by automatic interpola- tion of their heights on the calibration graph.The spectra used for the determination were obtained by subtracting the background signal due to the colour-forming, non-catalytic oxidation of hydroxylamine from the readouts recorded on passage of the sample through the flow cell. Likewise, the FI recordings reflected the changes with time in the differences between the signal at the absorption maximum (542 nm) and that at another wavelength at which readings were made on the baseline (700 nm). Experimental Reagents All reagents were of analytical-reagent grade and all solutions were prepared in distilled water. Phosphate buffer, 0.1 mol 1- 1. Prepared from dipotassium hydrogen phosphate (Merck) and 1.75 x 10-3 mol I-' EDTA solution obtained from the disodium salt (Probus); the pH was adjusted to 12.5 with NaOH solution.Iron(1Ir) standard solution containing 1000 pg ml-1 of Fe"' in 1% H N 0 3 . Prepared from iron(rI1) nitrate (Probus). Hydroxylamine solution, 0.05 mol 1-1. Prepared from the hydrochloride (Probus) and neutralized with NaOH solution. Sulphanilamide (SPA) (Merck) and N-( 1-naphthy1)ethyl- enediamine (NED) (Merck) solutions containing 0.3% of either substance in 0.8 mol 1-1 HCl (Merck). The last three solutions were freshly prepared each day in order to avoid baseline noise as far as possible.914 ANALYST, SEPTEMBER 1991, VOL. 116 ml min-1 mostatic bath -------- Peristaltic Pump SD Waste U Fig. 1 Flow injection manifold used for the direct determination of Fe. I = Sample injector; SD = spectrophotometric detector; lI = 5 m; l2 = 1.5 m; l3 = 2 m; [EDTA] = 1.75 x 10-3 moll-1; [NH20H] = 0.05 mol 1-1; [SPA] = 0.3%; [NED] = 0.3%; and T = 45°C Apparatus A customized semi-automated system equipped with an IBM-compatible computer for controlling injections and acquiring and processing the spectrophotometric data was used.Fig. 1 shows a schematic diagram of the FI manifold, which consisted of an eight-way Gilson Minipuls 3 peristaltic pump, a Rheodyne 50 injection valve which was controlled by the computer via a mechanical actuator, a Techtron thermostatic bath and a Hewlett-Packard (HP) 8452A diode array spectro- photometer furnished with a flow cell with a pathlength of 10 mm and a void volume of 18 y1. All reactors and the injection loop were made from poly( tetrafluoroethylene) (PTFE) tubing of 0.5 mm i.d.Data Acquisition and Processing A special program was developed for instrumental control and data acquisition and processing that takes advantage of the full potential of the HP 8452A diode array detector for the rapid acquisition of spectra and optionally also allows complete spectra to be acquired at every point along the FI recording and absorbances measured at different wavelengths to be used. In order to take full advantage of these assets and reduce noise from the spectrophotometer lamp as far as possible, the FI recordings were obtained by assigning each point as the difference between the absorbance measured at the maximum absorption wavelength (542 nm) and that at a zero-absorption wavelength (700 nm).Such values were calculated by the computer and stored in a file for subsequent use. Optionally, all the acquired spectra can be stored in another file, which can be used for assaying different measurement wavelength combinations without the need to repeat the experiments. The program also automatically processes the FI recordings obtained by detecting peaks, determining the baseline, calculating peak heights and areas, constructing calibration graphs and calculating the required concentrations. Peaks were detected by using the first two derivatives of the experimental data, which were calculated by the Savitzky- Golay algorithm. 17 If the first and second derivatives were larger than pre-set values, the corresponding point was taken as the start of the peak; also, the point at which such a condition was met was taken as the end of the peak.Then, the peak baseline was determined from points in front of and behind the peak, and the peak height was calculated as the distance between the peak maximum, the point at which the first derivative changed sign, and the previously calculated baseline; if the height was lower than a pre-set value, the peak in question was rejected. Next, the areas of the peaks were determined by the trapezium method and the concentration corresponding to each peak was calculated by interpolation of the calibration graph, which was also