ISSN:0003-2654    

DOI:10.1039/AN98712BP001

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

The AnalystThe Analytical Journal of The Royal Society 0: ChemistryAdvisory Board*Chairman: J. D. R. Thomas (Cardiff, UK)J. F. Alder (Manchester, UK)D. Betteridge (Sunbury-on-ThanE. Bishop (Exeter, UK)*C. Burgess (Ware, UK)D. T. Burns (Belfast, UK)G. D. Christian (USA)*M. S. Cresser (Aberdeen, UK)L. de Galan (The Netherlands)*A. G. Fogg (Loughborough, UK)*C. W. Fuller (Norringham, UK)V. D. Goldberg (London, UK)T. P. Hadjiioannou (Greece)A. Hulanicki (Poland*C. J. Jackson (London, UK)*P. M. Maitlis (Sheffield, UK)E. J. Newman (Poole, UK)T. B. Pierce (Hawell, UK)Tes, UK) E. Pungor (Hungary)J. ReiiEka (Denmark)R. M. Smith (Loughborough, UK)W. I. Stephen (Birmingham, UK)M. Stoeppler (Federal Republic of Germany)K. C. Thompson (Sheffield, UM*A.M. Ure (Aberdeen, UK)A. Walsh, K.B. (Australia)G. Werner (German Democratic Republic)T. S. West (Aberdeen, UK)*P. C. Weston (London, UK)J. D. Winefordner (USA)Yu. A. Zolotov (USSR)P. Zuman (USA)'Members of the Board serving on the Analytical Editorial BoardRegional Advisory EditorsFor advice and help to authors outside the UKDr. J. Aggett, Department of Chemistry, University of Auckland, Private Bag, Auckland, NEWDoz. Dr. sc. K. Dttrich, Analytisches Zentrum, Sektion Chemie, Karl-Marx-Universitat, Talstr.Professor L. Gierst, Universit6 Libre de Bruxelles, Facult6 des Sciences, Avenue F.-D.Professor H. M. N. H. Irving, Department of Analytical Science, University of Cape Town,Dr. 0. Osibanjo, Department of Chemistry, University of Ibadan, Ibadan, NIGERIA.Dr.G. Rossi, Chemistry Division, Spectroscopy Sector, CEC Joint Research Centre,Dr. 1. RubeSka, Geological Survey of Czechoslovakia, Malostransk6 19, 118 21 Prague 1,Professor K. Saito, Coordination Chemistry Laboratories, Institute for Molecular Science,Professor M. Thompson, Department of Chemistry, University of Toronto, 80 St. GeorgeProfessor P. C. Uden, Department of Chemistry, University of Massachusetts, Amherst,Professor Dr. M. Valchrcel, Departamento de Quimica Analitica, Facultad de Ciencias,Professor Yu Ru-Qin, Department of Chemistry and Chemical Engineering, Huna'n University,ZEALAND.35, DDR-7010 Leiptig, GERMAN DEMOCRATIC REPUBLIC.Roosevelt 50, Bruxelles, BELGIUM.Rondebosch 7700, SOUTH AFRICA.EURATOM, lspra Establishment, 21 020 lspra (Varese), ITALY.CZECHOSLOVAKIA.Myodaiji, Okazaki 444, JAPAN.Street, Toronto, Ontario M5S 1A1, CANADA.MA 01003, USA.Universidad de Cbrdoba, 14005 Chrdoba, SPAIN.Changsha, PEOPLES REPUBLIC OF CHINA.Editor, The Analyst:Philip C.WestonSenior Assistant Editors:Judith Brew, Roger A. YoungAssistant Editors:Anne Horscroft, Pamil SehmiEditorial Office: The Royal Society of Chemistry, Burlington House,Piccadilly, London, W1V OBN. Telephone 01-734 9864. Telex No. 268001Advertisements: Advertisement Department, The Royal Society of Chemistry, BurlingtonHouse, Piccadillv, London, W1V OBN. Telephone 01-437 8656. Telex No. 268001The Analyst (ISSN 0003-2654) is published monthly by The Royal Society of Chemistry,Burlington House, London W1V OBN, England.All orders accompanied with payment shoulcbe sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road,Letchworth, Herts. SG6 lHN, England. 1987 Annual subscription rate UK f160.00, Rest 01World f179.00, USA $315.00. Purchased with AnalyticalAbstracts UK f364.00, Rest of Workf403.00, USA $709.00. Purchased with Analytical Abstracts plus Analytical Proceedings UKf411 .OO, Rest of World f455.00, USA $801 .OO. Purchased with Analytical Proceedings UKf200.00, Rest of World f224.00, USA $394.00. Air freight and mailing in the USA byPublications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003.USA Postmaster: Send address changes to: The Analyst, Publications Expediting Inc., 20CMeacham Avenue, Elmont, NY 11003.Second class postage paid at Jamaica, NY 11431. Allother despatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Posloutside Europe. PRINTED IN THE UK.Information for AuthorsFull details of how to submit material forpublication in The Analyst are given in theInstructions to Authors in the January issue.Separate copies are available on request.The Analyst publishes papers on all aspects ofthe theory and practice of analytical chemistry,fundamental and applied, inorganic andorganic, including chemical, physical, biochem-ical, clinical, pharmaceutical, biological, auto-matic and computer-based methods. Papers onnew approaches to existing methods, newtechniques and instrumentation, detectors andsensors, and new areas of application with dueattention to overcoming limitations and to un-derlying principles are all equally welcome.There is no page charge.The following types of papers will be con-sidered:full papers, describing original work.Short papers: the criteria regarding origin-ality are the same as for full papers, but shortpapers generally report less extensive investi-gations or are of limited breadth of subjectmatter.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.Reviews, which must be a critical evaluationof the existing state of knowledge on a par-ticular facet of analytical chemistry.Every paper (except Communications) wilt besubmitted to at least two referees, by whoseadvice the Editorial Board of The Analystwill beguided as to its acceptance or rejection. Papersthat are accepted must not be pubiished else-where except by permission. Submission of amanuscript will be regarded as an undertakingthat the same material is not being consideredfor publication by another journal.Regional Advisory Editors.For the benefit ofpotential contributors outside the United King-dom, a Panel of Regional Advisory Editorsexists. Requests for help or advice on anymatter related to the preparation of papers andtheir submission for publication in The Analystcan be sent to the nearest member of the Panel.Currently serving Regional Advisory Editors arelisted in each issue of The AnalystManuscripts (three copies typed in double spac-ing) should be addressed to:The Editor, The Analyst,Royal Society of Chemistry,Burlington House,Piccadilly,LONDON W1V OBN, UKParticular attention should be paid to the use ofstandard methodsof literature citation, includingthe journal abbreviations defined in ChemicalAbstracts Service Source Index. Wherever pos-sible, the nomenclature employed should fol-low IUPAC recommendations, and units andsymbols should bethose associated with SI.All queries relating to the presentation andsubmission of papers, and any correspondenceregarding accepted papers and proofs, shouldbe directed to the Editor, The Analyst (addcessas above). Members of the Analytical EditorialBoard (who may be contacted directly or via theEditorial Office) would welcome comments,suggestions and advice on general policy mat-ters concerning The Analyst.Fifty reprints of each published contribution aresupplied free of charge, and further copies canbe purchased.@ The Royal Society of Chemistry, 1987. 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/AN98712FX005

出版商:RSC    

年代:1987

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98712BX007

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 113 Electrode Membrane and Solvent Extraction Parameters Relating to - the Potentiometry of Polyalkoxylates P. H. V. Alexander, G. J. Moody and J. D. R. Thomas Department of Applied Chemistry, Redwood Building, UWIST, P.O. Box 13, Cardiff CFI 3XF, UK Poly(viny1 chloride) (PVC) matrix membrane electrodes containing various amounts and compositions of liquid ion-exchanger were studied in order to investigate their response to aqueous solutions of Antarox CO-880 as a model polyalkoxylate. The liquid ion-exchanger used in the membrane was based on solutions in 2-nitrophenyl phenyl ether of the tetraphenylborates of complexes with barium of either Antarox CO-880 or Antarox CO-430 polyethoxylates. Barium ion-responses were also noted. The electrodes recommended for use with polyalkoxylates in aqueous solutions are formed from an Antarox CO-430 based liquid ion-exchanger.The master membrane used for making the electrodes should consist of 0.40 g of the liquid ion-exchanger in 0.17 g of PVC, that is 70% of sensor in 30% of the immobilising polymer. Neither Antarox CO-430 (Series A) nor Antarox CO-880 (Series 6) type electrodes respond to the methyl glucoside polypropoxylates Glucam P10 and P20, but, conversely, Glucam PI0 and P20 (Series C) type electrodes respond more effectively to Antarox CO-880 polyethoxylate in aqueous solution than they do to either of the propoxylates. This and the better response to Glucam P20 compared with Glucam P10 is consistent with the relative stability of metal - polyalkoxylate complexes according to data obtained on the solvent extraction of metal ions into dichloromethane by the polyalkoxylates, based on studies of the large polarisable picrate and dipicrylamine anions.Solvent extraction data show that the complexing tendencies of the various polyalkoxylates studied with barium, magnesium and zinc ions are generally in the order of Antarox CO-730 > PEG 1500 > Antarox CO-880 > Glucam P20 > Glucam P10 > Antarox CO-430. There is a correlation between these and electrode qualities for cations and also with regard to the response of these electrodes towards alkoxylates. However, other factors can also be involved such as those impressed by the substituent tail attached to the end of the alkoxylate chain. Keywords: Non-ionic surfactants; alkoxylates; barium ion-selective electrodes; critical micelle concentration; solvent extraction The potentiometric response to polyalkoxylates of poly- (vinylchloride) (PVC) matrix membrane electrodes based on sensors of the tetraphenylborates (TPB) of complexes of barium with various polyalkoxylates has already been charac- terised.1-3 Thus, a membrane system based on barium complexed with Antarox CO-880 [a nonylphenoxypoly- ethoxylate (NP) with 30 ethoxylate units (EOU)], which usually functions as a barium ion-selective electrode, responds potentiometrically to various polyalkoxylates, showing linear relationships between e.m.f.and log[alkoxylate].l There are usually breaks in the linearity that are attributed to the critical micelle concentration (CMC) of this class of non-ionic surfact ants .1 Further studies of the potentiometric behaviour of other barium - polyethoxylate complexes in electrode membranes, that is, barium complexes with Antarox CO-430 (an NP with four EOUs), Antarox CO-730 (an NP with seven EOUs), Lutensol-A07 (an ethoxylate with seven EOUs and C13H27- C15H31 as the hydrophobe) and Monflor-51 (an ethoxylate with 23 EOUs and CsFls at each end of the ethoxylate chain) have shown the barium - Antarox CO-430 system to be superior in its response to non-ionic surfactants.2 All four alternative electrode types are inferior to the Antarox CO-880 system for sensing barium ions.2 The Antarox CO-430 electrode system has been evaluated for the determination of Dobanol 25-7, Synperonic 7 and Lutensol-A07 in detergent powders and good recoveries of the non-ionic surfactants were observed.2 In each instance, the electrode membranes were based on ion-exchangers consisting of saturated solutions of the TPB of the barium - alkoxylate complex in 2-nitrophenyl phenyl ether.The PVC matrix master membranes consisted of 0.40 g of the ion-exchanger with 0.17 g of PVC.4,5 However, there is a need to determine whether this is the best choice of membrane composition and also whether propoxylates, as alternatives to the ethoxylate systems mentioned above, are appropriate sensor membrane components. This paper describes studies carried out by varying the proportions of membrane components in electrode systems based on the polyethoxylates Antarox CO-430 (Series A electrodes) and Antarox CO-880 (Series B electrodes).For the studies involving polypropoxylate electrodes (Series C), methyl glucoside polypropoxylates { CH3C6Hlo05[CH- (CH3)CH20InH} were selected on the basis of the availability of systems with two different lengths of propoxylate chains, namely, Glucam P10 and Glucam P20 with 10 and 20 propoxylate units, respectively. Although the primary objectives of this study were to determine the appropriate compositions for a response towards alkoxylates in solution, the scope of complexes of the propoxylate ionophores with cations smaller than barium, that is, magnesium and zinc, as cation sensors was also studied. Finally, in an attempt to elucidate the mechanism of the potentiometric response, some studies were also made of solvent extraction by the polyalkoxylates of metal ions from water into dichloromethane.The principal characteristics of the polyalkoxylates dis- cussed are summarised in Table 1. Experimental Materials and Reagents Members of the Antarox series of ethoxylate non-ionic surfactants were gifts from GAF (Great Britain) (Manchester, UK) and the Glucam polypropoxylates were gifts from D. F.114 ANALYST, FEBRUARY 1987, VOL. 112 Table 1. Features of the polyalkoxylates studied in this work Nature and number of alkoxylate units Commercial name* (AOUs) AntaroxCO-430 . . Ethoxylate (4) AntaroxCO-730 . . Ethoxylate (15) AntaroxCO-880 . . Ethoxylate (30) PEG1500 . . . . Ethoxylate (34) GlucamP10 . . . . Propoxylate (10) GlucamP20 . . . . Propoxylate (20) * Listed in reference 6.Nature of h ydrophobe Non ylphenox y Nonylphenoxy Nonylphenoxy None Methylglucoside Meth ylglucoside Average relative molecular mass 396 880 1540 1500 775 1355 Ba2+ : AOU ratio 4 15 12 10.5 10 8 Anstead (Billericay, Essex, UK). All other materials and reagents were of the best available laboratory-reagent grade. Preparation of Complexes The metal - polyalkoxylate metal chloride complexes were prepared in the manner previously described,49778 that is, by adding 0.1 M aqueous solutions (10 cm3) of the metal chloride to 0.01 M polyalkoxylate solutions (30 cm3). Saturated sodium tetraphenylborate was added to the solution of the complex and the resulting precipitate was filtered, washed with water and finally dried in a vacuum oven. The stoicheiometries of the alkoxylate - metal - TPB products were determined by NMR spectrometry as previously described,7 using a Perkin- Elmer Model R32 (90 MHz) NMR spectrometer.Electrode Membranes These were based on various amounts of liquid ion-exchanger, taken with 0.17 g of PVC (for fabricating the master membranes according to established proced~res4~5) for assembly into electrodes. The liquid ion-exchanger consisted of a saturated solution of the TPB of the metal complex with the appropriate polyalkoxylate in 2-nitrophenyl phenyl ether (NPPE). In some instances, the liquid ion-exchanger was diluted with NPPE for the fabrication of the master mem- brane, whereas in others 0.04 g of the appropriate TPB complex was taken with 0.36 g of a mixture (1 + 1 V/V) of NPPE and dioctylphenyl phosphonate.Assembly and Operation of Electrochemical Cells The poly(viny1 chloride) matrix membrane electrodes were assembled as described previ0usly4~5 from discs cut from the master membranes. The inner electrode reference solution was a 0.1 M solution of the appropriate metal chloride for the metal ion-sensing electrodes and 1 M metal chloride solution for the alkoxylate surfactant responsive electrodes, except for the zinc-responsive electrodes in which zinc nitrate solutions with small added volumes of saturated zinc chloride were used. The electrodes were used in conjunction with a low-leak (approximately 0.01 cm3 h-1) calomel reference electrode (Corning 476 10 900). The cell e.m.f.s were recorded to k0.5 mV of the steady values for stirred solutions at 25 k 0.1 "C with a high-impedance millivoltmeter (Corning, Model 112 or Radiometer PHM 64) reading to 0.1 mV and used in conjunction with a Servoscribe Model RE 541.20 potentio- metric chart recorder.The metal ion-sensing studies were carried out using cation standards prepared by serial dilution (if necessary) of the appropriate 0.1 M metal chloride solutions, or of zinc nitrate for the zinc ion calibrations. The activities of the metal ions were based on the activity coefficient (f) data calculated from the following modified form of the Debye - Hiickel equation, which is applicable to any ion. logf= - 2 2 - AvI -0.2) (1 + VI . . . . where A is the Debye - Hiickel constant, I is the ionic strength and z the valency. Alkoxylate surfactant response studies were made with 10-2 M surfactant standards.Appropriate known aliquots of the relevant standards were spiked into 25 cm3 of doubly de-ionised water. As previously reported for alkoxylate responses ,1J the e.m.f. data for summarising the responses of the electrodes to alkoxylate surfactants were for second and subsequent read- ings, the first reading being used for guidance only. Whereas the metal ion e.m.f. response readings were relatively fast, those for the alkoxylate surfactant test solutions were gener- ally slow; hence the e.m.f. readings for surfactant responses were taken 5 min after the addition of the previous surfactant spike. The increases in the cell e.m.f. for surfactant solutions corresponded to the difference between the recorded e.m.f.and that for the steady base line corresponding to that of the electrode immersed in water (water was used throughout this study). Solvent Extraction Procedures These were carried out for Antarox CO-430,730 and 880 and PEG 1500 polyethoxylates and Glucam P10 and P20 poly- propoxylates and followed the pattern previously described8 for the partitioning of the picrates of metal ion - poly- ethoxylate complexes from aqueous chloride solutions between water and dichloromethane. This method is based itself on metal ion complexation by macrocyclic polyethers9 and macrotretrolide antibiotics. 10.11 Theory The extraction equilibrium involving a lipophilic coloured anion A-, such as picrate, based on a 1 mol proportion of metal cation M2+ of valence z corresponds to . .Mz+ Y - L* + zA- ML;L* + zA-* n for y alkoxylate units (AOU) per metal cation. In equation (2), L is the neutral ion binding molecule (the polyalkoxylate) and n is the number of AOUs (or polyether oxygens) per mole of L. Thus, for Antarox C-880 (with 30 ethoxylate units per mole) complexes, n is 30 and y is 12 for the bivalent barium, magnesium and zinc cations under consideration here .6J1 An asterisk (*) denotes species characteristic of the organic solvent phase and species not so designated refer to the aqueous phase. The bulk extraction constant, KIM, for reaction (2) is given by (3)ANALYST, FEBRUARY 1987, VOL. 112 115 where the a terms are the activities of the species denoted by subscripts representing the main participants in reaction 2. Other possible equilibria can either be discounted or allow- ance made for them, such as any ion-pair formation between the ML$* species and the A- anions (reaction 4) in solvents of low permittivity (E is 9.05 for dichloromethane) ML$* + zA-* = (MLy/, A,)* .. . . (4) Reactions 2 and 4 adequately represent the relevant equilib- ria9J0 K* as the ion-pair formation (reaction 4) is given by . . . . (6) KM(~M> (uA)' ( ~ L * ) Y ' ~ Assuming ideal behaviour of all species in the organic solvent phase and all the neutral species in the aqueous phase, the activity terms can be replaced by concentrations such that - ~ M L A * - and As CA* = zCbL, equation 8 reduces to which is taken to represent the solvent extraction equilibrium as previous data on picrate extractions involving poly- ethoxylatess indicate negligible ion pairing (see Figs.1 and 2 of reference 7). Replacing the activity (a) terms in equation 9 by the activity coefficient u> times concentration (C) gives Table 2. Summary of response of potentiometric PVC matrix membrane electrodes of TPBs of Antarox ethoxylate : barium complexes to barium ions and Antarox CO-880 in aqueous solutions Membrane compo- Electrode or sition (liquid membrane ion-exchanger added Response to number to 0.17 g of PVC/g barium ions A. Electrodes based on Antarox CO-430: 1 2 3 4 5 6 7 8 9 10 11 12 0.05 0.20 0.30 0.40 0.101 0.55 0.40 + 0.20 SM* 0.27 + 0.13 SM 0.04 Complex TPB + 0.36 MSM* 0.04 Complex TPB + 0.36 DOPP* Unsatisfactory Poor Poor Poor Good. Slope: 28-29 mV Poor. Linearity: 10- 3-10- 1 M .Long response times Slope: 27.5-29.5 mV for 5 x 10-5-10-7~ Very good. Poor B. Electrodes based on Antarox CO-880: 13 14 ::?;} 15 0.20 16 0.30 17 0.40 21 0.20 + 0.20 SM No useful response Very poor. Severe Similar to 15 but Very good. Response drift. Slope: 20-25 mV no drift times d 1 min. Slope: 26-28 mV for 10 -'-lo- M Good but response time 27.5-30.0 mV Similar to 16 2 5 min. Slope: 22 0.27 + 0.13 SM Very good. Similar to 17 23 0.4 + 0.20 SM Similar to 16 but slow in 24 0.04 Complex TPB Poor. Slow in response. 25 0.04 Complex TPB Similar to 15 response Linear 10-5-10-2 M + 0.36 MSM + 0.36DOPP Response to Antarox CO-880 None Poor CMC break Very good. Slope: 37 mV Very good and similar to 5. Slopes: 35-44 mV Poor Adequate but inferior to 5. Slope: 46 mV 1 I Poor None None Poor CMC break Poor CMC break Good, but inferior to 5.Slope: 42 mV Superior to 17 and similar to 5. Slopes: 37-44 mV Linear plots with no CMC breaks Linear plots with only occasional CMC breaks Similar to 17. Slope: 40 mV Linear plots but not as reproducible as 21 Poor CMC break and variable slopes Lifetime for continuous use/d 4-7 for CO-880 4-7 for CO-880 4-7 for CO-880 -30 for Ba2+ 4-7 for CO-880 -30 for Ba2+ - 4-7 for CO-880 4-7 for CO-880 -30 for Baz+ 4-7 for CO-880 -30 for Ba2+ 4-7 for CO-880 4-7 for CO-880 -30 for Ba2+ -30 for Ba2+ 4-7 for CO-880 4-7 for CO-880 4-7 for CO-880 * SM = Solvent mediator (2-nitrophenyl phenyl ether); MSM = mixed solvent mediator (Znitrophenyl phenyl ether - dioctylphenyl phosphonate, 1 : 1 VIV); DOPP = dioctylphenyl phosphonate.116 G O A 60 80 90 > E L 0 v) vj s 2040 50 0 c .- * E ui -20 0 10 ANALYST, FEBRUARY 1987, VOL.112 - - 1 1 -4 -3.5 Metal ion distribution coefficients, D M , may also be calculated from the concentrations of extracting anion, A-. As the anions in this work (based on picric acid and dipicrylamine) are univalent, then the amount of divalent metal ion extracted must satisfy the requirements of electro- neutrality and be equal to one half of the anion extracted. Thus, where Cm is the initial metal ion concentration taken, CM is the metal concentration in the aqueous phase after equilibra- tion, CM* is the metal concentration in the organic phase after equilibration and CA*/z is the concentration of the extracting anion in the organic phase divided by the metal ion valence, z .However, a more appropriate approach used in this work is the percentage of polarisable extracting anion, A- , trans- ferred by the metal - polyalkoxylate complex from the aqueous to the organic phase per mole of polyalkoxylate. Experimental Procedure The appropriate aqueous solutions containing the metal cation (5 cm3) and coloured anion (5 cm3), and the poly- Fig. 1. Increases in e.m.f.s for Antarox CO-880 added to water, for cells consisting of the best types of Series A and Series B electrodes (see Table 2) alkoxylate in dichloromethane (10 cm3) were mechanically shaken in a separating funnel in order to achieve equilibration. The optical absorbance of the coloured anion in the aqueous phase was determined with a Shimadzu Model 120-02 UV - visible spectrophotometer.Picric acid absorbances were measured at 357 and 378 nm for the aqueous and organic phases, respectively. The respective wavelengths for the dipicrylamine absorbances were 429 and 421 nm. The amount of coloured anion in the aqueous phase was calculated after allowance had been made for the reagent blanks. Results and Discussion Electrode Systems Based on Polyethoxylates Various amounts and compositions of liquid ion-exchanger were taken with 0.17 g of PVC in the master membranes (Table 2) and electrodes were prepared from discs cut from such membranes. The various membrane composition experi- ments were checked by replication. The e.m.f. responses of the resulting electrodes to barium ions and to water with added (spiked) Antarox CO-880 non-ionic surfactant are summarised in Table 2 for electrodes based on the TPBs of the Antarox CO-430 - barium complex (Series A) and the Antarox CO-880 - barium complex (Series B), respectively.Fig. 1 shows increases in cell e.m.f.s for the best types of Series A and Series B electrodes for Antarox CO-880 spiked into water. Response to barium ions The two series of electrodes showed marked differences; just two types of Series A electrodes (Types 9 and 11) responded well to barium ions compared with seven types of Series B electrodes (Types 16-22) , confirming previous observa- tions1-498 of the relative superiority of the barium - Antarox CO-880 complex sensor for barium ions. Good electrodes for barium ions are characterised by short response times (ca.1 min), lifetimes of ca. 30 d and responses of near-Nernstian slope down to ca. 10-5 M. A remarkable feature is the “reversed role” function of the electrodes of Types 11 and 24 compared to the pattern for Series A and B electrodes. Thus, Type 11 responds well to barium ions and poorly to Antarox CO-880, whereas Type 24 responds poorly to barium ions and moderately to Antarox CO-880 (Table 2). This effect is related to the partial replacement of the 2-nitrophenyl phenyl ether solvent media- tor with dioctylphenyl phosphonate, which leads to different solubility relationships of the sensor in the solvent mediator.4 1 I I I I I I -3.5 101 -4.5 -4 -3.5 -4 -3.5 -4 Log([Antarox C0-8801/~) Fig. 2. Illustration of inflection patterns for increases in e.m.f.s for Antarox CO-880 added to water, for cells consisting.of various types pf Se+s A and Series B electrodes (Table 2).(a), Group illustrating little or no inflection (Type 15); (b), group illustrating good Jnflection (Type 5 ) ; and (c), group illustratin gradual change of slope rather than a sharp inflection (Type 16). The illustrations are for newly-made electrodes (second runs) and the 8, A and 0 symbols represent electrodes from different master membranes in each groupANALYST, FEBRUARY 1987, VOL. 112 10 > E & a 5 - 117 - A contributory factor can also be the different degrees of complexation of the barium ions with Antarox CO-430 and CO-880. According to solvent extraction data of picrate complexes,8 the degree of complexation is less for lower relative molecular mass ethoxylates.Response to Antarox CO-880 Electrodes from membranes of Type 1-3 and 12-14 were of generally poor quality towards both barium ions and Antarox CO-880, with large variations in response for successive runs and between electrodes. For example, electrodes of Type 3 gave variations of up to 40 mV in the presence of 10-4 M added Antarox CO-880 compared with 4 5 mV for those of Types 5 and 17. Electrodes of Types 9, 11, 15 and 16 also gave poor responses towards Antarox CO-880. The relatively short lifetimes of the electrodes towards Antarox CO-880 are a general feature of non-ionic surfactant response. The inflections seen in the e.m.f. versus log[surfactant] calibrations, corresponding to the CMC, are of various degrees of prominence for the alkoxylates.l.2 Those for electrodes of Types 5-8, 10, 17-20 and 23 are discernible by characteristic abrupt changes of slope, and are indicative of good all-round electrode quality for non-ionic surfactants (see Fig.1 where CMC = 2.25 X 10-4 M, s.d. = 0.06 X M). The inflections of electrodes of Types 4, 15 and 16 are very difficult to locate because of a more gradual (or no) change of slope. However, it should be noted that even the best types show substantial variations in the extent of inflection between one electrode and another (Figs. 1 and 2). Response times towards non-ionic surfactants are generally longer2 than for normal ion-selective electrode types and normally exceed 10 min. In practical applications for the determination of non-ionic surfactants, electrodes are calibrated before and after each determination.2 In some instances, especially for Types 21,22,24 and 25, there were very small changes in slope so that the CMC breaks could not be located.Different kinds of patterns are illustrated in Fig. 2. The most functional systems in terms of response to Antarox CO-880 were of Types 5-8, 10, 17-20 and 23 (see Fig. 1). Although these are of nearly equal quality there seems to be no sound reason to deviate from the more straightfor- ward membrane composition of Type 5 for routine use in work with ethoxylates. Electrode Systems Based on Polypropoxylates Metal ion sensing The results for the Series C electrodes based on Glucam propoxylate ionophores and exposed to standards of barium, magnesium and zinc ions are summarised in Table 3.In no instance was a good response obtained for either magnesium or zinc, despite the more favourable polypropoxylate ligand, which on solvent extraction evidence (see under Solvent Extraction Studies) might have shown some potentiometric sensing properties for these ions. A barium ion response was evident in all instances (Table 3), although these did not Table 3. Responses of polypropoxylate and ethoxylate based electrodes (Series C) to barium, magnesium and zinc ions Responses to* Electrode or membrane number and type BaZ+ (BaC1,) MgZ+ (MgC12) Zn2+ [Zn(N03),] 26 (Glucam P20)0,4 Ba.2TPB Poor slope: -20 mV. Linear: 5 x 10-4-10-1 M -20 mV. Linear: 5 x 10-4-10-1 M -20 mV. Linear: 5 x 10-4-10-1 M -20 mV. Linear: 5 x 10-4-10-1 M Zn.2TPB -20 mV.Linear 5 x 10-4-10-1 M 27 (Glucam P20)0.4 Mg.2TPB Poor slope: 28 (Glucam P20)0.4 Zn.2TPB Poor slope: 29 Glucarn P10. Ba.2TPBT Poor slope: 30 (Antarox CO-880)0,4 Poor slope: Minimal Minimal Very poor, Minimal nonlinear and erratic Minimal Very poor Minimal Minimal Minimal Poor slope: -22 mV. 10-3-10-1 M Linear: * Response times were longer and readings were taken 15 min after addition of metal ion spike. t Zinc and magnesium complexes could not be prepared using the Glucam P10 ionophore. 0- -4.5 -4 -3.5 Log([Glucam PIOI/M) Fig. 3. Response of Series C, Type 26 electrode to Glucam P10 propoxylate. A, First run; 0, second run; and 0 third run -4.5 -4 -3.5 Log ([ G I ucam P~O]/M) Fig, 4. Response of Series C, Type 26 electrode to Glucam P20 propoxylate.A, First run; 0, second run; and 0 third run118 150 > 100 E G a 50 ANALYST, FEBRUARY 1987, VOL. 112 - - - - -4.5 -4 -3.5 Log( [CO-880]/~) Fig. 5. Res onse of Series C, Type 29 electrode to Antarox CO-880 ethoxylate. 1, First run; 0, second run; and 0, third run 150 ,100 E G Q Fig. 6. Relative response of Series C, Type 29 electrode to ethoxylate and propoxylate. A fresh electrode membrane was taken for each run. A, Antarox CO-880; 0, Glucam P20; and 0, Glucam P10 Table 4. Extraction of picrate into dichloromethane by Antarox CO-880 and Antarox CO-730 in the presence of barium ions Bulk extraction Picrate constant, [Ba Cl$M [NIS]/M extracted, Yo K M * Antarox CO-880 - Ba: 5 x 10-3 7.04 x 10-4 19.6 0.42 M-0-4 5 x 10-4 7.04 x 10-4 10.1 0.35 M-0.4 Antarox CO-730 - Ba: 5 x 10-3 7.04 x 10-4 30.6 0.48 ~ - 1 5 x 10-4 7.04 x 10-4 11.5 0.094 M-1 * Mean of three runs for each system at each concentration of barium chloride.approach the quality of response characteristic of the estab- lished system4 based on the tetraphenylborate of the barium complex with Antarox CO-880. The static response times were very long (up to 15 min). An electrode based on the TPB of the zinc complex with Antarox CO-880 gave a calibration for zinc in addition to barium, but of rather poor quality in each instance (Type 30 in Table 3). These attempts at producing an ion sensor for small cations are more disappointing than those for producing a magnesium ion sensor from the tetraphenylborate of the complex of magnesium with PPG 1025 propoxylate.13 However, PPG 1025 provided the bonus of a reasonable calcium ion-selective electrode.13 It seems that bivalent ions near to barium in size offer the best prospects for ion-sensors.Thus, a good lead(I1) ion-selective electrode, which is linear from 10-5 to 10-1 M lead nitrate with a response of -26 mV decade-1 and which is stable for more than one week, has been developed12 from the tetraphenylborate of lead( 11) complexed with Antarox CO-880. Polyalkoxylate sensing This is illustrated in Figs. 3-6, new electrodes being used for each study. Thus, Figs. 3 and 4 summarise the responses of Type 26 electrodes to Glucam P10 and Glucam P20 propoxy- lates, respectively. Fig. 5 summarises the responses of Type 29 electrodes to Antarox CO-880 and Fig. 6 compares the responses of the Glucam P10 electrode towards Antarox CO-880 ethoxylate and the Glucam P10 and P20 propoxylates.The response towards Glucam P10 is relatively small, as shown in Figs. 3 and 6, and the same characteristic applies to all the electrode types of Table 3. When the electrodes are exposed to Glucam P20 there is a sharper response (Figs. 4 and 6), and the response for Type 26 electrodes is highly reproducible. The responses towards Antarox CO-880 were very pro- nounced, especially for first runs (Fig. 5). There is a similarly high response towards Antarox CO-880 by the Type 26 electrodes. The response by the Glucam propoxylate electrodes towards Antarox CO-880 contrasts with the poor response of the Antarox ethoxylate based electrodes towards Glucam- containing solutions.Thus, no significant response was noted when Types 3, 5, 11, 15, 17, 19 and 24 electrodes (Table 2) were exposed to Glucam P20 between 5 X 10-5 and 5 x 10-3 M concentrations. These data point to an interaction process at the membrane - analyte interface that may be controlled by the relative stabilities of the metal - alkoxylate complexes concerned, and Table 5. Extraction of dipicrylamine into dichloromethane by polyalkoxylates in the presence of barium, magnesium and zinc ions. [M*+] = 5 x M. [Dipicrylamine] = 5 x 10-5 M. Equal volumes (10 cm3) of water and dichloromethane used M2+ Alkox ylate Ba . . . . . . . . Antarox CO-730 Mg Zn Ba . . . . . . . . PEG1500 Mg Zn Ba . . . . . . . . Antarox CO-880 Mg Zn Ba . . . . . . . . Antarox CO-430 Mg Zn Ba .. . . . . . . GlucamP10 Mg Zn Ba . . . . . . . . GlucamP20 Mg Zn [Alkoxylate]/ 10-5 M 0.704 0.600 0.636 7.58 1.12 2.07 K h 4 1.38 x 106 1.38 x 103 0.55 3.64 0.70 0.27 2.29 0.58 0.04 1.94 x 103 6.68 x 102 1.59 x 103 47.2 184 7.86 9.45 7.08 8.05 Units of K M M-1 M-0.3 M-0.4 M-1 M-1 M-0.4 Dipicrylamine extracted, YO 92.2 61.9 7.9 51.9 40.2 32.4 40.6 30.3 17.9 64.2 53.8 61.9 20.5 28.8 16.7 62.2 59.4 60.1 Dipicrylamine extracted per [alkoxylate]/l06% ~ - 1 13.1 8.8 1.1 8.7 6.7 5.4 6.4 4.8 2.8 0.85 0.71 0.82 1.8 2.6 1.5 3.0 2.9 2.9ANALYST, FEBRUARY 1987, VOL. 112 119 Table 6. Solvent extraction data for the extraction of thallium(1) nitrate into dichloromethane by various polyalkoxylates using dipicrylamine. [Dipicrylamine] = 5 X 10-5 M; [T1NO3] = 5 X M; [polyalkoxylate] = 5 x 10-6 M Polyalkoxylate Antarox CO-880 Glucam P20 PEG 1500 Glucam P10 Antarox CO-730 Antarox CO-430 YO Dipicrylamine extracted from aqueous phase per [ polyalkoxylate]/ lo6 .. . . . . 11.2 8.2 7.5 5.9 3.7 2.4 can be related to the solvent extraction trends discussed under Solvent Extraction Studies. Solvent Extraction Studies Solvent extraction data based on the partitioning of picrates of metal ion complexes with Antarox CO-880 and certain other members of the Antarox series have shown8 the affinity of bivalent metals for the polyethoxylates to be Be>Ba>Sr>Ca>Mg. These studies have been extended here, but significant extractions could only be obtained for picrate with Antarox CO-880 and Antarox CO-730 complexed with barium from among the materials studied in this investigation (Table 4).Owing to the differences in the units of the bulk extraction constants [arising from power terms such as yln in equation (9)] between the barium complexes of the two polyethoxylates, it is not possible to draw conclusions from the bulk extraction constant data except to comment that the complexing tendencies are similar for the two poly- ethoxylates. In order to obtain fuller data for each of the polyalkoxylates complexed with barium, magnesium and zinc ions, dipicryl- amine was substituted for picrate as a large polarisable anion, dipicrylamine having been previously shown by Jawaid and Ingman14 to be about one hundred times more powerful in its extractions of crown ethers than picrate. Table 5 summarises the principal solvent extraction data and the bulk extraction constants, KM, the units of KM being related to the stoi- cheiometry of the relevant metal ion - polyalkoxylate com- plex.Again it is difficult to compare the different groups of KM data because of differences in units. Also, much higher concentrations of alkoxylate have had to be employed for Antarox CO-430, Glucam P10 and Glucam P20 in order to obtain sufficient extraction of the dipicrylamine. Within groups, barium forms the strongest complex for each poly- alkoxylate, except for the Glucam P10 system, where the magnesium complex is the strongest. Zinc forms the weakest complexes of each group except for Antarox CO-430 and Glucam P20. A very useful comparison of solvent extraction data can be made by considering the “complex” dipicrylamine extracted per mole of alkoxylate (see the last column of Table 5).This measure of complexing tendency with the bivalent metals shows, in general, that Antarox 730>PEG 1500>Antarox CO 880>Glucam P20>Glucam PlO>Antarox CO 430. Correlation of Potentiometric and Solvent Extraction Data Antarox CO-880 forms the basis of a very good barium ion-selective electrode (Table 4). This is in accordance with the KM value for the barium complex and with the good extraction of the complex into the organic phase as depicted by dipicrylamine (Table 5). Although this electrode system gives e.m.f. changes for ethoxylates in aqueous solutions, it is inferior to the Antarox CO-430 electrode system for this purpose. The better performance of the barium - Antarox CO-430 system as an electrochemical sensor for poly- ethoxylates may be at least partly due to the relatively poor affinity exhibited by the organic phase for this complex, as depicted by the last column of Table 5.Also, just four ethoxylate units are complexed to the barium cation, com- pared with the larger number of ethoxylate units complexed for other members of the Antarox series, which promote the transfer of the polyethoxylate being sensed towards the barium ions in the membrane. Within each polyalkoxylate - metal ion group, the barium systems provide the best metal ion-selective electrodes (Tables 2 and 3), which is in general accordance with the solvent extraction data (Table 5). In relation to the potentiometric response being associated with complex stability, it is interesting to note the facile conversion15 of an Antarox CO-880 based barium ion-selec- tive electrode to a thallium(1) electrode by soaking the barium electrode in thallium(1) nitrate for 24 h.Such a conversion is consistent with the dipicrylamine solvent extraction data of this work (Table 6). From the standpoint of the potentiometric response to polyethoxylates, it is significant that the Glucam-based electrodes readily respond to added Antarox CO-880 poly- ethoxylate, but that the Antarox-based electrodes do not respond to the polypropoxylated Glucam materials. However, in relation to the differences in complexing tendency and extraction into the hydrophobic phase there is also the matter of the less hydrophobic nature of the more oxygenated methyl glucoside tail of the Glucam materials compared with the hydrocarbon nonylphenoxy tail of the Antarox materials.Conclusion Allowing for the optimum proportions of potentiometric sensing materials in PVC membranes, the above studies indicate a definite relationship between the potentiometric behaviour of an alkoxylate ionophore and the stability and solvent extraction of the ionophore - cation complex. The authors thank the Science and Engineering Research Council and Unilever Research, Port Sunlight Laboratory for a studentship (to P. H. V. A.) under the Cooperative Awards in Engineering Scheme. Professor A. A. S. C. Machado and Dr. J. L. F. C. Lima of the University of Oporto are also thanked for information relating to the Antarox CO-880 based thallium electrode made possible by NATO Grant No. 0069. References 1. Jones, D. L., Moody, G. J., and Thomas, J. D. R., Analyst, 1981, 106,439. 2. Jones,D. L.,Moody,G. J.,Thomas, J.D. R., andBirch,B. J., Analyst, 1981, 106, 974. 3. Moody, G. J., and Thomas, J. D. R. in Cross, J., Editor, “Non-Ionic Surfactants: Chemical Analysis,” Marcel Dekker, New York, 1987, p. 117. 4. Jaber, A. M. Y., Moody, G. J., and Thomas, J. D. R., Analyst, 1976, 101, 179. 5. Craggs, A., Moody, G. J., and Thomas, J. D. R., J. Chem. Educ., 1974, 51, 541. 6. Hollis, G., Editor, “Surfactants UK: Directory of Surface Active Agents Available in the United Kingdom,” Second Edition, Tergo Data, Darlington, 1979.ANALYST, FEBRUARY 1987, VOL. 112 7. Delduca, P. G . , Jaber, A . M. Y., Moody, G. J., andmomas, J. D. R., J. Znorg. Nucl. Chem., 1978, 40, 187. 8. Jaber, A. M. Y . , Moody, G. J., andnomas, J. D. R., J. Inorg. Nucl. Chem., 1977, 39, 1689. 9. Pedersen, C. J . , Fed. Proc. Fed. Am. SOC. Exp. Biol., 1968,27, 1305. 10. Eisenman, G., Ciani, S. M., and Szabo, G., Fed. Proc. Fed. Am. SOC. Exp. Biol., 1968, 27, 1289. 11. Eisenman, G., Ciani, S., and Szabo, G., J. Membrane Biol., 1969, 1,294. 12. Jones, C. G. D., Moody, G. J., and Thomas, J. D. R., unpublished work. 13. Jaber, A. M. Y., Moody, G. J., and Thomas, J. D. R., Analyst, 1977, 102,943. 14. Jawaid, M., and Ingman, F., Tulantu, 1978,91,25. 15. Lima, J. L. F. C., and Machado, A. A. S. C., to be published. Paper A61258 Received July 31st, 1986 Accepted September 16th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200113