constructed by using the program from one or several standard FI recordings on the basis of peak heights or areas according to the user's choice. Procedure The determination was started by circulating the different reagent streams through the FI system in order to allow the background colour due to the non-catalytic oxidation of hydroxylamine to develop.Once the absorbance was stable, the blank signal was obtained and assigned a zero value in the spectral recording. Then, the baseline of the FI recording was obtained and the program commanded injection of the sample (200 yl), the Fe concentration of which must be within the linear determination range. After the residence time had elapsed, a peak reflecting the colour increase resulting from the catalytic effect of the FelII-EDTA complex on the oxidation of hydroxylamine was obtained. The corresponding FI recording was stored in a pre-set file and processed in order to obtain the peak height over the baseline and the injected sample concentration by interpolation of the height on the calibration graph.Results and Discussion Optimization of Variables As previously found for the thermometric method,12 the maximum absorbance was obtained by using a 0.1 mol 1-1 phosphate buffer of pH 12.5 as the medium for the catalysed reaction. Other buffers consisting of borax or NH4+-NH3 yielded poorer results, while tris(hydroxymethy1)amino- methane provided approximately the same readings as those obtained with the phosphate buffer. The spectrophotometric signal yielded by the azo dye was found to remain constant at HC1 concentrations above 0.8 moll-', which was therefore chosen as the working concentra- tion. The spectrophotometric signals provided by Fell1 standards were identical with those yielded by Fell solutions prepared from Mohr's salt; hence all the Fe present under the working conditions was in the Fe"' form.The simplex method18 was used to optimize simultaneously a set of eight parameters consisting of the three reactor coil lengths (Il, 12 and 13) and five chemical variables, viz., the concentrations of EDTA, hydroxylamine, SPA and NED, and the temperature of the thermostated bath in which the three reactors were immersed. The simplex programme was run by using one-directional simple advancement, contraction and no quadratic interpolation. The simplex optimization was performed by measuring the height of the peak obtained on injection of 100 p1 of a 50 ng ml-1 standard of Fell1 and plotting the difference between the absorbance of the catalysed and uncatalysed (baseline) reaction.A maximum working temperature of 45 "C was used in order to avoid an excessively noisy baseline. The optimum conditions thus determined are shown in Table 1 and correspond to step 11 in Fig. 2, which shows the variation of the signal as a function of the simplex evolution. Finally, the variation of the signal as a function of the injected sample volume was studied. It was found that while the signal increased by 40% when 150 rather than 100 p1 was used as the injected volume, an increase in the injected volume, from 150 to 200 pl increased the signal by only 25%. In order to avoid potential diffusion problems and an inordinate decrease in the sample throughput, it was decided not to use larger injected volumes.After the optimum working conditions had been establi- shed, the linear determination range was evaluated, and was found to be between 3.5 and 150 ng ml-1 of Fe, the detection limit being 2.0 ng ml-1. The equation obtained was: A = 0.019ANALYST, SEPTEMBER 1991, VOL. 116 915 ~ ~ ~~ Table 1 Optimum working conditions as established by the simplex method [EDTA]/ [NH,OH]/ [SPA] [NED] Tl Ill l ~ l 131 ' C m m m Normal FI 1.75 x 10-3 0.05 0.3 0.3 45 5.0 1.5 2.0 Reversed FI 1.85 x lo-' 0.01 0.5 0.5 20* 4 0 0 moll-' moll-' (Yo) (Yo) * Room temperature. 