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 121 Liquid Membrane Ion-selective Electrodes for Diquat and Paraquat G. J. Moody, R. K. Owusu and J. D. R. Thomas* Department of Applied Chemistry, Redwood Building, UWIST, P.O. Box 13, Cardiff CFI 3XF, UK Three main types of PVC - liquid membrane ISEs for diquat and paraquat are described. They are based on the crown ether DB30C10 and/or the diquat - paraquat ion pair with tetraphenylborate (DQT.2TPB or PQT.2TPB). With DB30C10-based ISEs (type l), selectivity for diquat and paraquat over other cations is greatly improved after the adddition of DQT.2TPB to the membrane to yield type 2 electrodes. Electrodes based entirely on DQT.2TPB or PQT.2TPB (type 3) also give excellent characteristics with appropriate solvent mediators. Selectivity coefficient data (eiT,B and $&B) have been determined with respect to B = Li, Na, K, Mg, Ca, Ba, ammonium, di-methyl ammonium, anifinium and guanidinium. Although the selectivity coefficients frequently exceeded unity for interferents of singly charged B ions, the modifying influence of the square power term in the selectivity relation enhances the selectivity towards diquat and paraquat. For doubly charged B ions the selectivity coefficients for types 2 and 3 electrodes were normally <1 for diquat.Type 2 electrodes with o-nitrophenyl phenyl ether as the plasticising solvent mediator in the PVC matrix membrane with DB30C10 and DQT.2TPB are recommended as diquat and paraquat ion-selective electrodes, whereas type 3 electrodes with either o-nitrophenyl phenyl ether or o-nitrophenyl octyl ether as the solvent mediator are suitable as paraquat ion-selective electrodes.Keywords: Crown ether ion sensors; ion-selective electrodes; neutral carrier electrodes; paraquat determination; diquat determination Ion-selective electrodes (ISEs) based on neutral carrier ligands are well established for alkali metal and alkaline earth metal cations and for thallium(I), cadmium(II), uranyl and chiral ammonium ions. 1-3 Essentially, the neutral carriers have the characteristics of being uncharged, lipophilic and of undergoing reversible complexation with selected cations and hence of promoting cation transfer between the aqueous sample phase and the organic membrane phase by carrier translocation. The early examples of neutral carrier based ISEs involved the use of valinomycin and other naturally occurring antibiotic compounds, e.g., macrotetrolides, for potassium and ammo- nium ISEs.495 Synthetic neutral carriers, such as the cyclic polyethers (crown ethers) ,6 polyheterocyclic systems (cryptands)7 and acyclic polyether diamidesg have since been produced. The last group has given rise to highly specific sensors, whereas the cryptands have not, although the latter form the more stable complexes.Of the synthetic crown ethers, dodecylmethyl-14-crown-4 has yielded an interesting lithium ISE,9 and bis(crown ether) derivatives of 15-crown-5 and 18-crown-6 have yielded potas- sium and caesium ISEslO and a macrocyclic polyether-amide with two polyether rings has shown enhanced calcium ion-selectivity when used in calcium ISEs.llJ2 The 2 : 1 crown ring : cation complexes formed between the two polyether ring systems and cations are considered to have a significant effect in promoting electrochemical ion-selectivity.11 The selection of suitable neutral carriers for ion sensing can be helped by structural studies on interactions between the carriers and ions. In this respect, Stoddart and co-workerslS28 have shown interesting features concerning the interactions between macrocylic polyethers and molecules and cations. They classified the bonding under three main headings: (i) coordination via [N-H - - - 01 hydrogen bonds as in the complexation between polyether and primary alkylammonium ions13J4 and neutral species such as borane - ammonial5-17 and where the polyether serves as the second sphere ligand by interaction with -NH3 ligands in the first coordination sphere of cationic and neutral transition metal complexesls22; (ii) coordination through [C-H .- - 01 linkages to the diquat dication ,23 alkyl phosphonium cations24 * To whom correspondence should be addressed. and alkyl sulphonium cations25; and (iii) metal cation coordi- nation to oxygen atoms of the crown ether compound and to the anions of salts such as NaPF6,26 Li picrate22 and Ba(SCN)2.28 These studies, 13-28 supported by various crystallographic structural data,13-28 and the involvement of [N-H . 01 interactions in the complexation of primary alkylammonium ions to 18-crown-614.29 and its many derivatives,l3~14~29 suggest that a much wider range of materials can be detected by potentiometric electrodes than is evident from previous studies, which have been essentially concerned with ionophore-type potentiometric metal cation sensors.373O33l Among the few studies on the sensing of larger organic ions and molecules is the application of chiral macrocyclic poly- ethers for sensing chiral primary alkylammonium ions, e . g . , (R)- or (S)-PhCHMeNH3+.3J2 However, there is a need for a wider range of ion sensors for larger organic ions. Hence, there is considerable incentive for using the structural information obtained by Stoddart and co-workers13-28 and in this first approach ion sensors for diquat and paraquat dications are described. The determination of these compounds is important in view of their use as contact herbicides.They are also toxic to animal life with respective values of LD50 (i. e . , 50% lethal dose) in the rat being 231 and 100 mg kg-1 of body mass.33 The use of dibenzo-30-crown-10 (DB30C10) as a neutral carrier sensor results from discussions with Stoddart and Williams. These were based on an X-ray crystallographic structure determination of the complex between the diquat dication and DB30C10 where the gross host - guest structural features are of the category (ii) type of hydrogen bonding and charge transfer and give stable and ordered 1 : 1 solution complexes between DB30C10 and diquat bis(hexafluorophosphate) .23 Experimental Reagents Dibenzo-30-crown-10 (DB30C10) was synthesised by Stoddart and co-workers at the University of Sheffield from the 2-benzyloxyphenol of 1,2-dihydroxybenzene and polyethylene glycol ditoluene-4-sulphonate.20 Diquat dibromide (6,7- dihydrodipyrido[ 1,2-a : 2‘, 1 ‘-clpyrazinediium dibromide) and paraquat sulphate (1’ , l’-dimethyl-4,4’-bipyridinium sulphate)122 ANALYST, FEBRUARY 1987, VOL.112 were kindly presented by Dr. J. F. Stoddart, University of Sheffield. A list of other reagents and suppliers is as follows: o-nitrophenyl phenyl ether (NPPE) (Eastman Kodak, Rochester, NJ, USA); o-nitrophenyl octyl ether (NPOE) (Fluka, Switzerland); dioctylphenylphosphonate (DOPP) (Lancaster Synthesis, Lancashire, UK); dibutyl phthalate (DBP) and dinonyl phthalate (DNP) (BDH Chemicals, Poole, Dorset, UK); and poly(viny1 chloride) (PVC) Breon SllO/OP (BP Chemicals International, Barry, UK).Best laboratory-reagent grade chemicals were used, including chlorides of lithium, sodium, potassium, ammonium, magne- sium, calcium and barium (BDH Chemicals) and guanidinium chloride and sodium tetraphenylborate (Sigma Chemicals, Poole, Dorset, UK). Preparation of Diquat Tetraphenylborate and Paraquat Tetra- pheny lborate A two-fold mole excess of sodium tetraphenylborate (NaTPB) dissolved in a minimum volume of water was added to aqueous diquat dibromide and paraquat sulphate (10-1 M, 100 m 3 ) as appropriate. The resulting brown precipitates were filtered, washed with water (3 x 20 cm3) and dried overnight in a vacuum oven (39 "C, and 150 mmHg setting). PVC Membrane Fabrication and Use The method for preparing PVC immobilised neutral carrier membranes and their assembly into the electrodes is as described previously.34J5 Table 1 summarises the three main types of membrane compositions for diquat and paraquat but, in addition, some PVC matrix membranes were made with just solvent media- tors.Electrodes were cpnnected to measuring instruments by shielded copper coaxial cables. They were normally con- ditioned before use by immersion for 24 h in 0.1 M diquat and paraquat salts as appropriate. Table 1. Composition of PVC matrix membrane types used for electrode construction Membrane composition,% mass ratio Solvent DQT.2TPB or Membrane type mediator PVC DB30C10 PQT.2TPB 1 67.0 31.6 1.4 - 2 62.3 29.4 1.3 7.0 7.0 3 63.0 30.0 - Electrochemical measurements were made with a Radio- meter PHM 64 pH - millivolt meter (Radiometer A / S , Copenhagen, Denmark) in association with a potentiometric recorder (Rikadenki Kogyo Co., UK).The reference electrode used was an EIL Type RJ23 with a remote microjunction. Diquat dibromide or paraquat sulphate, as appropriate (10-3 M in 10-1 M HCl), was the internal filling solution of the ISEs. Electrode calibrations were carried out by spiking with successive aliquots of known concentrations of sample into doubly de-ionised water thermostated at 25 k 0.1 "C using a Paratherm heater - stirrer (Seelbach Uber Lahr, Schwarzwald, FRG). Some calibrations were also carried out in reverse, i. e., by diluting standards, and the response behaviour wasgenerally similar to the spiking procedure. All samples were prepared in doubly de-ionised water. Selectivity Coefficient Determination Potentiometric selectivity coefficient (kgk.B) values were determined using the separate solution met od @GT,B = 10[(AEB - ')''I /(a)~"y .. . . (1) where AEB is the electrode response to a separate solution activityofinterferention (aB),z = 2,yis thechangeonionB and S and Care linear regression parameters for electrode response to the primary ion (AE = S log a + C). The numerator in equation (1) corresponds to the activity of DQT to which the electrode gives the same response as AEB. Membrane Resistances The resistances of freshly prepared (dry) membranes were measured with a dry metal contact in conjunction with a 0-39.9 V d.c. source, a picoammeter (Model 485, Keithley Instrument, Cleveland, OH, USA) and a digital multimeter (Model 191, Keithley Instrument).The resistances of the membranes (wet) after soaking in 0.1 M analyte for 48 h were determined using a potentiostat (Briicker, Model E130M). Results and Discussion Diquat Electrodes The response characteristics of the three electrode types (1-3) are summarised in Table 2. The electrode response to diquat is described by the linear regression line AE(mV) = Sloga + C . . . . . . (2) where S, a and C, respectively, are the electrode calibration slope (mV/log a), the activity of the sensed ion (mol dm-3, represented by M) and the response intercept (mv). The Table 2. Response characteristics of ion-selective electrodes for diquat. Standard deviations in parentheses are for n = 5-7, spread over 7 d. mediator* SImV CImV amin,/10-6M r S/mV CImV amin./10-6M r SImV CImV amin.I10-6M r DBP .. . . . . 35.6 179 9.7 0.991 25.2 135 4.2 0.998 34.1 186 3.5 0.998 DNP . . . . . . t t t 1- 17.5 8.1 24 0.89 27.3 114 430 0.987 DOPP . . . . 47.0$ 233 11 0.998 21.8 105 16 0.997 27.0 149 3.3 0.997 0.998 NPOE . . . . 32.6 151 31 0.989 28.8 170 1.5 0.999 31.2 160 18 1.8 0.999 NPPE . . . . t t t t 29.8 169 1.4 0.999 26.7 153 (3.3) (15.7) (2.2) (0.35) (3.1) (0.4) (0.6) (5.3) (0.8) (1.1) (1.6) (11) (1.2) (10) (360) (0.2) (2.0) (2) (2.2) (3.4) (2) (20) (2.1) (18.7) (24) (0.6) (6.4) (1.1) (1.9) (27) (1.4) (11) (0.4) (0.9) (5.3) (0.4) * DBP = dibutyl phthalate, DNP = dinonyl phthalate, DOPP = dioctylphenyl phosphonate, NPOE = o-nitrophenyl octyl ether and NPPE = t Functional electrodes not obtained, see text. 3 Electrode slope unstable, see text.o-nitrophenyl phenyl ether.ANALYST, FEBRUARY 1987, VOL. 112 minimum activity detectable, amin., is estimated from amin, = 10-C’S. Activity coefficients, f, were calculated from the modified Debye - Hiickel relation used previously35 logf= 0.511 z21$/(1 + 14) . . . . . . (3) where z is the valence and Z is the ionic strength. Most electrodes examined gave a 95% steady-state dynamic response in about 30 s and a full response in less than 2 min. Electrodes transferred from test solutions into water quickly attained the “background” potential. The notable exception was DOPP-based electrodes with a steady-state response of about 5 min at low sample ion activities (a d 5 X The pH response profiles for electrodes were examined using 10-2 M aqueous solutions of diquat bromide (40 cm3).The pH was adjusted by introducing small drops of hydro- chloric acid (10 M) or sodium hydroxide (4 M). All ISEs studied had a useful pH range of about 1.2-8.5. Generally between pH 9.0 and 9.5 (and at higher pHs) there was a sudden change in the colour of diquat solutions from light yellow to red. There then followed a monotonic increase in the electrode response with time, but normal response was restored by immersion in solutions of pH 4.5-6.0. Type 1 electrodes tended to be more unstable. Notably, type 1 electrodes with NPOE as the plasticising solvent mediator had a lifetime of only 96 h. The electrode intercept decreased at a rate of 0.9 mV h-1 over this period. The corresponding type 3 electrode with NPOE exhibited an intercept change of -2.5 mV h-1 and also had a lifetime of about 96 h.Type 2 electrodes with NPOE also showed some drift. An important membrane parameter is probably the plasti- cising solvent mediator viscosity.36 All types of electrodes based on DBP, DNP and DOPP had lifetimes exceeding 2 weeks. Electrodes based on NPPE were used continuously for over a month. Type 1 electrodes with DNP or NPPE plasticising solvent mediator proved too noisy for a full evaluation. Such characteristics may be attributed to a high membrane resis- tance (Table 3) and may be counteracted by the addition of background (lipophilic) electrolytes to the membrane phase.37-40 Accordingly type 2 electrodes, in which DQT.2TPB and DB30ClO were present in the membrane (Table l), gave rise to fully functional electrodes with DNP or NPPE as the solvent (Table 2).Notably, type 1 electrodes with DBP as the plasticising solvent mediator gave a fully func- tional electrode. Likewise, the substitution of the octyl substituent for a phenyl group in the NPPE solvent to give NPOE appears to have a significant effect on the membrane resistance (Table 3). Type 1 electrodes with DOPP as the plasticising solvent mediator gave unusual responses. For instance, the calibra- tion slope for diquat dibromide samples was 33.2, 47.0 and 50.3 mV (log a)-1 for 2, 48 and 96 h-old electrodes, M). 123 respectively. Responses to paraquat sulphate were also variable with S = 56.5,67.1 and 69.1 (+ 2.0) mV (log a)-1 for 24, 48 and 120 h-old electrodes. Super-Nernstian values for S were also frequently observed for other new DOPP-based electrodes in response to para- quat, e.g., type 2 [ A E (mV) = 58.0 log a + 338; r = 1.0001 and type 3 [ A E (mV) = 51.2 log a + 298; r = 0.9961 electrodes.In types 2 and 3 electrodes, S for calibrations with diquat was S30 mV (log u)-1 and did not increase with the age of the electrode. Super-Nernstian slopes observed for neutral carrier based ISEs for the divalent U022+ and Cd2+ cations have been attributed to U020H+ and Cd Cl+.41>42 Such monovalent complexes may also be the predominant membrane permeat- ing species in classical liquid ion-exchanger electrodes for divalent metal cati0ns.h this instance a Nernstian slope [ S = 29.6 mV (log a)-’] is obtained only where the anion transferance number z is 0.5 z = u(s)/[u(is) + u(s)] .. . . . . (4) where s = the ion-exchange sites, e.g., TPB-, is = the monovalent ion-exchange site - cation complex and u = the mobility in the membrane phase.43 A super-Nernstian slope might, therefore, arise in situations where u(s) >> u(is), i.e., t = 1.0. The above possibilities may underly the non-ideal response characteristics noted for the DOPP-based elec- trodes. The time dependence of S in calibrations with diquat and paraquat samples for type 1 electrodes may be the result of slow changes in the environment of charged species within the membrane phase, e.g., as a result of slow saturation with water. The presence of DQT.2TPB in the membrane preven- ted the time dependent increases in the calibration slope for the diquat-based ISEs.The selectivity characteristics of type 1 and 2 electrodes (Table 4) will, as mentioned earlier, partly depend on the complexation characteristics of DB30C10. Crown ethers DB3nCn (n = 6-12) form complexes with DQP+ and the most stable complexes arise23 for n = 10 (log K, = 4.24 in acetone). X-ray crystallographic studies revealed a ligand - cation complex ion formation in which the major plane of DB30C10 was curved into a C-shaped “pseudo-cavity”; within this sits DQF+ with a plane at 90” to that of the ligand.23 The greatest dimension of the pseudo-cavity was 6.8 8,, with an average contact distance between DB30C10 donor (oxygen) atoms and diquat (two quaternary ammonium centres) of 3.3 A.23 The conformation of the DB30C10 D Q F f complex and of other DB30C10 or DB21C7 complexes, e.g., with Cs+, do not correspond to the “cation-within-ligand cavity” conformation observed for small ring size ligands (e.g., n = 4-6).2+ Nevertheless, the relationship between a cavity - cation size ratio of unity and optimum complex stability is expected to apply to large ring size ligands.The ligand pseudo-cavity Table 3. Membrane resistances of PVC membranes for diquat and paraquat for fresh dry membranes and for membranes (wet) after soaking in 0.1 M analyte for 48 h. Membrane surface area = 0.28 cm3 and thickness - 0.02 cm; s.d. in parentheses Membrane resistancelMS1 Solvent Dry Wet DBP . . . . 29(10) 22 (1.2) DNP . . . . 2200(700) 200 (11) NPPE . . . . 9.0(5) 36 (1) DOPP . . . . lO(1.5) 0.6 (0.2) NPPE * . . NPOE .. . . 2.0(0.6) 12 (0.05) NPOE* . . - - - - Type 2 - Dry Wet 6.3 (1.5) 3.0 (0.2) 0.82 (0.2) 13 (2.7) 0.20 (0.02) 0.054 (0.004) - 0.38 (0.01) 0.24 (0.05) 0.054 (0.01) 0.55 (0.2) 0.46 (0.02) 100 (4.2) 46 ( 5 ) Type 3 No sensor Dry Wet Dry 74 (10) 43 (2.3) - 0.25 (0.01) 0.1 (0.02) - 0.20 (0.05) 0.26 (0.03) - - 0.70 (0.1) - 1.4 (0.5) - - 13.1 (1) 3.1 (0.3) 130 (50) 0.40 (0.01) 0.26 (0.05) 31 (4) * PQT-based sensors; remainder are for DQT-based sensors.124 ANALYST, FEBRUARY 1987, VOL. 112 Table 4. Selectivity coefficients for diquat electrodes determined by the separate solution method Log for the various electrode types based on different solvent mediators Type 1 Type 2 Type 3 Bspecies,[B]=9.lm~ DBP NPOE DBP DNP DOPP NPOE NPPE DBP DNP DOPP NPOE NPPE Li+ . . . .. . . . 2.3 0.5 -0.85 0.34 5.5 Na+ . . . . . . . . 4.1 4.0 -0.56 1.4 3.0 K+ . . . . . . . . . . 7.2 5.3 0.6 6.8 1.4 NH4+ . . . . 3.5 3.5 0.33 4.3 3.5 Guanidi&m(Gu+) . . 3.2 3.2 2.2 5.7 7.3 C6HSNH3+ , . . . . . 4.5 4.0 3.4 7.6 9.3 (GH5)2NHz+ . . . . . . 2.8 1.4 1.8 5.2 4.7 PQF+ . . . . . . . . 0.04 -0.5 -0.05 1.6 6.8 MgZ+ . . . . . . . . -1.9 -1.7 -3.0 -2.2 1.6 Ca2+ . . . . . . . . -1.1 -1.2 -2.9 -1.7 3.3 Ba2+ . . . . . . . . 0.2 0.5 -1.3 -0.95 1.2 -1.7 -1.7 -2.0 -1.6 -2.0 -1.3 -0.80 -1.1 -3.6 -3.9 -1.3 -1.7 1.1 -0.1 2.7 -1.0 -1.7 1.7 1.2 1.7 -1.2 -1.1 2.7 2.9 0.9 -1.5 -1.7 0.7 1.3 2.4 -1.7 -1.7 2.2 4.2 5.0 -1.5 -1.5 2.9 4.3 4.1 -1.5 -1.0 2.0 4.0 3.0 -0.4 0.04 0 0 2.1 -2.0 3.7 -3.4 -2.2 0.8 -3.3 -3.7 -2.7 -2.4 2.2 -3.3 -3.7 -1.5 -2.0 0.4 -1.5 -1.4 -1.4 -0.84 -1.7 -1.3 -1.3 -0.85 -0.04 -3.5 -3.5 -1.8 diameter (d‘) and true cavity diameter (d) are approximately related by d = Mzd’.DB30C10 is known to form strong complexes with potas- sium, rubidium, caesium and barium (e.g., log K, = 4.2 for the K+ complex in MeOH). Their respective diameters, 2.66, 2.98, 3.30 and 2.68 approach the optimum size for accommodation within a spherical region (-- 3.3 A) from the ligand donor atoms. (Incidentally, the true cavit diameter of DB30C10 is expected to be of the order of 10.4 1). DB30C10 has been used as a basis for a neutral carrier based ISE for potassium.47 Strong interference by Group 1 and Group 2 cations is, therefore, to be expected in PVC DB30C10-based ISEs for diquat and paraquat (Table 4). For type 1 electrodes there is interference from all the monovalent cations assessed.The very strong interferences by K+ and Ba*+ ions are notable. Electrodes with PVC Membranes Containing Plasticising Solvent Mediators For solvent mediators alone, a PVC matrix containing only DBP responded to diquat dibromide [e.g., E(mV) = 18.7 log a + 92.5; r = 0.9901. A PVC - NPOE membrane, on the other hand, gave a near-Nernstian response to paraquat sulphate [e.g., E(mV) = 34.0 log a + 195.7; r = 0.9981, but the response to diquat was marginal. In the latter instance, kgOdT,B values (Figs. 1 and 2) were estimated48 from log @&,B = log %T,B - log %T,DQT The role of DBP and NPOE membrane components on the selectivity of the various PVC.DB30C10-based electrodes may be seen in Figs. 1 and 2, respectively, whereas the main results for DBP and NPOE solvents alone in the electrode membranes are as mentioned above.Thus, in the absence of sensor (i.e., DB30C10 or DQT.2TPB) selectivity is a function of cation partitioning at, for example, the PVC.DBP or PVC.NPOE membrane aqueous solution interfaces.49-51 For the DBP or NPOE alone in membranes the greatest interfer- ences arise from the more lipophilic cations, e . g . , CbHS.NH3 and (&H&NH2. The strongly hydrophilic nature of divalent metal cations is also evident from the low + + values. Effect of Adding Diquat Tetraphenylborate to Membranes The addition of DQT.2TPB to membranes resulted in a marked reduction in the degree of cation extraction, for example, relative to PVCor PVC.DB30C10 systems (Figs. 1 4 and Tables 2 and 4).For the PVC.DB30ClODQT.2TPB type 2 membrane with NPPE or NPOE solvent, the resulting electrodes were virtually insensitive to all metal cations assessed (a <lo-2 M) (Fig. 3 and Tables 2 and 4). There is also a significant reduction in the interference from organic DOT POT Mg Ca Ba Li Na K NH4 GU PhNH3Et2 B H2 Fig. 1. Selectivit characteristics of ISEs for diquat showing effect of PVC matrix memgrane composition inco rating dibutyl phthalate gasticising solvent mediator. 0, Type 1; T T y p e 2; 0, Type 3; and , PVC with just DBP. Lines joining points connect the various electrode types 4 -1 J - 2 i:i -4 I I I I I I ’ I ’ I I I I ] DOT POT Mg Ca Ba Li Na K NH4 GuPhNH3 Et,NH, B Fig. 2. Selectivity characteristics of ISEs for diquat showing effect of PVC matrix membrane composition incorporating o-nitrophenyl octyl ether plasticising solvent mediator. 0, Type 1; A , Type 2; 0, Type 3; 0, PVC with just NPOE; and A, Type 2 but with PVC DB30C10/ PQT.2TPB.Lines joining points connect the various electrode types cations. Similar trends apply also to type 3 electrodes (Fig. 4 and Tables 2 and 4) but the type 2 electrode is the more satisfactory in use. Such effects seem consistent with a generalised decrease in the ability of the membrane to solvate interfering cations. It might be that the introduction of a highly lipophilic neutral salt (DQT.2TPB) results in a reduc- tion of the membrane phase associated water. In support of such an interpretation it is important that innately more hydrophobic membrane systems are more resistant than polar membranes to permeation by lipophilic sample anions.1ANALYST, FEBRUARY 1987, VOL.112 125 8 - 6 - F 4 - BX 4 0-- 2 - rn - 2 - - 4 - DQT POT Mg Ca Ba Li Na K NH4 Gu PhNH3EtZNH2 B Fig. 3. Effect of plasticising solvent mediator on selectivity oft diquat ISEs. 0, NPOE; 0, NPPE; A, DOPP A, DNP; and., &$%? Lines joining points connect the various solvent mediator types 3 - 2 - m -< 1 :g 8 - 0 - - -I -1 - 2 - -3 - - -41 I ' ' I ' I I I I J B Fig. 4. Effect of plasticising solvent mediator on selectivity oft diquat ISEs. 0, NPOE; 0, NPPE; A, DOPP; A, DNP; and I, I!%'? Lines joining points connect the various solvent mediator types DQT PQT Mg Ca Ba Li Na K NH4 GuPhNH3 Et2NH2 Paraquat Electrodes The responses to type 2 and 3 membrane electrodes based on paraquat (PQT) are summarised in Table 5; Fig.5 compares the calibrations of the electrodes based on an NPPE plasticis- ing solvent mediator with the corresponding diquat elec- trodes. Type 1 PQT electrodes were of poor quality. The responses towards paraquat sulphate in the presence of interferences are frequently different from the corresponding DQT types for diquat dibromide (Table 6 and Fig. 6). As can be seen from Table 6 and Fig. 6, l@?dT frequently exceeds unity, whereas ~ t ~ , ~ with the DQ'r electrodes (Table 4 and Figs. 3 and $%oes not do so for types 2 and 3 electrodes with NPOE and NPPE as the solvent mediators. This greater interference of PQT electrodes can arise from the elongated shape and longer PQF+ ion compared with DQF+ and associated differences in electron distribution resulting from the wider separation of the charged nitrogen sites.Me I 8 N+ I h e Diquat Paraquat Table 5. Response characteristics of electrodes for paraquat based on PQT.2TPB. Standard deviations in parentheses are for n = 4 spread over 48 h Electrode type Parameter and solvent mediator SJmV ClmV ami,./10-6~ r Type 2: NPOE . . . . 19.0 82 NPPE . . . . 27.1 140 (0.8) (2.2) (5.0) (30) (1.1) (18) (4.0) (38) Type 3: NPOE . . . . 30.1 181 NPPE . . . . 27.2 155 3.4 0.999 0.981 9 (4) (6) 1.2 0.991 0.989 7 (0.8) (12) 1 > E iii Q -Log([al/M) Fig. 5. Responses of types 2 and 3 diquat and paraquat electrodes towards diquat dibromide and para uat sulphate, respectively. The electrode membranes are based on JPPE plasticising solvent media- tors.0, Type 2 (DQT); 0, type 3 (DQT); e, type 2 (PQT); and ., type 3 (PQT) lo F n PQT DOT Mg Ca Ba Li b Na K NHo Gu PhNH3 Et, Fig. 6. t 8 P E . For type 3 electrodes: a, NPOE; and 0, NPPE Selectivity characteristics of ISEs for paraquat according to e and solvent mediator. For type 2 electrodes: 0, NPOE; and 0, Related to these effects are the fact that type 3 PQT electrodes are of better quality than those of type 2 (Table 6 and Fig. 6). With regard to univalent ion interferences, the electrodes based on PQT and on DQT are still selective towards PQF+ and DQF+ even when the respective selectivity coefficients are 3 1 . This is because of the modlfying effect of the squared power terms in the relationship (for DQT)126 ANALYST, FEBRUARY 1987, VOL.112 Table 6. Selectivity coefficients by the separate solution method for paraquat electrodes with different solvent mediators based on PQT.2TPB log kg“dT,B Type 2 Type 3 B species, [ B ] = 9 . l m ~ NPOE NPPE NPOE NPPE Li+ . . . . . . 0.7 -1.0 -2.0 -1.2 Na+ . . . . 3.0 -1.0 -2.0 -1.7 K+ . . . . . . 10.1 5.0 -1.0 -0.4 NH4+ . . . . 6.0 3.6 -1.7 -2.0 Gu+ . . . . 4.1 2.5 0 -0.8 Ph.NH3+ 4 . 7.1 - 1.3 0.3 (Et)2NH2+ . . 2.6 0.7 2.0 1.2 DQP+ . . . . 4.2 3.0 -0.6 1.1 Mg2+ . . , . -1.7 -1.1 -4.0 -3.4 Ca2+ . . . . -0.6 -1.5 -4.0 -3.4 Ba2+ . . . . 1 .o 0.5 -2.4 -0.5 Conclusions The mole ratio of DQT.2TPB (or PQT.2TPB) to DB30C10 used in these studies was 3.1 : 1 (membranes contained 1.4 x mol of DB30C10 and/or 4.3 x 10-5 mol of DQT.2TPB or PQT.2TPB). A neutral carrier ligand function is undoubtedly retained in type 2 electrodes. This is apparent from the consequent alterations in selectivity profiles in Figs.1, 2, 3 and 6. The relation of type 1 electrodes to type 3 electrodes is analogous to that of a potassium tetraphenylborate (KTPB) - valinomycin based ISE4,37-40,52953 to a KTPB or potassium tetra-p-chlorophenylborate (KTPClPB) based electrode .38754 In the present example, the response characteristics of type 3 electrodes rival, and for PQT are better, than those of the neutral carrier based systems (type 2) (Tables 2,4 and 6, Figs. 1-6). The unusually low urnin. values for type 3 electrodes are probably the result of (a) a large association constant between TPB- and DQ’P+ (or PQT;?+) and (b) a high lipophilic character of the resulting neutral salt.In conclusion, the findings of this work for diquat are in general agreement with those of previous studies on modified neutral carrier membranes.37,39,40,52153 Type 2 DQT elec- trodes, containing both liquid ion-exchanger and neutral carrier, have superior characteristics compared to ion-selec- tive electrodes based entirely on neutral camer membranes. These type 2 electrodes have especially favourable charac- teristics as diquat ISEs when the membrane plasticising solvent mediator is o-nitrophenyl phenyl ether followed by o-nitrophenyl octyl ether. With regard to paraquat, the electrodes are frequently poor in quality, especially with regard to selectivity, but among these the type 3 electrodes, albeit without neutral carrier, are the best.The authors thank the Science and Engineering Research Council for financial support and for a post doctoral research associateship to R. K. 0. Dr. J. F. Stoddart of the University of Sheffield is also thanked for much practical help and advice relating to the supply of DB30C10 and other materials. 1. 2. 3. 4. 5. References Morf, W. E., and Simon, W., in Freiser, H., Editor, “Ion Selective Electrodes in Analytical Chemistry 1 ,” Plenum Press, New York and London, 1978, p. 211. Morf, W. E., “The Principles of Ion-selective Electrodes and of Membrane Transport ,” Elsevier, Amsterdam and Oxford, 1981, p. 274. Amman, D., Morf, W. E., Anker, P., Meier, P. C., Pretsch, E., and Simon, W., Ion-Sel. Electrode Rev., 1983, 5,3. Stefenac, Z., and Simon, W., Microchem. J., 1967, 12, 125.Stefenac, Z., and Simon, W., Chimia, 1966, 20,436. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. Pedersen, C. J., J. Am. Chem. Soc., 1967,89,7017. Dietrich, B., Lehn, J. M., and Squage, J. P., Tetrahedron Lett., 1969,21,2885. Amman, D., Pretsch, E., and Simon, W., Tetrahedron Lett., 1972,24,2473. Kitazawa, S., Kimura, K., Yano, H., and Shono, T., Analyst, 1985, 110,295. Kimura, K., Ishikawa, A., Tamura, H., and Shono, T., J. Chem. SOC., Perkin Trans., 2 , 1984, 447. amura, K., Kumami, K., Kitazawa, S., and Shono, T., J. Chem. SOC., Chem. Commun., 1984,442. Kimura, K., Kumami, K., Kitazawa, S., and Shono, T., Anal. Chem., 1984,56,2639.Stoddard, J. F., Chem. SOC. Rev., 1979,8, 85. Maud, J. M., Stoddard, J. F., and Williams, D. J., Acta Cryst., 1985, C41, 137. Colquhoun, H. M., Jones, G., Maud, J. M., Stoddard, J. F., and Williams, D. J., J. Chem. SOC., Dalton Trans., 1984, 63. Allwood, B. L., Shahriari-Zavareh, H., Stoddard, J. F., and Williams, D. J., J. Chem. SOC., Chem. Commun., 1984, 1461. Alston, D. R., Stoddard, J. F., Wolstenholme, J. B., Allwood, B. L., and Williams, D. J., Tetrahedron, 1985,41,2923. Colquhoun, H. M., Lewis, D. F., Stoddard, J. F., and Williams, D. J., J. Chem. SOC., Dalton Trans., 1983, 607. Colquhoun, H. M., Doughty, S. M., Stoddard, J. F., and Williams, D. J., Angew. Chem. Znt. Ed. Eng., 1984,23,235. Colquhoun, H. M., Stoddard, J. F., Williams, D.J., Wolsten- holme, J. B., and Zarzycki, R., Angew. Chem. Znt. Ed. Eng., 1981,20,1051. Colquhoun, H. M., Stoddard, J. F., and Williams, D. J., J. Chem. SOC., Chem. Commun., 1981,847. Alston, D. R., Slawin, A. M. Z., Stoddard, J. F., and Williams, D. J., Angew. Chem. Int. Ed. Eng., 1984,23, 821. Colquhoun, H. M., Goddings, E. P., Maud, J. M., Stoddard, J. F., Williams, D. J., and Wolstenholme, J. B., J. Chem. SOC., Chem. Commun., 1983, 1140. Allwood, B. L., Colquhoun, H. M., Crosby, J., Pears, D. A., Stoddard, J. F., and Williams, D. J., Angew. Chem. Int. Ed, Eng., 1984,23, 824. Allwood, B. J., Crosby, J., Pears, D. A., Stoddard, J. F., and Williams, D. J., Angew. Chem. Int. Ed. Eng., 1984,23, 977. Maud, J. M., Stoddard, J. F., Colquhoun, H. M., and Williams, D.J., Polyhedron, 1984, 3, 675. Doughty, S. M., Stoddard, J. F., Colquhoun, H. M., Slawin, A. M. Z., and Williams, D. J., Polyhedron, 1985,4, 567. Allwood, B. L., Fuller, S. E., Ning, P. C. Y. K., Slawin, A. M. Z., Stoddart, J. F., and Williams, D. J., J. Chem. SOC., Chem. Commun., 1984, 1356. Cram, D. J., and Cram, J. M., Acc. Chem. Res., 1978,11, 8. Petranek, Y., and Ryba, O., Anal. Chim. Acta, 1974,72,375. Bogatsky, A. V., Likyanenko, N. G., Golubev, V. N., Nazarova, N. Yu., Karpenko, L. P., Popkov, Yu. A., and Shapkin, V. A., Anal. Chim. Acta, 1984, 157, 151. Bussmann, W., Lehn, J.-M., Oesch, U., Plumere, P., and Simon, W., Helv. Chim. Acta, 1981,64, 657. Windholz, M., Editor, “The Merck Index,” 19th Edition, Merck Co., Inc., Rahway, NJ, 1983, p. 3369 and 6896. Moody, G. J., Oke, R. B., and Thomas, J. D. R., Analyst, 1970,95, 910. Craggs, A., Moody, G. J., and Thomas, J. D. R., J. Chem. Educ., 1974, 51, 541. Jaber, A. M. Y., Moody, G. J., and Thomas, J. D. R., Analyst, 1976,101,179. Oehme, M., and Simon, W., Anal. Chim. Acta, 1976,86, 21. Armstrong, R. D., Covington, A. K., and Evans, G. P., Anal. Chim. Acta, 1984, 166, 103. Meier, P. C., Morf, W. E., Laubli, M., and Simon, W., Anal. Chim. Acta, 1984, 156, 1. Nieman, T. A., and Horvai, G., Anal. Chim. Acta, 1985, 170, 359. Senkyr, J., Ammann, D., Meier, P. C., Morf, W. E., Pretsch, E., and Simon, W., Anal. Chem., 1979,51,786. Schneider, J. K., Hofstetter, P., Pretsch, E., Ammann, D., and Simon, W., Helv. Chim. Acta, 1980, 63, 217. Morf, W. E., and Simon, W., in Freiser, H., Editor, “Ion Selective Electrodes in Analytical Chemistry 1 ,” Plenum Press, New York and London, 1978, p. 232. Dobler, M., “Ionophores and their Structures,” Wiley, New York, 1981.ANALYST, FEBRUARY 1987, VOL. 112 127 45. 46. 47. 48. 49. 50. 51. 52. Pedersen, C. J., J . Am. Chem. SOC., 1967,89,7017. Dietrich, B., J . Chem. Educ., 1985, 62, 554. Petranek, J., and Ryba, O . , Anal. Chim. Acta., 1974,72,375. Senkyr, J., and Kouril, K., J . Electroanal. Chem. Interfacial Electrochem., 1984, 180, 383. Matsui, M., and Freiser, H., Anal. Lett., 1970, 3 , 161. Paper A61265 Scholer, R., and Simon, W., Helv. Chim. Acta, 1972,55,1801. Baum, G., and Lynn, M., Anal. Chim. Acta, 1973, 65, 393. Morf, W. E., Ammann, D., and Simon, W., Chimia, 1974,28, Received August 8th, 1986 65. Accepted September 16th, 1986 53. Mod, W. E., Kahr, G., and Simon, W., Anal. Lett., 1974,7,9. 54. Davies, J. E. W., Moody, G. J . , Price, W. M., and Thomas, J. D . R., Lab. Pract., 1973, 22,20.