0.08 - 0.06 - aJ C (D .f! a 2 0.04 . I) 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 Simplex number Fig. 2 Variation of peak height as a function of the simplex evolution. [Fe"'] = 50 ng ml-1; and injected sample volume = 100 pl 1.10 I p) 0.90 s 0 (D .f! 0.70 D 4 0.50 0.30 0.10 c E L -0.10 ' I I I I I 0 5 10 15 20 25 Time/mi n Fig.3 Flow injection recording obtained by using the normal FI manifold for the construction of the calibration graph from triplicate injections ( V , = 200 pl) of A, 5; and B, 10 ng ml-1 of Fe'" and duplicate injections of C, 25; D, 50; E, 100; and F, 150 ng ml-1 of Fe"l. The working conditions used are given in Table 1 + 7.52 x lo-3c; (Y = 0.9992), where c is the FelI1 concentration in ng ml-1 (Y = correlation coefficient). The detection limit was taken as five times the standard deviation of the base line noise ( o = 2.5 x lO-3), and the quantification limit as ten times the same parameter. Fig. 3 shows six points of the corresponding FI recording.The repeatability of the proposed method was found to be 6.7 and 2.4% for 5 and 50 ng ml-1 of Fe"' (n = ll), respectively. In order to compare the working conditions established by the procedure described above with those of a reversed FI system, the manifold shown in Fig. 4 was used, the representa- tive variables of which were also optimized by the simplex method, but using the ratio of the signals yielded by the catalysed and uncatalysed reaction (i.e., the peaks obtained by injection of an Fe standard and distilled water, respectively, through the sample channel) instead; this ensured that the highest sensitivity in the lowest height of the peak blank was [ NED Om60 Peristaltic Pump L Fig. 4 Reversed FI manifold used for the determination of Fe.I I = 4 m; [EDTA] =1.85 x 10-3 mol 1-l; [NH20H] = 0.01 mol 1-l; [SPA] = 0.5%; and [NED] = 0.5% obtained. The conditions thus established are also sum- marized in Table 1. As the increased sensitivity obtained also provided a more convenient operational procedure resulting from the possibility of using the baseline as the blank, it was decided to use the manifold depicted in Fig. 1 for all subsequent experiments. Effect of Foreign Ions The potential interfering effect of various ions that were added to solutions containing 50 ng ml-1 of Fell1 in different proportions in relation to the analyte concentration was studied. Table 2 lists the results obtained. A given ion was considered to interfere with the determination if it resulted in a signal variation greater than k20.As can be seen, few ions were found to interfere with the proposed method. The most serious interferences were from Co", Crlll and Cu", which increased the height of the peak yielded by the analyte alone. The interference from the divalent ions at concentrations up to 40 times higher than that of the analyte was readily overcome by injecting 1% ammonia solution (Merck) into the stream containing EDTA and the buffer; direct incorporation in the sample caused the analyte to precipitate. On the other hand, the interference from Crlll was eliminated at interferent-to- analyte ratios below 10 if ammonia solution was added to the EDTA channel and 2 ml of 0.05 moll-' tartrate solution were added to the sample. The weak interference from Ca and Mg was investigated in detail because of their significance in the analysis of very hard waters.This interference was found to arise from the complexation of the two alkaline earth metal ions by EDTA; hence, relatively large concentrations of these two ions will lead to an insufficient concentration of EDTA for complete formation of the FeI'I-EDTA complex, thereby preventing catalysis of the oxidation of hydroxylamine. The interfering effect of Ca and Mg can be halved by doubling the EDTA concentration used or, provided the analyte concentration allows, by diluting the sample; even if the interferent-to- analyte ratios remain constant, the amount of EDTA removed from the medium will obviously be reduced by 50%. Fig. 5 shows the interfering effect of Ca and Mg ,on the determina-916 ANALYST, SEPTEMBER 1991, VOL.116 Table 2 Tolerated foreign ion-to-analyte ratios in the determination of 50 ng ml-1 of Fellr Foreign ion [Ion] : [Fe"'] C1- , NO3-, SO&, C032-, Ca", Ball, Pb" 1000* Mg", Cd", Zn", NilL, Al"' 400 Mn" 5 CU" 1 (4O)t CO" 0.5 (40)t Cr"' 0.5 (lo)$ * Maximum concentration tested. t In the presence of 1% ammonia solution. $ In the presence of 1% ammonia solution and 2 x 10-3 mol 1-1 tartrate. 0.25 0 -0.25 Q) -0.50 * s a Q -0.75 5.0 7.0 9.0 11.0 13.00 1 (b) 0.25 -0.25 15.0 16.0 17.0 18.0 19.0 Tim e/m i n Fig. 5 Determination of Fe in tap water. (a) Interference from Ca and Mg. (b) Elimination of the two interferents by a 1 + 5 dilution of the sample tion of Fe in hard tap water and the elimination of such an effect by diluting the original sample five-fold.Applications The proposed method was applied to the determination of Fe in waters from various sources and in white wines. For drinking waters, international legislation classifies Fe as an undesirable component. Hence, the proposed method is interesting because it allows the determination of this metal in a simple and fast way within legally permitted ranges. Table 3 lists the dilutions used and the results obtained by using both the proposed method and a standard atomic absorption procedure, with standard additions, as reference. 