ISSN:0003-2654    

DOI:10.1039/AN9871200121

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 129 Optimisation of the Micro-scale Determination of Phosphates by Direct Potentiometric Titration with Silver Ions and its Application to the Determination of Phosphorus in Organic Compounds Agostino Pietrogrande and Mirella Zancato Department of Pharmaceutical Sciences, University of Padua, Via Marzolo 5, 3573 7 Padua, Italy and Gin0 Bonterqpelli Institute of Chemistry, University of Udine, Wale Ungheria 43, 33700 Udine, Italy The micro-scale determination of orthophosphate ions by argentimetric titration with potentiometric detection of the end-point using a combined massive silver electrode was investigated. The best results were obtained when this titration was carried out in water - isopropanol (50% VN) buffered at pH 8.80. Under such conditions the titration error arising from the partial coprecipitation of silver oxide near the equivalence point is minimised and the results obtained have a satisfactory accuracy (errors within f0.30%) and good reproducibility. The concentration range in which reasonable accuracy and precision are attained and the most suitable titration rate to be adopted when an automatic titrimetric device is used were also considered.This optimised titration method was successfully applied to the determination of phosphorus in organophosphorus compounds after mineralisation in a suitably modified Schoniger flask. Keywords : Phosphate determination; argentimetric titration; potentiometric titration; micro-scale analysis; organophosphorus determination All analytical procedures developed for the determination of phosphorus in organic compounds involve the initial decom- position of the substance followed by the complete conversion of phosphorus to orthophosphate, for which many determina- tion methods are available.On reaction with ammonium molybdate, this species can be determined either directly by gravimetry as ammonium phosphomolybdatel or spectro- photometrically after selective reduction to molybdenum blue.2 Complexometric titration procedures3 or the iodimetric determination of hydrogen ions present in orthophosphoric acid4 can also be employed. Two methods based on the precipitation of phosphates as the silver salt have been suggested. In the first, the precipitate is dissolved with a nickelocyanide solution and the de-masked nickel is then titrated with EDTA,S whereas in the second method the silver ions set free are determined by standard Volhard titration.6 More recently, the determination of phosphate by cetylpyridi- nium employing constant-current potentiometry has been proposed.7 The determination of orthophosphates by a more con- venient and rapid direct argentimetric titration with potentio- metric detection of the end-point has been proposed,%lO and more recently applied to micro-scale determinations.1 1 ~ 2 In these papers, however, no detailed investigations aimed at finding the optimum experimental conditions (namely pH and medium) for carrying out such a titration were reported. In the most recent study,12 the use of an expensive silver sulphide ion-selective electrode for monitoring the e.m.f.during the titration was recommended in order to achieve satisfactory results. This paper reports the results obtained during the titration of phosphates with silver ions using the commercially available and less expensive massive silver electrode for the poten- tiometric detection of the end-point. The purposes of this investigation were: (i) to establish the most convenient pH values for performing such a titration; (ii) to choose the solvent that enables the titration to be optimised; (iii) to determine the concentration range in which reasonable accuracy and precision are attained; and (iv) to determine the most suitable titration rate to be adopted when an automatic titrimetric device is used. The optimised procedure was applied to the determination of orthophosphate ions derived from the decomposition of some phosphorus-containing organic compounds.Experimental All titrations were carried out using a Metrohm 636 Titro- processor connected to a Metrohm 635 automatic burette (total volume 20 ml) and a Metrohm EA 246 massive silver combination electrode. The titrant employed was a 0.01 M silver nitrate solution standardised by levamisole hydro- chloride (standard deviation fO.O6%). Procedure for Phosphate Samples Samples containing 3-6 mg of phosphate ion were weighed, placed in a 50-ml beaker and dissolved in 30 ml of a 0.0125 M sodium tetraborate solution in water - isopropanol (50% VIV). Dilute acetic acid (20%) was then added dropwise until the pH reached 8.80 (controlled by a Metrohm EA 120 combined glass pH electrode connected to the Titro- processor).The silver electrode and the outlet tube of the burette were inserted to within 1 cm of the bottom of the beaker. Magnetic stirring and “control card operation” were then started and automatic titration with 0.01 M silver nitrate solution was carried out. Titration curves, potential breaks and end-point volumes were recorded and printed. The use of the “control card operation” programmed for “dynamic titration, Kinetic T3” is recommended when the above automatic titrimetric device is used. This setting gives a fixed time interval (20 s) between every titrant addition, the volume of which is progressively decreased automatically as the end-point is approached. At the end of each titration series (about 12 determinations) it is advisable to wash the electrode surface by dipping it into dilute ammonia solution.130 ANALYST, FEBRUARY 1987, VOL.112 2 4 6 8 10 12 PH Fig. 1. Comparison of the variation with pH of the Ag+ concentra- tion required to begin the precipitation of Ag,O (A) with that exhibited by Agf from the dissolution of Ag3P04 (B). pK, = 19.9 for Ag3P04 and pK, = 7.6 for Ag20 (relative to the equilibrium 95 Ag,O + '/2 HzO + Ag+ + OH-) Procedure for Determination of Phosphorus in Organic Compounds Mineralisations were performed in a modified Schoniger flask at the bottom of which was a small beaker-shaped area where the electrode and the outlet tube of the burette were inserted after combustion. This modification, described in a previous paper,13 makes the combustion vessel suitable for the direct titration of the phosphates formed, thus avoiding transfers from the flask into a separate titration beaker, which may give rise to negative errors.The combustion step was carried out by charging the flask with 12 ml of water and 0.8 ml of 30% hydrogen peroxide solution. The samples (containing not less than 1 mg of phosphorus) were weighed, wrapped in a piece of Schleicher and Schull5892 paper and then placed in the platinum basket. Oxygen was blown into the flask and the samples were burned in the usual way. After about 30 min, the flask was opened and the stopper, basket and the inner side of the flask were washed with 15 ml of isopropanol and then 3 ml of 0.125 M aqueous sodium tetraborate were added.A few drops of acetic acid (about 20%) were added, if necessary, to adjust the pH to 8.80 and the titration procedure given under Procedure for Phosphate Samples was followed. Results and Discussion In the titration of phosphates with silver ions, the pH can be expected to be significant as a consequence of the strongly basic character of the titrated anion. This can be seen in Fig. 1, which compares the variation with pH of the concentration of free silver ions required to begin the precipitation of Ag20 (line A) with the trend exhibited (line B) by the concentration of silver ions coming from the dissolution of silver phosphate (this occurs at the equivalence point of the titration). It can be seen that in acidic media, where the formation of the interfering silver oxide does not occur, the high solubility of Ag3P04 prevents reliable titrations being carried out.Also, in alkaline media (pH > 10.9) these titrations become unfeasible owing to the precipitation of Ag2O before the equivalence point, as AgzO becomes much less soluble than Ag3P04 as the pH increases. The most convenient pH interval for carrying out such a titration is from 8.5 to 10.5, where Ag3P04 still exhibits a Table 1. Data obtained in the titration of K2HP04 (phosphate content: 54.54%) in 50% V/V aqueous methanol solutions containing the tetraborate buffer (0.0125 M). Number of determinations (n) at each pH value: 6 Phosphorus found (YO) at 95% confidence interval (mean value +&/-\/n) PH 8.80 55.56 f 0.42 9.70 54.99 f 0.15 10.50 55.11 _+ 0.11 Table 2.Data obtained in the titration of K2HP04 (phosphate content 54.54%) in 50% V/V aqueous isopropanol solutions containing the tetraborate buffer (0.0125 M). Number of determinations (n) at each pH value: 6 Phosphorus found (YO) at 95% confidence interval PH (mean value +&/Vn) 8.80 54.67 & 0.08 9.70 54.69 _+ 0.11 10.50 55.22 f 0.12 ~~~ ~ ~ ~ ~ Table 3. Results obtained in the titration of K2HP04.3H20 (phos- phate content 41.62) and NaH2P04.H20 (calculated phosphate content 68.81) samples dissolved in 50% WV aqueous isopropanol solutions containing the tetraborate buffer (0.0125 M) at pH 8.80. Number of determinations (n) for each compound: 6 Phosphorus found (YO) at 95% confidence interval Compound (mean value f2dVn) KZHP04.3H20 41.85 f 0.02 NaH2P04.H20 68.56 f 0.11 sufficiently low solubility and, at the same time, Ag20 begins to precipitate only beyond the equivalence point.Hence, the formation of mixed precipitates should be minimised and the presence of an inconveniently short vertical break at the end-point, due to the closeness of the solubility of the two silver compounds formed sequentially, should be avoided. Some preliminary titrations were performed in this pH range on stock solutions of anhydrous K2HP04 (phosphate content 54.54%) dissolved in water containing tetraborate buffer adjusted to the desired pH values. Although the most satisfactory results were obtained at the lower limit of this pH range, the results obtained at all the pHs employed were characterised by high values for phosphates and a low reproducibility, probably due to the coprecipitation of Ag20 and Ag3P04 near the equivalence point.With the aim of improving these results, we tested the medium (containing up to 70% methanol) recommended by Selig for the potentiometric titration of orthophosphates with silver nitrate using a silver sulphide electrode.12 A series of titrations were performed in the pH range 8.50-10.50 using media containing different concentrations of methanol. In spite of the appreciable improvement observed under these experimental conditions, both the accuracy and precision remained unsatisfactory, as shown in Table 1. Subsequently, mixtures of water and isopropanol were used as the solvent in order to take advantage, as reported in previous work on the sequential titration of halides with silver ions,13J4 of the favourable effect of this alcohol on the solubility constants of the two silver compounds.The surface-active properties of isopropanol should also reduce the incidence of coprecipita- tion. Hence, in order to establish the best conditions for minimising the titration error, water - isopropanol with isopropanol concentrations ranging from 0 to 50% V/V was tested in the pH range 8.50-10.50 (isopropanol contentsANALYST, FEBRUARY 1987, VOL. 112 131 Table 4. Results obtained in the determination of phosphates derived from combustion in a Schoniger flask of some phosphorus-containing organic compounds. Number of determinations (n) for each compound: 6 Compound Phosphorus content, ‘YO Found at 95% confidence interval (mean value Calculated k W V n ) Triphenylphosphine, P(C6H5)3 .. . . . . 11.81 11.86 f 0.13 1,2-Bis(diphenylphosphine)butane, (C~H~)~P(CH~)~P(C~HS)~ . . . . . . 14.55 14.52 k 0.11 1,2-Bis(diphenylphosphine)ethane, (C6H5) zP ( CHz) zP (C6H-5) 2 . . . . . . 15.5 7 15.52 k 0.08 higher than 50% VIV caused the partial precipitation of tetraborate salts). This series of tests showed that the best results were achieved for the highest concentration of isopropanol (50% VIV), irrespective of the pH. Also, when this medium is used, both the accuracy and precision decrease with an increase in pH from 8.80 as shown in Table 2. It is seen that the relative error and the confidence interval increase from 0.13 to 0.53% and from 0.08 to 0.12, respectively, on increasing the pH from 8.80 to 10.50. Consequently, the most favourable pH in this medium is 8.80, at which the minimum error and the smallest scatter were obtained.The confidence interval found at this pH is only slightly higher than that expected on the basis of the uncertainty relative to the standardisation of the titrant solution (see under Experimental), indicating that there is a very low intrinsic error in the phosphate titration. This finding confirms the theoretical expectations reported above on the precipitation of Ag3P04 and Ag20 as a function of pH, in spite of the presence of isopropanol, the effect of which on the equilibria (Fig. 1) is apparently not so strong as to cause a substantial change of the basic characteristics of this particular solubility problem.The results reported were obtained by adjusting the titration rate such that the automatically pre-selected titrant additions were delivered every 20 s (setting “dynamic titra- tion, Kinetic T3,” see under Experimental), allowing the titrations to be completed in about 12 min. More frequent titrant additions caused higher positive titration errors, whereas lower titration rates were not used in order to avoid excessively long execution times. Slow rates also led to the formation of a solid deposit on the electrode surface. Before beginning each series of determinations, the electrode was conditioned by titrating at least one solution containing a roughly weighed amount of K2HP04. The reliability of the above procedure was tested on two other compounds with different phosphate contents [NaH2P04.H20 (68.81% phosphate) and KzHPO4.3H20 (41.62% phosphate)].These results are reported in Table 3 and indicate that the careful choice of experimental conditions on the above basis enables satisfactory phosphate determina- tions to be achieved (errors within &0.30%). This accuracy and precision were obtained for all the test compounds employed in the concentration range 0.1-0.2% (100-200 p.p.m.) in 30 ml of sample. Lower concentrations led to poorly reproducible determinations and to unsatisfactory titration errors, whereas higher concentrations were not explored as they required sample amounts exceeding the usual limits of micro-scale determinations. The proposed method can be successfully employed for the direct determination of phosphates dissolved in water.For this purpose it is sufficient to dilute the aqueous samples (15 ml) with a suitable amount of isopropanol and then to adjust the pH of the resulting solution to 8.80 by adding tetraborate and acetic acid (see under Experimental). Such a procedure may be applied only if the resulting phosphate concentration falls within the range cited above. Determination of Phosphorus in Organic Compounds The method was applied to the determination of phosphorus in some organic compounds in order to test its suitability for the determination of the element. The results, obtained after combustion carried out as described under Experimental, are reported in Table 4. Satisfactory accuracy and precision were obtained. The proposed titration method therefore allows reliable results to be obtained simply and quickly using inexpensive apparatus. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Lieb, H., and Wintersteiner, O., Mikrochemie, 1924, 2 , 78. Roth, H., Mikrochemie, 1944,31, 290. Piischel, R., and Wittmann, H., Mikrochim. Acta, 1960,670. Gawargious, Y. A., andFarag, A. B., Microchem. J . , 1971,16, 333. Maric, Lj., Siroki, M., and Stefanac, Z., Microchem. J., 1976, 21, 129. Shanine, S., and El Medany, S., Microchem. J., 1979,24,212. Selig, W., Mikrochim. Acta, 1984,II, 133. Flatt, R., and Brunisholz, G., Anal. Chim. Acta, 1947,1, 124. Christian, G. D., Knoblock, E. C., and Purdy, W. C., Anal. Chem., 1963,35, 1869. McColl, D. H., and O’Donell, T. A., Anal. Chem., 1964,36, 848. Selig, W., Mikrochim. Acta, 1970, I, 564. Selig, W., Mikrochim. Acta, 1976, 11, 9. Pietrogrande, A., Zancato, M., and Bontempelli, G., Analyst, 1985,110,993. Pietrogrande, A., and Zancato, M., Mikrochim. Acta, 1985, 11, 283. Paper A61147 Received May 16th, 1986 Accepted September 24th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200129

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 133 Open Thin-layer Cell-A Flow-through Electrode for Potentiometric Stripping Analysis Wolfgang Frenzel” and Gerhard Schulzet lnstitut fur Anorganische und Analytische Chemie, Technische Universitat Berlin, Strasse des 17, Juni 135, 0-1000 Berlin 12, FRG An electrochemical flow-through cell is described that is substantially different from all conventional cells. It has an open structure and is made from a piece of Plexiglass that contains the working and auxiliary electrodes. The flow channel is replaced by a strip of filter paper, which adheres to the cell body and covers the inlet hole and both electrodes. The carrier solution is therefore precisely directed along the filter strip, which dips into the waste reservoir containing the SCE.Potentiometric stripping analysis (PSA) was used to evaluate the performance characteristics of the cell in terms of flow-rate, injection volume and carrier composition. Linear calibration graphs were obtained in two concentration ranges (ICr500 pg 1-1 and 1-10 mg 1-1) for Cd, Pb and Cu under slightly different experimental conditions. The practical advantages and the long-time response stability of the cell were demonstrated by the PSA determination of Zn in tap water. In the range 0.1-20 mg I-’ a sample throughput of 120 h-1 is obtainable with a reproducibility better than 5%. Keywords: Potentiometric stripping analysis; electrochemical flow-through cell; heavy metal determination; zinc determination; tap water Electrochemical stripping analysisl-2 is a useful tool for the determination of trace amounts of heavy metals.The high sensitivity, specifity and the possibility of simultaneous determination inherent in stripping analysis are distinctive features that make it attractive in the environmental, clinical, industrial and other fields of analytical chemistry. Anodic stripping voltammetry (ASV) in the linear scan and differen- tial-pulse mode is so far the most used technique. However, during the last decade potentiometric stripping analysis (PSA)3*4 has found increasing interest because of some superior attributes ( i e . , no interference by oxygen, less susceptibility to adsorption phenomena, inexpensive and simple instrumentation).s7 Progressive interest in the conti- nuous monitoring of heavy metals has resulted in the adaption of ASV and PSA to flow-through measurements.8-10 The performance of stripping analysis in flowing systems offers several advantages, e.g., ease of automation, high sample throughput and medium exchange.However, the design of flow cells is still an interesting subject as the question of which cell is best suited to a particular application has not yet been answered conclusively. Mercury films on glassy carbon supports (MFGCE) are often the preferable working elec- trode because of their extreme sensitivity, good stability under flow conditions and ease of preparation. They have been incorporated into a variety of different flow-through cells, including thin-layer cells79J1J2 tubular,13 wall-jet ,14115 and porousl6J7 electrodes. The main aspects in the choice of the cell are an effective mass transport and well-defined hydrody- namic conditions. One drawback of all conventional flow- through cells is that they require dismantling when air bubbles appear at the electrode surface or when the electrodes have to be cleaned.In this paper we report a novel cell design referred to as an “open thin-layer cell” that overcomes this disadvantage. The flow channel is replaced by a strip of filter paper that adheres to the cell body containing the glassy carbon working and * Present address: Hahn-Meitner Institut GmbH, AG Spuren- elementforschung in der Biomedizin, Glienicker Strasse 100, D-1000 Berlin 39, FRG. f To whom correspondence should be addressed. auxiliary electrodes. The response characteristics of the cell are evaluated with respect to the size of the filter strip, carrier composition, flow-rate, deposition time and metal ion concen- tration.The utility of this electrode and its advantages when compared with other cell designs are demonstrated by the determination of zinc in tap water. Experimental Instrumentation The potentiometric stripping analysis was performed with a Striptec System (Tecator, Hoganas, Sweden). The open thin-layer cell was simply constructed by using the upper part of the thin-layer cell described previously11 containing a 1-mm glassy carbon electrode (GCE) and a platinium wire auxiliary electrode. A common laboratory adhesive was used to fix this part vertically, with the sample inlet at the upper side (Fig. 1). Ordinary filter paper (Schleicher and Schiill, No 5892) was cut into small strips, wetted with distilled water and placed on the Plexiglass surface in order to cover the inlet hole, glassy carbon and auxiliary electrodes. The lower end of the strip was dipped into the waste reservoir, which contained the saturated calomel reference electrode (SCE).In I I J SCE i Fig. 1. Schematic diagram of open thin-layer cell. GCE = 1 mm glassy carbon working electrode; Pt = 0.2 mm awilary electrode; SCE = saturated calomel reference electrode; and F = filter strip134 ANALYST, FEBRUARY 1987, VOL. 112 The carrier solution was propelled by means of variable air pressure using a FIAstar unit (Tecator AB), which allows flow-rate adjustments in the range 1-10 ml min-1. Lower flow-rates (0.2-2 ml min-1) were maintained by gravity flow.All flow-rates were calibrated volumetrically . The intercon- nections between the carrier reservoir, valve and cell were made of 0.5 mm i.d. PTFE tubes. Sample volumes down to 14 yl were injected by varying the loop length or the diameter of the valve. Chemicals and Solutions Appropriate metal ion concentrations were prepared by dilution of stock solutions (Merck, Titrisol) with bi-distilled water each day. The mineral acids and salts used were high purity reagents (Merck, Suprapur). All other chemicals were of analytical-reagent grade. The carrier solution contained 0.1 mol 1-1 HCl and 20 mg 1-1 Hg2+. For continuous measure- ments it was spiked with appropriate amounts of Pb2+. Some experiments were performed with a de-aerated carrier solu- tion; nitrogen was bubbled through the carrier reservoir prior to analysis for at least 3 h and throughout the measurements.Procedure The filter strip was fixed at the Plexiglass surface of the cell body. A mercury film was pre-plated at the GC electrode by Time Fig. 2. Five of about 50 PSA curves recorded during a 2 h period. The 500 pl sample contains 0.5 mg 1-1 of Zn2+ and 1 mg 1-1 each of Cd2+ and Pb2+. Flow-rate, 1.8 ml min-1; deposition potential, - 1.3 V vs. SCE. The peaks in the curves are the result of the fact that the Tecator system records the sum of the potential and its derivative on the ordinate setting a potential of -1.0 V to the working electrode for at least 3 min while the carrier solution was flowing at a rate of 1-2 ml min-1.The circuit was then disconnected and a background curve recorded. Immediately afterwards a nega- tive potential of -0.2 V was applied to the working electrode and continuously maintained in order to prevent the forma- tion of calomel. This is very critical as without this step the electrode response fails. In order to carry out the continuous flow measurements Pb2+ was added to the carrier giving a final concentration of 0.5 mg 1-1. Flow injection experiments were carried out as follows. The flow-rate was adjusted to the desired value, the deposition potential was applied and the standard or sample solutions were injected simultaneously. For the determination of zinc the deposition potential was - 1.4 V; for other metals it was - 1.0 V vs. SCE. The deposition time set at the instrument was adapted to the residence time of the sample.The latter depends on the flow-rate, injection volume and dispersion and was measured prior to the determinations as described previously. 10 Results and Discussion Performance Characteristics Cell design The influence of the size of the filter strip on the electrode response was investigated by cutting small slices of variable width. Three fundamentally different instances can be distin- guished, i.e., when the width of the strip is smaller, equal or broader than the diameter of the working electrode. Surpris- ingly, the response of the electrode was nearly identical in all three instances. However, increased carryover was observed in flow injection measurements using broader strips, i.e., the residence time of the sample plug was increased and longer washing times were required between successive samples.The reproducibility of the stripping signals obtained in the continuous flow measurements was determined from 10 successive runs at 60 s deposition time and 0.5 mg 1-1 Pb2+ in the carrier solution. The mean stripping times found were 466 ms (3.4% standard deviation) and 452 ms (4.3% standard deviation) for measurements using the same 0.5 mm strip and using 10 different strips of equal size, respectively. These data clearly demonstrate that the preparation of the actual open thin-layer cell is not critical and can be reproducibly repeated. Strips of 0.8-1.0 mm width were used in the remainder of the work. Practical aspects Our experience with several flow-through electrodes18 in electrochemical stripping analysis has shown that problems may occur during prolonged measurements due to the accumulation of air or hydrogen bubbles at the electrode surface.This leads to response failure and in such instances it is necessary to dismantle the cell, clean the electrode and plate a fresh mercury film. Recalibration is required to obtain reliable results. Table 1. Regression data for the calibration of Cd, Pb and Cu in the (i) 10-500 kg 1-1 and (ii) 1-10 mg 1-1 range. Experimental conditions: (i) Flow-rate, 0.3 ml min-1; injection volume, 1 mi. (ii) Flow-rate, 1.8 ml min-1; injection volume, 400 PI Slopel Intercept/ s 1 mg-l S Variance/ Regression Element Range S coefficient 23.71 0.204 -0.0094 0.0521 0.0816 0.9999 0.2293 0.0048 0.0788 0.0289 0.0371 0.9993 17.95 0.142 0.0204 0.0363 0.0564 0.9999 0.1153 0.0019 0.1028 0.0118 0.0151 0.9994 37.13 0.501 0.2203 0.1277 0.2004 0.9997 0.2807 0.0088 0.1930 0.0532 0.0684 0.9983 Cd .. . . (9 Pb . . . . (i) c u . . . . (9 (ii) (ii) (ii)ANALYST, FEBRUARY 1987, VOL. 112 135 When the open thin-layer cell was used, even the injection of a l-ml air plug instead of the liquid sample did not harm the electrode, although, of course, only a blank signal was obtained. More interestingly, no hydrogen bubbles were visible at the electrode under conditions where hydrogen evolution is to be expected, i.e., injection of strong acid samples (1 moll-1 HC1) and a deposition potential of -1.4 V vs. SCE. In batch measurements the electrode was completely covered with hydrogen bubbles after a few seconds under these experimental conditions. In Fig.2, five of fifty successive recorded signals are shown, which were obtained during an uninterrupted run lasting about 2 h. The sample contains 500 pg 1-1 of Zn2+ and 1 mg 1-1 of Cd2+ and Pb2+, respectively. The experimental conditions are given in the figure. It is obvious that the potential steps become less pronoun- ced, probably due to an increased background, but that the stripping time, i.e., the distance between two peaks of the derivative curve, is reproducible within 5% standard deviation for all three elements. After several working hours swelling of the filter paper was observed, leading to an increased residence time of the sample. In such instances the filter strip was replaced by a new one.Carrier composition As has been shown in previous work the composition of the carrier solution can be optimised in order to obtain better sensitivity11 or to improve the selectivity. 19 De-aeration of the carrier diminishes the concentration of oxidising agent, i. e., 02, and thus leads to longer stripping times. However, because of the open structure of the cell at least partial re-aeration of the carrier or diffusion of atmospheric oxygen through the filter paper can be expected. The influence of de-aeration on the stripping time was studied at different flow-rates in continuous flow measurements. At flow-rates below 0.6 ml min-1 no difference between de-aerated and non-de-aerated carrier solutions was observed. With increas- ing flow-rate the de-aeration effect was more pronounced.At 8 ml min-1 the stripping time of a de-aerated sample was 15 times that of the sample which had not been de-aerated. However, it has to be borne in mind that the sensitivity in flow injection potentiometric stripping analysis (FIPSA)10 at con- stant injection volume is inversely proportional to the flow-rate (see under Flow-rate). Therefore in FIPSA the advantage gained by de-aeration of the carrier solution at high flow-rates is at least partially cancelled out by the decreased residence time and hence the decreased stripping time. Flo w-rate Theoretical considerations7~20 predict no influence of the flow-rate on the stripping time in continuous flow measure- ments. This was confirmed for various Pb2+ concentrations at flow-rates in the range 0.3-8 ml min-1.However, at flow-rates higher than 5 ml min-1 the filter strip was sometimes flushed away, which leads to erroneous results. At flow-rates below 0.3 ml min-1 back diffusion from the waste reservoir was observed, causing higher blank values if the waste solution was contaminated. In addition, carryover effects were consider- ably enhanced. In flow injection measurements the stripping time is inversely proportional to the flow-rate in the range 0.34 ml min-1, as was also found for other electrode configurations.15J8 Deposition time In PSA the stripping time is directly proportional to the deposition time. This could be verified in continuous flow measurements for stripping times between 10 and 600 s for different Pb*+ concentrations.In FIPSA the actual deposition time is given by the residence time of the sample, which depends on the flow-rate, injection volume and dispersion of the sample. Low flow-rates increase the residence time of the sample and hence the deposition time. At a constant flow-rate the residence time is directly proportional to the sample volume. A linear dependence of the stripping time on injection volumes in the range 14 pl to 1 ml was found irrespective of the flow-rate chosen. Calibration procedure The choice of the experimental parameters (i.e., flow-rate, injection volume, carrier composition) easily allows the analysis procedure to be optimised for different concentration ranges. At a flow-rate of 0.3 ml min-1 and a l-ml injection volume the residence time of the sample is about 4 min.Under these conditions rectilinear calibration graphs are obtained for Cd, Pb and Cu in the range 10-500 pg 1-1. The regression data are given in Table 1. The dynamic range of PSA covers at least three orders of magnitude so that even higher concentrations can be deter- mined under identical experimental conditions. However, with respect to sample throughput and possible saturation of the mercury film21 it is advisable to change the parameters. Using a 1.8 ml min-1 flow-rate and 400-p1 injection volume the residence time of the sample is below 15 s. This allows the determination of the three elements in the range 1-10 mg 1-1 at a rate of approximately 150 h-1. The regression data for this concentration range are also given in Table 1.Application Public Health Regulations in the Federal Republic of Ger- many prescribe limiting values for several metal ion concen- trations in potable water.22 For zinc the maximum permitted value is 2 mg 1-1. In order to demonstrate the utility of the open thin-layer cell for tap water analysis the experimental conditions were chosen to allow the determination of zinc in the range 0.1-20 mg 1-1. This was achieved with a 500-yl injection volume at a flow-rate of 0.4 ml min-1. The deposition potential was -1.3 V vs. SCE. Prior to analysis standard solutions in the desired concentration range were injected three times. The mean values are given in Table 2. Least-squares treatment of the Table 2. Dependence of the stripping time (rs) on zinc concentration.Deposition potential, -1.3 V; injection volume, 500 p1; flow-rate, 1.4 ml min-1 Zinc Recalibration values concentration/ rs (mean)/ mgl-l S 1 2 3 4 5 - - 0.1 0.22 0.24 0.22 0.23 0.5 1.03 1.07 1.05 1.05 1.02 1.03 1.0 2.16 2.13 2.15 2.20 2.18 2.21 5.0 10.10 10.00 10.00 9.87 10.19 10.02 20.41 20.18 20.54 20.62 20.40 - 10.0 15.0 31.24 30.84 30.92 31.00 20.0 39.83 40.10 40.06 39.76 39.89 - - -136 ANALYST, FEBRUARY 1987, VOL. 112 data yield a linear calibration graph with a slope of 2.093 k 0.017, an intercept of -0.077 k 0.1775 and a variance of 0.330. Recalibration was performed with one of the six standard solutions after every tenth injection of a sample. No corrections were made for the zinc concentrations as long as the recalibration yielded signals within the standard deviation of the original calibration graph.This was usually applied for several hours as is indicated by the results shown in Table 2. Potable water samples were collected in 100-ml polyethylene bottles at private taps in West Berlin, spiked with 500 p1 of concentrated €€NO3 and measured within 3 d of collection. All samples were injected at least twice. In order to check the reproducibility one sample was subsequently injected 20 times. The concentration found was 0.65 f 0.03 mg 1-1 (4.6% standard deviation). The well known interference of copper in the determination of zinc by electrochemical stripping tech- niques23.24 caused no problems in our studies, as the copper concentration, which was determined simultaneously, was well below that of zinc in all samples.Conclusions From a practical point of view the open thin-layer cell offers several advantages when compared to conventional flow- through detectors. It is cheap and easy to make, can be cleaned without dismantling and is not susceptible to bubble formation at the electrode surface. This is particularly important in the determination of zinc in acid solutions when hydrogen evolution occurs. The response stability of this cell under prolonged measure- ments is excellent, as is demonstrated in the determination of zinc in tap water at high sampling rates. The poor sensitivity due to fast oxidation by atmospheric oxygen can be improved by the use of computerised PSA.25 Thus, other elements such as Cd and Pb that are present at much lower levels could also be monitored.The authors thank Frank Chisela for his interest in our work and helpful discussions in the final stage of preparation of this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. References Vydra, F., Stulik, K., and JulAkovi, E., “Electrochemical Stripping Analysis,” Ellis Horwood, Chichester, 1976. Wang, J., “Stripping Analysis,” Verlag Chemie, Weinheim, 1985. Jagner, D., Analyst, 1982, 107, 593. Jaya, S., and Rao, T. P., Rev. Anal. Chem., 1982, 6, 343. Florence, T. M., J. Electroanal. Chem., 1984, 168,207. Wahdat, F., and Neeb, R., Fresenius 2. Anal. Chem., 1983, 316, 770. Han, E., Teachers Diploma Thesis, Techical University Berlin, 1985. Wang, J., Dewald, H. D., and Green, B., Anal. Chim. Acta, 1983, 146, 45. Anderson, L., Jagner, D., and Josefson, M., Anal. Chem., 1982, 54, 1371. Frenzel, W., and Bratter, P., Anal. Chim. Acta, 1986,179,389. Schulze, G., and Frenzel, W., Fresenius 2. Anal. Chem., 1983, 316, 26. Wise, J. A., Heineman, W. R., and Kissinger, P. T., Anal. Chim. Acta, 1985, 172, 1. Seitz, W. R., Jones, R., Klatt, L. N., and Mason, W. D., Anal. Chem., 1973,45, 840. Gunasingham, H., Ang, K. P., and Ngo, C. C., Anal. Chem., 1985,57,505. Schulze, G., Husch, M., and Frenzel, W., Mikrochim. Acta, 1984, I, 191. Blaedel, W. J., and Wang, J., Anal. Chem., 1980, 52,76. Cohen, A., and Bartak, D. E., Anal. Lett., 1983,16,429. Frenzel, W., Thesis, Technical University Berlin, 1984. Schulze, G., Bonigk, W., and Frenzel, W., Fresenius 2. Anal. Chem., 1985,322, 255. Frenzel, W., unpublished work. Frenzel, W, , Diploma Thesis, Technical University Berlin, 1981. “Verordnung uber Trinkwasser und uber Brauchwasser fur Lebensmittelbetrieber (Trinkwasser-Verordnung),” Bundes- gesetzblatt Teil I, Z 1997A, Nr. 16, 1975,453, Bonn. Shuman, M. S., and Woodward, G. P., Anal. Chem., 1976,48, 1979. Schulze, G., and Frenzel, W., Frensenius 2. Anal. Chem., 1983,314,459. Jagner, D., Trends Anal. Chem., 1983,2, 53. Paper A61216 Received July 7th, 1986 Accepted September 26th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200133

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 137 Determination of Potassium Bromate in Flour by Flow Injection Analysis Brian G. Osborne Flour Milling and Baking Research Association, Chorleywood, Hertfordshire WD3 5SH, UK Potassium bromate was determined in flour by aqueous extraction and photometric FIA based on its reaction with acidified potassium iodide and starch. The recovery was 90.0% and the FIA linear range was 0.5-5.0 mg I-' with between-injection and between-extraction standard deviations of 0.04 and 0.11 mg 1-1, respectively. For 22 samples of commercial flour the mean differences between the results obtained by the FIA method and by the current official titration method was 0.14 mg kg-1 and the standard deviation of differences between the two methods was 0.74 mg kg-1 in the range 0-20 mg kg-1.Keywords : Potassium bromate determination; flour; flow injection analysis The breadmaking quality of freshly milled flour tends to improve with storage for a period of up to two months but this process may be accelerated by the addition of chemical substances called improvers. Among other substances, the Bread and Flour Regulations 19841 permit the use of potassium bromate (BR) up to a maximum level in bread of 50 mg kg-1 of flour mass. Potassium bromate may be added to the flour at the mill or to the dough at the bakery or both provided that the total level is within the limit given above. It is necessary therefore for both millers and bakers to be able to analyse flour rapidly in order to determine the BR content. A semi-quantitative test for BR in flour, based on its reaction with acidified potassium iodide, has been carried out in mill laboratories for many years2 and the same reaction forms the basis of several quantitative methods of determina- tion.In one method, the liberated iodine is determined colorimetrically after the formation of a complex with starch,3*4 but the colour intensity is unstable due to the slow release of iodine from the reagents. Precise timings are therefore necessary, which make the method vulnerable to differences in technique between analysts. It is more satisfac- tory to react the liberated iodine with sodium thiosulphate and to titrate the excess thiosulphate against potassium iodate.5fj Ion chromatography has been used for the determination of BR in flour7 and compound bread improversstg but its application is not straightforward because of interference from chloride ions.This paper describes how flow injection analysis10 (FIA) may be used to develop the colorimetric procedure394 into a method which is as reproducible as, but more rapid than, the official titrimetric method.6 Experimental Apparatus The FIA system used was a Tecator FIAstar 5020 Analyzer with a 5032 Detector Controller and Chemifold Type 11. The pump tubes, injector, mixing coils and detector were thermo- stated to 37 k 1°C. Reagents All reagents were of analytical-reagent grade. Zinc sulphate solution, 20 g 1-1. Prepared in distilled water. Sodium hydroxide solution, 21 g 1-1. Prepared in distilled Dilute sulphuric acid solution, 1 + 9. Prepared to contain Potassium iodide solution, 10 g 1-1.Prepared in distilled water. 0.3 g 1-1 of ammonium molybdate. water. Prepare fresh daily or stabilise with a few millilitres of sodium hydroxide. Starch solution, 10 g 1-1. Prepared by mixing 1 .O g of soluble starch to a creamy paste with a little water and pouring the paste in 100 ml of briskly boiling water. Prepare fresh every few days. Carrier solution. Prepared from 0.45 g of zinc sulphate, 0.59 g of sodium sulphate (anhydrous) and 25 ml of starch solution in 100 ml of distilled water. Potassium bromate solutions. Dry potassium bromate at 110 "C for 1 h and prepare a stock solution of 0.2500 g 1-1. Prepare by dilution working standards of 25, 50, 75, 100 and 125 mg 1-1. Procedure Weigh 5 g of flour into a 100-ml screw-capped bottle and add by pipette 20 ml of 20 g 1-1 zinc sulphate solution.Stopper the bottle and shake to disperse the flour, open and add by pipette 4 ml of 21 g 1-1 sodium hydroxide solution and 1 ml of distilled water. Carry out a similar procedure using untreated, unbleached (UU) flour with 1 ml of distilled water and 1 ml each of the BR working standards. Shake the bottles for 5 min on a wrist-action shaker, transfer the contents into centrifuge tubes and centrifuge for 5 min at 1150 g. Filter the supernatant through a Whatman No. 4 paper and analyse the filtrate by FIA using the manifold shown in Fig. 1 with the detector set to gain factor 2. By means of the peak heights for the UU flour (B), sample (S) and the working standard (R) of concentration C mg 1-1 having a peak height closest to that of the sample, calculate the BR content of the flour as S-B C R-B 5 BR = - x - mg kg-1 Before closing down the FIA system, flush all the lines with the 21 g 1-1 sodium hydroxide solution for a few minutes then flush with distilled water.Fig. 1. FIA manifold for the determination of potassium bromate. C = camer solution at 1.1 ml min-1; R1 = dilute sulphuric acid (1 + 9) containing 0.3 g 1-1 ammonium molybdate at 1.5 ml min-l; R2 = 10 g 1-1 potassium iodide at 0.8 ml min-l138 ANALYST, FEBRUARY 1987, VOL. 112 Results and Discussion Choice of FIA Manifold If BR is treated in acid solution with a large excess of iodide ion, the bromate ion is quantitatively reduced and an equivalent amount of iodine liberated. KBr03 + 6KI + 3H2S04+ KBr + 312 + 3K2S04 + 3H20 Ammonium molybdate is usually added to increase the rate of this reaction. In photometric FIA, the liberated iodine may be reacted with starch in the presence of iodide ions to form a complex with an absorption maximum at 570 nm (Fig. 2).Various manifold designs with different orders of reagent addition and sizes of mixing coils were investigated. As potassium iodide is unstable in acid solution,4 these two reagents must be introduced separately, although varying the order of addition did not affect the response because the reaction is fast; stopped-flow measurements showed that it was essentially complete in 11 s. It proved unnecessary to add the starch and ammonium molybdate as individual reagent streams and these were placed in the carrier and sulphuric acid streams, respectively.Varying the coil dimensions had very little effect on the maximum absorbance of a 5 mg 1-1 BR standard (Table 1). The FIA manifold was therefore designed with the aim of achieving a medium dispersion by maximising the sample volume and minimising the length of the reaction coils. The final design (Fig. 1) gave a residence timelo of 15 s and a dispersion coefficientlo of 4.47; the maximum sampling rate was 240 samples h-1 and the linear range 0.5-5.0 mg 1-1. Wavelengthlnrn Fig. 2. Absorption spectrum of starch - iodine complex [Starchjig I-’ 0 1 2 3 4 5 6 7 8 9 10 I 60 [Kll/g I-’ 0 2+98 4 + 9 6 6 + 9 4 8+92 1+9 Dilute sulphuric acid Fig. 3 Effect of reagent concentrations on maximum absorbance of starch - iodide complex.A, [starch]; B, [KI]; C , dilute sulphuric acid Selection of Reagent Concentrations The effects on the absorbance of a 5 mg 1-1 BR standard of various concentrations of starch, potassium iodide and sul- phuric acid are shown in Fig. 3. Preliminary experiments showed the starch concentration to be the most critical so this was studied first with 100 g 1-1 of potassium iodide solution and dilute sulphuric acid (1 + 9). On the basis of the results given in Fig. 2, a starch concentration of 2.5 g 1-1 was chosen while the other components of the carrier solution arise from the reagent blank. The effect of potassium iodide concentration was investi- gated next, with 2.5 g 1-1 of starch and dilute sulphuric acid (1 + 9), and 10 g 1-1 of starch was selected as the optimum concentration.Finally, the effect of sulphuric acid concentra- tion was studied and a (1 + 9) dilution of the concentrated acid was selected as the optimum concentration. Recovery of Potassium Bromate from Flour After the optimisation of the FIA procedure, attention was directed to the recovery of BR from flour. The manual procedures was followed but scaled down ten-fold in order to improve the over-all speed and convenience of the method. The mean recovery of BR added to UU flour in the range 10-50 mg kg-1 was 90.0%, which is comparable to that obtained using the standard method.6 The reason for the scale of the manual method appears to be the theoretical problem of sampling errors due to the small amount of BR present and the density difference between it and flour.4Y5 However, the precision and accuracy of the FIA method given below suggest that sampling is not a serious problem. Precision and Accuracy The precision of the method was evaluated by carrying out a number of determinations and recoveries with duplicate extraction and injection (Table 2).The standard deviation of Table 1. Effect of sizes of reaction coils on absorbance of starch - iodine complex Coils R1 Length/ 1.d.l mm mm 300 0.5 300 0.5 600 0.5 300 0.7 300 0.7 600 0.7 R2 Length/ 1.d.l mm mm Absorbance x 2 300 0.5 0.915 600 0.5 0.910 600 0.5 0.874 300 0.7 0.894 600 0.7 0.883 600 0.7 0.827 Table 2. Recovery and precision for FIA determination of potassium bromate in flour Potassium bromate determinedmg kg-1 Potassium bromate Extraction 1 Extraction 2 added mg kg-1 10 15 20 25 12.5 25 37.5 50 Sample Sample Sample Injection 1 Injection 2 Injection 1 Injection 2 10.7 10.4 9.0 9.2 14.1 14.5 12.6 12.6 16.6 17.3 16.2 16.5 22.1 22.1 22.9 23.2 10.6 10.6 11.0 11.2 23.0 22.6 22.6 22.6 33.5 33.2 33.7 33.9 45.7 45.3 45.3 44.6 8.8 8.4 8.6 8.3 8.8 9.0 8.3 8.3 10.8 10.8 10.6 10.6ANALYST, FEBRUARY 1987, VOL.112 139 Interferences As the FIA method has been adapted from the existing manual procedure for BR in flour, which is based on a non-specific iodimetric reaction, it might not be applicable to flour samples containing a mixture of oxidising or reducing agents. However, blank determinations on UU flour showed only minimum values (mean 0.4 mg kg-1) and, according to current practice, BR is unlikely to be added to flour in admixture with other bread improvers, except possibly azodicarbonamide (ADA).Azodicarbonamide is occasionally added up to a concentration of 10 mg kg-1, together with BR up to 15 mg kg-1, in order to produce flour that is suitable for a variety of breadmaking processes. Table 4 shows the mean effect of 0 and 10 mg kg-1 of ADA on the absorbance obtained when the FIA determination was carried out in duplicate on flours containing 0 and 15 mg kg-1 of BR; ADA caused a mean increase in absorbance, 0.010 of which is equivalent to 0.32 mg kg-1 of BR. This interference is unlikely to give rise to errors of practical importance. Table 3. Accuracy of FIA determination of potassium bromate in flour Potassium bromate/mg kg-1 FIA 8.6 0 6.9 5.0 4.5 15.1 10.4 8.9 17.4 6.9 18.4 Titration 8.5 1.1 5.98 4.2 4.3 13.9 10.8 8.5 18.9 7.0 20.0 FIA 8.3 16.0 12.1 17.6 8.5 18.7 14.2 10.7 6.9 13.3 16.0 Titration 8.5 16.4 11.8 17.4 8.5 19.2 14.6 10.8 7.8 12.6 16.7 Table 4.Effect of azodicarbonamide (ADA) on the FIA determination of potassium bromate (BR) in flour Absorbance X 2 [ADA]/mg kg-1 [BR]/mg kg-1 Difference 10 0 0 0.075 0.068 0.007 15 0.479 0.467 0.012 Conclusion A method for the determination of potassium bromate in flour using photometric FIA has been developed. Once the FIA system has been set up and calibrated, the determination can be completed on a batch of eight samples in 15 min with a precision of k 1.2 mg kg-1. Mrs. G. M. Barrett is thanked for her technical assistance. replicates, s,, was then calculated as d(Edi212n) where di represents the individual differences between each pair of duplicates on n samples.The between-extraction s, was 0.11 mg 1-1 and the between-injection s, was 0.04 mg 1-1 on the extracts, which is equivalent to 0.60 and 0.22 mg kg-1, respectively, on the flour. The accuracy of the method was evaluated by comparing the results of determinations performed by the FIA method and by the standard AACC method 48-426; the results for 22 samples of commercial flour are given in Table 3. The mean difference between the two sets of results was 0.14 mg kg-1 and the standard deviation of differences, sd, given by -), was 0.74 mg kg-1. Fitting the data in Table 3 by the method of least squares gave calculated values for the slope and intercept which did not differ significantly (p>O.O5) from the “ideal” values of 1 and 0, respectively. Thus there is no evidence for systematic errors between the FIA and standard methods. Variance analysis showed that there was no significant difference (p>O.O5) between 8d2 and the sum of the between- extraction and between-injection variances. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References “The Bread and Flour Regulations 1984,” S.I. 1984, No. 1304, HMSO, London. Kent-Jones, D. W., and Amos, A. J., “Modern Cereal Chemistry,” Sixth Edition, Food Trade Press, London, 1967, p. 621. Hoffer, A., and Alcock, A. W., Cereal Chem., 1946,23, 66. Johnson, L. R., and Alcock, A. W., Cereal Chem., 1948, 25, 266. Armstrong, A. W., Analyst, 1952, 77, 460. “Approved Methods of the American Association of Cereal Chemists,” Eighth Edition, American Association of Cereal Chemists, St. Paul, MN, 1983, Method 48-42. Osborne, B. G., Anal. Proc., 1986, 23, 359. Haddad, P. R., and Jackson, P. E., Food Technol. Aust. 1985, 37, 305. COX, D., Harrison, G., Jandik, P., and Jones, W., Food Technol., (Chicago), 1985, 39 (7), 41. RWEka, J., and Hansen, E. H., “Flow Injection Analysis,” Wiley, New York, 1981. Paper A61221 Received July 9th, 1986 Accepted September 22nd, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200137