19 When a graphite furnace was used the ashing and atomizing temperatures were optimized (1200 and 2100 "C, respectively) and magnesium nitrate was added as a chemical modifier. As can be seen, the results obtained were consistent.In all the water samples analysed, the Fe concentrations found were below the maximum level allowed by the legislation (up to 200 ng ml-1). Table 3 Results obtained for the determination of Fe in real samples [Fe"']/pg ml-1 Sample Dilution Mineral water - Well water Tap water - (Palma de Mallorca) Zone 1 1 + 4 Zone 2 1 + 4 1 + 4 Zone 3 Zone 4 1 + 4 White wine 1 + 99 * Reference 19. t Using a graphite furnace. Proposed method 0.0043 0.017 0.10 0.072 0.047 0.059 2.8 Atomic absorption method* 0.0044.l 0.018t 0.1 I t 0.079t 0.045.l 0.065t 2.4 However, application of the proposed method to ros6 and red wines provided different results from those yielded by the reference method, probably because of the presence of an additive, the interference from which could not be overcome by masking or mineralization.Conclusions The working conditions used in the proposed method for the spectrophotometric determination of Fe in an FI system are very similar to those employed in the previously reported thermometric procedure12 as far as the development of the catalytic reaction involved is concerned. Operationally, the spectrophotometric method is more sensitive, reproducible and rapid than its thermometric counterpart, although the latter is subject to less serious interferences and hence is more selective. This is consistent with the features of the two techniques used. In addition, the thermometric method does not require the use of a coupled reaction; hence it is subject to fewer background perturba- tions.The normal FI configuration used was found to provide better results than the reversed FI manifold tested, partly because of the colour arising from the non-catalytic reaction taking place in the latter instance. The proposed method has a linear determination range with a lower limit similar to the lowest reported in the literature," but with the added advantage of being applicable to real samples. The authors thank the DGICyT (Spanish Council for Research in Science and Technology) for financial support granted for the realization of this work as part of Project PA 86-0033. References 1 RGiiCka, J., and Hansen, E. H., Anal. Chim. Acta, 1975, 78, 145. 2 RGiiEka, J., and Hansen, E. H.. Flow Injection Analysis, Wiley, New York, 1981. 3 RfiiiCka. J., and Hansen, E. H., Anal. Chim. Acta, 1986,179,l. 4 Valcarcel, M., and Luqcle de Castro, M. D., Flow Injection Analysis: Principles and Applications, Ellis Horwood, Chi- Chester, 1987. Quim. Anal., 1989, 8(2), Special Issue. Mortatti, J., Krug, F. J., Pessenda, L. C. R., Zagatto, E. A. G., and Jorgensen, S. S., Analyst, 1982, 107, 659. Bubnis. B. P., Straka, M. R., and Pacey, G. E., Talanta, 1983, 30, 841. Lynch, T. P., Kernoghan, N. J., and Wilson, J . N., Analyst, 1984, 109, 843. 5 6 7 8ANALYST, SEPTEMBER 1991, VOL. 116 917 9 10 11 12 13 14 15 Faizullah, A. T., andTownshend, A.. Anal. Chim. Acta, 1985, 167, 225. Rios, A., Luque de Castro, M. D.. and Valcarcel, M.. Analyst, 1985, 110, 277. Leach, R. A., RfiiiEka, J., and Harris, J. M., Anal. Chem., 1983, 55, 1669. Gomez. E., Estela, J. M., and Cerda, V., Thermochim. Acta, 1991, 176, 121. Cerda, V., and Ramis, G., An Introduction to Laboratory Automation, Wiley, New York, 1990. Maimo. J.. Cladera, A., Mas, F., Forteza, R., Estela, J. M., and Cerda, V., Int. J. Environ. Anal. Chem., 1989, 35, 161. Cladera, A., Caro, A.. Estela, J. M., and Cerda, V., Int. J. Environ. Anal. Chem., 1990, 43, 11. 16 Cladera, A., Estela, J. M., and Cerda, V., J. Autom. Chem.. 1990, 12, 108. 17 Savitzky, A., and Golay, M. J . E., Anal. Chem., 1964,36,1627. 18 Morgan, L., and Deming, S. N., Anal. Chem., 1974, 46, 1170. 19 Perkin-Elmer Handbooks: Analytical Methods for Atomic Absorption Spectrophotometry and HGA -400 Graphite Fur- nace. Operator’s Manual, Perkin-Elmer, Norwalk, CT, 1984. Paper 1 /00703C Received February 14th, 1991 Accepted April 25th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600913

出版商:RSC    

年代:1991

数据来源: RSC