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 141 Kinetic Method for the Determination of Nanogram Amounts of Lead(l1) Using its Catalytic Effect on the Reaction of Manganese(l1) with 5,10,15,20=Tetrakis(4=sulphonatophenyl)porphine Masaaki Tabata Department of Chemistry, Faculty of Science and Engineering, Saga University, Saga 840, Japan Lead(ll) accelerates the reaction of manganese(l1) with 5,10,15,20-tetrakis(4-sulphonatophenyl)porphine (H2tspp). The kinetics and mechanism of this reaction were studied at pH 6-8.7, 25 "C and I = 0.1 (NaN03). Lead(l1) rapidly forms Pbll(tspp) as a reaction intermediate; Pb" (tspp) is 7200 times as reactive as the free base porphyrin (H2tspp) in the reaction with manganese(l1). Lead(l1) concentrations as low as 10-7 mot dm-scan be determined from the decrease in absorbance at 413 nm (kmax.of Hztspp) at a fixed time after the start of the reaction of manganese(l1) with H2tspp. After the separation of lead(ll) from iron and silicate by solvent extraction using n-decanoic acid, the proposed method is highly selective and free from interference from most substances usually encountered in the samples studied if carried out in the presence of cyanide as a masking agent for cadmium(l1). The molar absorptivity and Sandell's sensitivity calculated from the calibration graph 15 min after the start of the reaction are 2.75 x lo5 mol-1 dm3 cm-1 and 0.75 ng cm-1, respectively. The method was applied to the determination lead(ll) in rain and drain water. Keywords: Lead( /I) determination; metalloporphyrin formation; catalytic determination; spectrophoto- metry; water analysis There has recently been much interest in the determination of toxic metal ions such as lead in the environment, but it is often difficult to establish reliable background levels of the toxic metal, mainly owing to a lack of analytical sensitivity and contamination during sampling and analysis.Catalytic determination has been used for the determina- tion of low concentrations of toxic metals.ly2 Most of these catalytic methods are based on redox reactions and the reaction mechanisms are sometimes complicated. In this paper we report a kinetic method for the determina- tion of sub-microgram amounts of lead. The method is based on the catalytic effect of lead(I1) on the formation of a complex between manganese(I1) and 5,10,15,20-tetrakis(4- sulphonatopheny1)porphine (Hztspp) .Porphyrins have a strong absorption band (E = ca. 5 X lo5 cm-1 mol-1 dm3) and they have recently been used as highly sensitive reagents for metals.3 The rate of metal incorporation into the porphyrin to form a metalloporphyrin is lO"lO9 times slower than that of acylic ligands4-6 and it is sometimes necessary to heat the reaction mixture in order to complete the reaction.7-9 It has been found that large metal ions such as lead(I1) catalyse the reaction of manganese(I1) with H2tspp.10 The kinetics and mechanism of the reaction catalysed by trace amounts of lead are reported in this paper with the aim of elucidating the chemical behaviour of the catalyst. On the basis of these kinetic studies a kinetic method for the catalytic determination of lead has been developed.Experimental Reagents All reagents were of analytical-reagent grade unless stated otherwise. Porphyrin, 1 X 10-4 mol dm-3. H2tspp was synthesised11.12 and its sodium salt purified as described previously.10 Lead(lr) and manganese(I0 stock solutions, 10-2 mol dm-3. Prepared by dissolution of the metal nitrates in weak acid aqueous solutions (pH 2) and the concentrations stan- dardised by EDTA. A working solution of Iead(I1) was freshly prepared by the appropriate dilution of the stock solution. n-Decanoic acid - chloroform solution, 1 mol dm-3. n-Decanoic acid was distilled twice and 43 g dissolved in 250 cm3 of chloroform. Hydroxylammonium sulphate solution, 0.05 mol dm-3. Sodium borate solution, 5.0 x 10-2 mol dm-3. Sodium nitrate solution, 1.0 rnol dm-3.The above three solutions were prepared by dissolution of their salts recrystallised from hot water. Procedure The reaction was started by mixing, with magnetic stirring, 1 cm3 of H2tspp solution with 50 cm3 of a solution containing lead@), manganese(II), hydroxylamine, borate buffer and sodium nitrate. The rate of the formation of the complex of manganese(I1) with H2tspp was monitored spectrophoto- metrically by recording the disappearance of free H2tspp at 413 nm (Amax. of Hztspp) as a function of the reaction time. The absorbance was measured in a 10-mm thermostated cell. All the experiments were carried out at 25 "C and an ionic strength of 0.1 with NaN03. Results and Discussion Catalytic Effect of Lead(I1) on Formation of Mnn(tspp) The reaction of H2tspp with manganese(I1) was studied in the presence of lead concentrations from 10-7 to 4 X 10-6 rnol dm-3 and at pH 4-8.7.The kinetic equation for the reaction is given by equation (l), where ko is the conditional rate constant involving concentrations of hydrogen ions, lead(I1) and manganese(I1). -d[H2tspp]ldt = k,[H2tspp] . . . . (1) The conditional rate constant was determined from the slope of the first-order rate graph. In Fig. 1, ko is plotted against pH. The rate of formation of MnIItspp in the absence of lead(I1) is independent of pH. However, the rate constant in the presence of lead(I1) does depend on pH. Lead(I1) accelerates the rate of formation of MnrI(tspp) at pH >6. The conditional rate constant, ko, increases with pH and then levels off.The formation of MnII- (tspp) is first-order with respect to lead(I1). For the cad- mium(I1)-catalysed incorporation of manganese(I1) into142 ANALYST, FEBRUARY 1987, VOL. 112 2 I B m z 1 6 a - c - v U 6 7 8 9 -Log [H'] Fig. 1. Effect of pH on the incorporation of man anese(I1) into and Bfthe absence of %:&) at 25 "C and I = 0.1 (NaN03). CHztSp = 1. 4 5 X 10-6; C,, = 6.23 X 10-4; and Cm,oH = 9.80 X 10-3 mof dm-3 in (A) the presence (Cpb = 4.20 X H,tspp, we previously postulated the following mechan- ism,13J4 which is also suggested for the lead(I1)-catalysed reaction: kl k-2 Mn2+ + H2tspp - MnII(tspp) + 2H+ Pb2+ + H2tspp k2, PbII(tspp) + 2H+ PbII (tspp) + Mn2+ 3 MnII(tspp) + Pb2+ Scheme 1 Assuming a steady-state approximation for PbII(tspp), we can assume that equation (2) represents the lead-catalysed reaction.A graph of kpb-l versus [H+]2 gives k2 as the intercept and k-2/k3 from the slope. kpb = k2k3[Mn2+][Pb2+]/(k-2[H+]2 + k3[Mn2+]) (2) At higher pHs, k-2[H+]2 << k3[Mn2+] and lead(I1) hydrolyses (Kh = 10-7.78).15 Thus equation (2) can be rewritten kpb = kZ[Pb2+] + k4[PbOH+] . . . . (3) where k4 is the reaction path for PbOH+ + H2tspp$ PbII(tspp) + H+ The rate constants for pH 7.8-8.7 were independent of [Mn2+] over a concentration range 1 x 10-4-10-3 mol dm-3, as expected from equation (3). Using the equilibrium constant of the formation of PbII(tspp), ([PbII(tsps)] [H+]2/ [Pb2+][H2tspp] = 3.38 x 10-10),16 the rate constant for each reaction step in Scheme 1 were determined as follows: kl = (2.33 k 0.11) x 10-2 mol-1 dm3 s-1, k2 = (2.16 rfr 0.08) X 102 mol-1 dm3 s-1, k-2 = (6.39 rfr 0.19) x 10-1' mol-2 dm6 s-1, k3 = (1.67 f 0.11) x 102mol-1 dm3 s-1 and k4 = (7.33 k 0.41) X 102 mol-1 dm3 s-1.The catalytic effect of lead(I1) can be explained by the reaction proceeding through PbII(tspp) as shown in Scheme 1. Because of the large ionic radius of lead(I1) (132 pm), the coordination of lead(I1) will give a highly expanded or distorted porphyrinato core, which facilitates the attack of manganese(I1) on the porphyrin nucleus from the back.SJ4J7 PbII(tspp) is 7200 times as reactive as H2tspp in catalysing the incorporation of manganese(I1). Lead(I1) liberated after the incorporation of manganese(I1) into the porphyrin reacts with the free base porphyrin and acts as a catalyst. At pH B7.8, the formation of PbII(tspp) is the rate-determining step. It was revealed by an independent experiment that manganese in the product is in the trivalent state and that the rate of oxidation of Table 1.Masking of cadmium(I1) by cyanide for the determination of lead(I1) using the proposed method at 25 "C and at I = 0.1 (NaN03). C H ~ ~ ~ ~ = 2.07 X Cm20H = 1.02 X lo-* and CNaCN = 1.04 X 10-2 moi dm-3 C,, = 6.23 X Pb" added CdI* added PblI found 10-6mol dm-3 10-6 mol dm-3 10-6mol dm-3 Error, % 0.64 0 0.62 -3.1 1.99 0.61 -4.7 3.98 0.60 -6.3 5.98 0.63 -1.5 7.97 0.59 -7.8 9.96 0.62 -3.1 1.28 0 1.24 -3.1 1.99 1.33 +3.9 3.98 1.25 -2.3 5.98 1.28 0 7.97 1.34 +4.6 9.96 1.20 -6.3 1.93 0 1.98 +2.6 1.99 1.86 -3.6 3.98 1.79 -7.3 5.98 2.01 +4.1 7.97 2.04 +5.7 9.96 2.03 +5.2 MnII(tspp) to MnIII(tspp) is rapid compared to the formation of the complex.16 Catalytic Determination of Lead(II) Calibration graph At pH 8-8.7, the observed rate constant is dependent only on the lead(I1) concentration and therefore allows the determina- tion of lead(I1) by means of the catalytic reaction. A net decrease in absorbance (AA = A(o> - A pb)) at a fixed time is also related to the concentration of lead(II), where Ap) and A(pb) denote the absorbances of solutions in the absence and in the presence of lead(I1) , respectively. Molar absorptivities and Sandell's sensitivities18 of the calibration graphs are: A, 2.75 x 105,0.753;B, 1.42 x 105,1.46;andC,9.05 x 104,2.28,at fixed reaction times of (A) 3, (B) 6 and (C) 15 min (units of molar absorptivity and Sandell's sensitivity are mol-1 dm3 cm-1 and ng cm-3, respectively).Masking of cadmium(I4 by cyanide Mercury(I1) and cadmium(I1) accelerate the formation of MnII(tspp) in the same way as lead(II), but the interference from mercury(I1) was avoided by the addition of hydroxylam- ine as a reducing agent. Cadmium(I1) concentrations compar- able to the concentration of lead(I1) assist the formation of MnII(tspp). Therefore, cyanide was used to mask cad- mium(I1). Some kinetic runs show no statistical difference in the rate of MnII(tspp) formation in the presence and absence of 0.01 mol dm-3 of cyanide. Table 1 shows the results of masking up to 1 x 10-5 mol dm-3 of cadmium(I1) with cyanide. Effect of foreign ions Table 2 shows the effect of various substances usually encountered in the determination of lead(I1) in natural waters.Ca2+, Mg2+, Sr2+, Ba2+ and Mn2+ do not interfere in the determination of lead(I1). Iron(II1) is reduced to iron(I1) with hydrolxylamine and forms Fe(CN)64- in the presence of cyanide. The iron(I1) complex suppressed the catalytic effect of lead(I1) owing to the formation of Pb2[Fe(CN),]. Silicate also reduced the catalytic effect of lead(I1). Interferences from these ions were removed by solvent extraction using n-decan- oic acid in chloroform.19y20ANALYST, FEBRUARY 1987, VOL. 112 143 Table 2. Effect of foreign ions on the determination of lead(I1) (1.26 X 10-6 mol dm-3) using the proposed method at 25 “C and at Z = 0.1 (NaN03). CH2fSPP = 2.05 x 10-6, CMn = 6.23 x and CNH20H = 1.02 X 10-2 mol dm-3 Ions Addedmol dm-3 Error, ‘/o Ca2+ .. . . 1.0 x 10-4 -5.7 Sr2+ . . . . 1.0 x 10-4 +3.3 Ba2+ . . . . 1.0 x 10-4 +1.0 Mg2+ . . . . 1.0 x 10-4 -1.5 Mn2+ . . . . 1.2 x 10-4 -1.7 co2+ . . . . 1.0 x 10-5 +2.3 Ni2+ . . . . 1.0 x 10-5 -4.0 cu2+ . . . * 1.0 x 10-6 -20.0 1.0 x 10-5 +2.0* Zn2+ . . . . 1.0 x 10-5 +2.0 Hg2+ . . . . 1.0 x 10-6 -0.1 Fe3+ . . . . 1.0 x 10-6 -54.0 AP+ . . . . 1.0 x 10-4 +4.1 1.0 x 10-5 -6.3T NH4+ . . . . 2.0 x 10-4 0.0 c1- . . . . 1.0 x 10-3 +5.5 Br- . . . . 1.0 x 10-3 +1.1 I- , . . . . . 1.0 x 10-3 -5.3 s~03~- . . . . 2.0 x 10-6 -1.0 SCN- . . . . 1.0 x 10-4 -2.0 CN- . . . . 1.0 x 10-2 +4.3 SQ2- . . . , 1.0 x 10-4 -32.0 1.0 x 10-4 -5.3t * Cyanide (1.0 x 10-2 mol dm-3) was added to remove the j- Lead(I1) was extracted by n-decanoic acid before the interference. determination.Table 3. Determination of lead(I1) in rain and drain water by the proposed method Sample* PbVpg dm-3 Rainwater . . . . Drainwater . . . . Correction Added Found for addition 0 33 66 100 0 80 160 239 3 38 70 107 30 109 191 272 3 5 4 7 Mean = 5 30 29 31 33 Mean = 31 * As the concentration of lead(I1) in the sample was low, 250 cm3 was taken and concentrated to 50 cm3 on a hot-plate. Separation of iron and silicate by solvent extraction A 50-cm3 aliquot of sample solution was taken in a 100-cm3 separating funnel and 1 cm3 of 0.1 mol dm-3 nitric acid was added to adjust the pH to 2.7-3.0. Iron(II1) was extracted twice with 10 cm3 of 1 mol dm-3 n-decanoic acid in chloroform. A 3-cm3 aliquot of 0.1 mol dm-3 sodium acetate and 10 cm3 of 1 molcm-3 n-decanoic acid were added to the aqueous solution (pH 5.0).Lead(I1) was extracted into chloroform. The organic phase was washed twice with 10 cm3 of water containing 10-3 mol dm-3 nitric acid; lead(I1) was then back-extracted to the aqueous phase. The aqueous solution was successively extracted with chloroform and toluene in order to remove the remaining n-decanoic acid in the aqueous phase. The solution was then analysed by the proposed method. In Table 2, results obtained with and without solvent extraction for iron and silicate are shown. A ten-fold excess of iron and 100-fold excess of silicate were completely removed by the solvent extraction method. Catalytic Determination of Lead(I1) in Water Samples The proposed kinetic method was applied to the analysis of rain and drain water.The samples were slightly acidifed (pH 3) with nitric acid soon after collection. As the concentration of lead(I1) in the sample was low, 250 cm3 of sample were taken and concentrated to 50 cm3 on a hot-plate. The results are shown in Table 3. It was shown to be possible to determine p.p.b. levels of lead(I1) by the method. Lead was not determined in these samples by atomic absorption spec- trometry because of the high background noise at the low concentrations of lead present, Sensitivity, Accuracy and Precision The proposed kinetic method allows the determination of small amounts of lead(I1). The lead(I1) catalysed metallo- porphyrin formation is 7200 times as fast as the reaction in the absence of lead(I1).If the limit of determination of the catalyst is defined as the concentration giving a rate constant equal to the blank reaction, the sensitivity of the catalytic reaction is 8.7 X 10-8 mol dm-3. The standard deviation for the observed first-order rate constant was +3.6%, corresponding to an error of k5.5 X 10-8 mol dm-3 in the determination of lead(I1). The dispersion of the data in the calibration graph also shows the precision of the method. The standard deviations are k4.9 X 10-8 mol dm-3 Pb”, k5.7 x 10-8 mol dm-3 PbII and 35.5 X 10-8 mol dm-3 PbII at reaction times of 3, 6 and 15 min, respectively. In this method the reaction must be recorded for a given time period. However, the method allows the spectropho- tometric determination of lead(I1) at concentrations as low as lo-’ mol dm-3 on the basis of a clearly defined chemical reaction.The experimental conditions can also easily be set from a knowledge of the rate and equilibrium constants. The author thanks Professor Motoharu Tanaka, Nagoya University, for helpful discussions and gratefully acknow- ledges the financial support of this study by a grant from the Japanese Ministry of Education, Science and Culture (61540450). References 1. Yonehara, N., and Kawashima, T., Bunseki, 1983, 418, and references therein. 2. Mottola, H. A., and Mark, H. B., Jr., Anal. Chem., 1986,58, 264R. 3. Yotsuyanagi, T., Hoshino, H., and Igarashi, S., Bunseki, 1985, 496, and references therein. 4. Lavallee, D. K., Coord.Chem. Rev., 1985, 61, 55. 5. Tanaka, M., Pure Appl. Chem., 1983, 55, 151. 6. Margerum, D. W., Caylay, G. R., Weatherburn, D. C., and Pagenkopf, G. K., in Martell, A. E., Editor, “Coordination Chemistry,” Volume 2, American Chemical Society, Washing- ton, DC, 1978, p. 1. 7. Itoh, J., Yotsuyanagi, T., and Aomura, K., Anal. Chim. Acta, 1975, 74, 53. 8. Ishii, H., and Koh, H., Mikrochim. Acta, 1983, I, 279. 9. Corsini, A., DiFruscia, R., and Herrmann, O., Talanta, 1985, 32,791. 10. Tabata, M., Tanaka, M., Anal. Lett., 1980, 13, 427. 11. Adler, A. D., Longo, F. R., Finarelli, J. D., Goldmacher, J., Assour, J., and Korsakoff, L., J. Org., Chem., 1967, 32, 476. 12. Fleisher, E. B., Palmer, J. M., Srivastava, T. S., and Chatterjee, A., J. Am. Chem. SOC., 1971, 93, 3162. 13. Tabata, M., and Tanaka, M., Mikrochim. Acta, 1982,II, 149. 14. Tabata, M., andTanaka, M., J. Chem. SOC., Chem. Commun., 1985, 42. 15. Smith, R. M., and Martell, A. E., “Critical Stability Con- stants,” Volume 4, Plenum Press, New York, 1976, p. 1. 16. Tabata, M., and Tanaka, M., J. Chem. SOC. Dalton Trans., 1983, 1955.144 ANALYST, FEBRUARY 1987, VOL. 112 17. Barkigia, K. M., Fajer, J., Adler, A. D., and Williams, G. J. B., Inorg. Chem., 1980, 19,2057. 18. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience, New York, 1965, p. 80. 19. Tanaka, M., Nakasuka, N., and Goto, S., in Dryssen, D . , Liljenzin, J.-O., and Rydberg, J., Editors, “Solvent Extraction Chemistry,” North-Holland, Amsterdam, 1967, p. 154. 20. Nakasuka, N., Nakai, M., and Tanaka, M., J . Inorg. Nucl. Chem., 1970, 32, 3667. Paper A611 75 Received June 2nd, 1986 Accepted September 25th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200141

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1987, VOL. 112 145 Determination of Hydroxyl Value in Fats and Oils Using an Acid Cat a I yst Leopold Hartman and Regina C. A. Lago National Research Centre of Agricultural and Food Technology, EMBRAPA, Av. das Americas, 29.50 I , Guaratiba, Rio de Janeiro, RJ, 23020, Brazil and Laerte C. Azeredo and Maria A. A. Azeredo Chemistry Department, Federal Rural University of Rio de Janeiro, Itaguai, RJ, 23.460, Brazil A method for the rapid determination of the hydroxyl value of oils, fats and related products is described. The method uses toluene-p-sulphonic acid as a catalyst and this allows the sample, dissolved in toluene, to be acetylated in 10 min at room temperature. The fatty acid anhydrides formed during the acetylation are decomposed with aqueous sodium hydroxide and tert-butanol, and are then titrated with hydrochloric acid. The addition of an aliquot of the sample to the blank after the decomposition of the acetic anhydride reduces the number of titrations required to two.Keywords: Hydroxyl value determination; fats and oils; toluene-p-sulphonic acid; acetic anhydride; acid catalysis The hydroxyl value, expressed in milligrams of potassium hydroxide and corresponding to the number of hydroxyl groups present in 1 g of a sample, is one of the traditional characteristics of oils and fats. The official methods of determination recommend heating the sample with acetic anhydride and pyridine at 100 "C, then decomposing the excess of anhydride by heating with water and titrating with an ethanolic solution of potassium hydroxide.1-3 This recommen- ded method requires, in addition to the titration of the sample and the blank, two titrations to determine the acid value of the sample and to verify the molarity of the ethanolic alkali. According to O'Connor,4 this determination, in duplicate, takes at least 3 h. The use of near infrared spectroscopy has also been proposed for the determination of hydroxyl groups in oils and fat~~5.6 but this method requires 2.5 h for completion4 together with special equipment, and even small amounts of moisture interfere with the determination. It has been shown by Conant and Bramann7 that perchloric acid in glacial acetic acid catalyses the acetylation more rapidly than pyridine, and subsequently other acids and solvents have been suggested for the determination of hydroxyl values.Sl1 Based on this work, Fritz and Schenkl2 developed a method of acetylating primary and secondary alcohols with acetic anhy- dride in ethyl acetate solution in the presence of 0.15 M perchloric acid, and this was further developed by Panaiotova et al.l3 It seems surprising, therefore, that acid catalysts have not as yet been applied to the determination of the hydroxyl value of oils and fats. In the work reported here, perchloric acid was not investigated, owing to its unpredictable reaction with glycerides under anhydrous conditions. After prelimi- nary tests with several fats and acids, toluene-p-sulphonic acid was selected as a catalyst, and the method of Fritz and Schenkl2 was followed. Various results were obtained depending on the acid value of the samples.Oils with a low acidity gave hydroxyl values approaching those obtained using a standard method,3J4 but with increasing acidity the values increased out of proportion. Thus commercial oleic acid, which gave a hydroxyl value of 4.6 using the standard method, gave a value nearly 20 times higher when an acid catalyst was used. It was found that the fatty acid anhydrides formed during the acetylation resisted the decomposition with aqueous pyridine specified by Fritz and Schenk,12 thus giving rise to elevated hydroxyl values. This was probably the reason for not applying acid catalysts to the determination of the hydroxyl values of oils and fats, although no indication of this problem could be found in the literature.Experimental Development of the Method In order to overcome the difficulty of decomposing the fatty acid anhydrides formed during the acetylation, an excess of aqueous sodium hydroxide was added and ethyl acetate was replaced by toluene in the preparation of the acetylation reagent to avoid the saponification of the solvent. This had some effect, but failed to achieve the complete decomposition of the fatty acid anhydrides dissolved in the toluene phase. The replacement of toluene by water-soluble solvents such as dioxane or tetrahydrofuran was tried without success, and the addition of primary or secondary alcohols was obviously out of the question as they would be esterified by the fatty acid anhydrides present , thus unduly increasing the hydroxyl value. tert-Butanol was finally chosen, as although it is readily esterified by anhydrides in an acidic medium, it resists esterification in an alkaline medium owing to steric hindrance, and facilitates the decomposition of the fatty acid anhydrides because of its solubility in both water and toluene.The excess of alkali can be subsequently titrated with aqueous hydro- chloric acid. Apparatus Conventional laboratory glassware, such as Erlenmeyer and calibrated flasks, beakers, pipettes and burettes, were used, together with an analytical balance. Materials The following oils and fats were used in the investigation: commercial samples of crude cottonseed, groundnut, soya bean, rapeseed, olive, palm, crude, hydrogenated and dehy- drated castor oils, lard, beef fat and oleic acid; laboratory- extracted avocado oil, butter and chicken fats; monodecanoyl (caproyl) glycerol and monopalmitoyl glycerol produced by a modified Emil Fischer procedure15; and 12-hydroxystearic acid produced from hydrogenated castor oil.Ethyl esters were obtained by the trans-esterification of soya bean oil with anhydrous ethanol. Commercial ethyl acetate, octanol and cetyl alcohol were also used.146 ANALYST, FEBRUARY 1987, VOL. 112 Table 1. Amounts of sample and reagent to be used Table 2. Hydroxyl values of various oils and fats determined by the standard and modified methods Expected h ydroxyl value 0-100 100-200 200-300 300400 400-600 Mass of sample dissolved in 25 ml of toluene/g 5 3.5 2.5 1.5 1 Volume of sample solution Volume usedml reagendml 5 5 5 5 5 5 5 5 5 5 Reagents and Solutions Acetylation reagent.Toluene-p-sulphonic acid, acetic anhy- dride and toluene mixed in the proportions 1 g : 5 ml : 15 ml, or multiples thereof. This solution, when stored in a dark bottle in a refrigerator, remains stable for two weeks. Sodium hydroxide solution. Carbonate-free, ca. 1.3 M, containing 10% mlV sodium sulphate. Hydrochloric acid. Standardised, ca. 0.5 M. Toluene. Laboratory grade. tert- Butanol. Phenolphthalein solution. 170 in ethanol. Procedure Samples soluble in toluene A 1-5 g mass of sample, the mass depending on the expected hydroxyl value (see Table l), is weighed into a 25-ml calibrated flask and dissolved in toluene, diluting to volume with this solvent. A 5-ml aliquot of the above solution is pipetted into a 200-ml Erlenmeyer flask, followed by 5.0 ml of the acetylating reagent added by pipette or burette.The mixture is shaken and left for 10 min. Sodium hydroxide solution (1.3 M, 25 ml) containing sodium sulphate is pipetted into the flask, the flask is shaken and 10 ml of tert-butanol are added with shaking, via a measuring cylinder. After 1 min, the excess alkali is titrated with 0.5 M hydrochloric acid in the presence of 0.5 ml of phenolphthalein indicator. The blank determination is carried out by adding 5 ml of the acetylating reagent to a 200-ml Erlenmeyer flask by pipette or burette, then adding 25 ml of the 1.3 M sodium hydroxide solution by pipette and 10 ml of tert-butanol with a measuring cylinder, with shaking. After 1 min, 5 ml of the original sample solution in toluene are added, and the resulting solution is shaken and titrated with 0.5 M hydrochloric acid in the presence of phenolphthalein. For samples with a low hydroxyl value (20 or less), 10 ml of the original sample solution in toluene, instead of 5 ml, can be used for both sample and blank determination.Samples soluble with difficulty in toluene A 1-5 g mass of sample (e.g., 12-hydroxystearic acid) is weighed in a 25-ml calibrated flask, dissolved in 10 ml of glacial acetic acid, if necessary with slight warming, and the volume completed with toluene. The procedure then follows that described under Samples soluble in toluene, with the difference that 50 ml of sodium hydroxide solution and 20 ml of tert-butanol are used instead of 25 ml and 10 ml, respectively.The hydroxyl value is calculated according to the equation (S - B ) X A4 X 56.1 H.V. = m where S and B denote the volume of hydrochloric acid used in the titration of the sample and blank, respectively, M the molarity of the hydrochloric acid and rn the mass in grams of the sample contained in the aliquot of toluene solution used. Hydroxyl value Oil Standard method3.14 Modified method Cottonseed . . . . . . Groundnut . . . . . . Soyabean . . . . . . Rapeseed . . . . . . Avocado . . . . . . Olive . . . . . . . . Palm . . . . . . . . Castor . . . . . . . . Hydrogenated castor . . Dehydratedcastor . . . . Beef fat . . . . . . . . Lard . . . . . . . . Butterfat . . . . . . Chickenfat . . . . . . 10.3 3.8 1.8 17.0 31.8 21.1 20.4 166.8 153.3 33.3 3.4 3.6 5.5 7.2 9.2 4.0 1.5 16.4 30.6 12.7 19.4 166.4 152.8 34.2 3.1 3.8 6.1 8.1 Results and Discussion The hydroxyl value, apart from providing one of the common characteristics of oils and fats, is used in the evaluation of the quality of methyl and ethyl esters obtained by the trans- esterification of vegetable oils for use as fuels.It is sometimes used to assess the efficiency of the dehydration process for castor oil. In the work reported here, samples of various vegetable and animal oils and fats and related products were used to test the applicability of the method developed. The authenticity of the commercial samples was confirmed by the determination of their fatty acid composition by gas - liquid chromatography and that of the samples produced in the laboratory by the determination of their chemical characteristics.The labora- tory-produced 12-hydroxystearic acid was found to be chemi- cally pure and the monoacylglycerols were of 98-9970 purity. The use of 1.3 M sodium hydroxide solution containing 10% sodium sulphate and tert-butanol allowed the determination of a sharp end-point in the titration with hydrochloric acid, owing to the rapid separation of the aqueous and toluene phases. Table 2 shows the results of the determination of the hydroxyl value (in duplicate) of various vegetable and animal oils and fats using a standard method3J4 and the modified procedure, and Table 3 shows the corresponding results for other lipids and related products. The difference between the results obtained by the two methods was in most instances less than one unit.The results obtained for monoacylglycerols and castor oil show that the modified method determines both primary and secondary hydroxyl groups equally well. Five determinations of the hydroxyl value were carried out for castor and groundnut oil in order to assess the reproduci- bility of results when using oils with high and low hydroxyl values, respectively. The results and their statistical evalua- tion are given in Table 4 and show that the standard deviations for the two oils were 0.48 and 0.37, respectively. These deviations are inside the limits reported for the German standard method16 and are less than those stipulated in the American Oil Chemists’ Society’s method,17 according to which two single determinations performed in one laboratory should not differ by more than 2.4 units.The use of an excess of sodium hydroxide to decompose the anhydrides formed from the free fatty acids of the oils examined involved the risk of a partial saponification of the glycerides. The prevention of such a saponification by substituting sodium hydroxide with sodium carbonate was tried, but was found to be impractical, owing to the difficulty of obtaining a satisfactory end-point in the titration with hydrochloric acid. It was found, however, that samplesANALYST, FEBRUARY 1987, VOL. 112 147 ~~ Table 3. Hydroxyl values of various acids, alcohols and esters Hydroxyl value Compound Standard method3914 Modified method 12-Hydroxystearic acid . . 186.9 186.2 Oleicacid . . . . . . 4.8 4.1 Octanol .. . . . . . . 428.9 427.5 Cetylalcohol . . . . . . 231.4 231.8 Monodecanoyl (caproyl) glycerol . . . . . . 514.5 513.6 Monopalmitoyl glycerol . . 339.2 341.4 Ethylacetate . . . . . . 0 0.5 Ethyl esters from soya beanoil . . . . . . 2.7 2.2 Table 4. Standard deviations and coefficients of variation for 5 hydroxyl value determinations of castor and groundnut oil Oil Number of determination Castor Groundnut 1 166.5 2 166.3 3 167.5 4 166.8 5 166.4 Average . . . . . . . . . . 166.7 Standard deviation . . . . . . 0.48 Coefficientofvariation . . . . 0.28% 3.5 4.4 3.5 3.7 4.5 3.9 0.37 9.51% titrated immediately after the addition of sodium hydroxide and samples titrated after 10 min standing time required the same amount of hydrochloric acid for neutralisation. This was probably due to the fact that the oils dissolved in the toluene phase escaped saponification.The advantages of this modified method compared with existing methods is primarily the economy of time and work. Whereas standard methods require, as mentioned before, about 3 h for a duplicate determination, the modified procedure requires only 30 min, if the blank determination is performed during the 10 min required for the acetylation of the sample. The four titrations required in the existing methods are reduced to only two, because the addition of a sample aliquot to the blank removes the need for the acid value determination and there is no need to check the molarity of the standardised hydrochloric acid. The fact that the irritant pyridine is not used is an additional advantage.Burton and Praill18 suggested that the rapid acylation of hydroxy compounds with acetic anhydride at room tempera- ture in the presence of perchloric acid is due to the formation of the “acetylium” cation, CH3CO+, according to the follow- ing reactions: (CH3CO)zO + H+ + (CH3C0)20H+ (CH3C0)20H+ CH3CO+ + CH3COOH CH3CO+ + ROH + CH3COOR + H+ The regenerated hydrogen ion would then produce a further supply of acetylium ions, hence re-initiating the reaction. On the basis of the results obtained in this work, toluene-p- sulphonic acid seems to give rise to a similar reaction path. Conclusion The procedure described, which makes use of toluene- p-sulphonic acid instead of pyridine as the catalyst, allows the determination of the hydroxyl value of oils, fats and related products at room temperature in a much shorter time than existing methods.It reduces the number of titrations required from four to two, without impairing the reproducibility of the results and can be carried out using conventional laboratory equipment. The authors express their thanks to Mr. E. P. M. Sarmento, General Manager, and to Mr. E. Sundfeld, Assistant Tech- nical Manager, of the National Research Centre of Agricul- tural and Food Technology, EMBRAPA, for their interest and encouragement during this investigation. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Mehlenbacher, V. C., “The Analysis of Fats and Oils,” Garrard Press Publishers, Champaign, IL, 1960, p. 490. “British Standard Methods of Analysis of Fats and Fatty Oils,” British Standard BS 684, 1982, Section 2.9. IUPAC, “Standard Methods for the Analysis of Oils, Fats and Derivatives,” Pergamon Press, Oxford, 1979, Section 2.241. O’Connor, R. T., J. Am. Oil. Chem. SOC., 1961,38, 641. Bayzer, H., Schauenstein, E., and Winsauer, K., Monatsh. Chem., 1958, 89, 15. Hilton, C. L., Anal. Chem., 1959, 31, 1610. Conant, J. B., and Bramann, G. M., J. Am. Chem. SOC., 1928, 50,2305. Toennies, G., Kolb, J. J., and Sakami, W., J. Biol. Chem., 1942, 144, 193. Pesez, M., Bull. SOC. Chim. Fr., 1954, 1237. Erdos, J. B., and Bogati, A. G., Rev. SOC. Quim. Mex., 1957, 1, 223. Mesnard, P., and Bertucat, M., Bull. SOC. Chim. Fr., 1959, 307. Fritz, J. S., and Schenk, G. H., Anal. Chem., 1959, 31, 1808. Panaiotova, E. N., Mincheva, M. F., and Dimitrov, D., Khim. Znd. Sofia, 1970, 6 , 251. Cocks, L. V., and Rede, C. V., “Laboratory Handbook for Oil and Fats Analysis,” Academic Press, London, 1966, p. 120. Hartman, L., Azeredo, L. C., and Szpiz, R. R., J. Am. Oil Chem. SOC., 1984,61,963. Pardun, H., “Analyse der Nahrungsfette,” Paul Parey, Berlin, 1976, p. 89. “Official and Tentative Methods of the American Oil Chemists’ Society,” American Oil Chemists’ Society, Champaign, IL, 1973, Method Cd 13-60. Burton, H., and Praill, P. F. G., J. Chem. SOC., 1950, 1203. Paper A61124 Received April 21st, 1986 Accepted July 31st, 1986

ISSN:0003-2654    

DOI:10.1039/AN9871200145

出版商:RSC    

年代:1987

数据来源: RSC