ISSN:0003-2654    

DOI:10.1039/AN97601FX041

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

ANALAO 101 (1 208) 833-920 (I 976)I SS N 0003-2654November 1976THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYC 0 NTE IUTSORIGINAL PAPERSAn Assessment of Various Types o f Reference Electrode f o r Use in ContinuousPotentiometric Analysis with Particular Application to Highly PureWaters-D. Midgley and K. Torrance848 Factors Affecting the Limit of Detection of Sodium-responsive Glass Electrodes-G. I. Goodfellow, D. Midgley and H. M. Webber856 Colorimetric Method for the Determination of Arsenic(l1l) in Potable Water-Shingara S. Sandhu860 Rapid Spectrophotometric Determination o f Cobalt in High-speed SteelsBased on the Formation o f Tricarbonatoeobaltate( 111)-A. Sanz Medel,A. Cob0 Guzmi5n and J. A. Pirrez-Bustamante867 Micro-determination o f Pyrrole Derivatives with N-Bromosuccinimide inAcetic Acid Medium-J. P.Sharma, V. K. S. Shukla and A. K. Dubey870 Losses o f Trace Metals During the Ashing o f Biological Materials-S. R. Koirtyo-hann and Carole A. Hopkins876 Determination o f Triamcinolone Acetonide in Cream and Suspension For mu-lations by High-performance Liquid Chromatogaaphy-G. G.;,don and P. R.WoodQuantitative Determination of Ethoxyquin in Apples by Gas Chromatography-Bjorn WinellDetermination o f Ozone by Gas Chromatography-R. V. Holland and P. W. Board833883887892 Appraisal o f the Determination of Molecular Size and Mass by Membrane901 X-ray Analysis o f Particulate Matter Collected in an Environmental MonitoringDiffusion-Peter E. O'Connor and Michael G. HarringtonDevice-G. C. S. Collins and D. NicholasCO M M U N ICATIO N91 2 Determination of Arsenic(ll1) and Total Arsenic by the Silver Diethyldithio-914 Book Reviews920 Erratum920 Notice t o Authorscarbamate Method-J. Aggett and A. C. AspellSummaries o f Papers in this Issue-Pages iv, v, viii, ix, xPrinted by Heffers Printers Ltd, Cambridge, EnglandEntered as Second Class at New York, USA, Post Offic

ISSN:0003-2654    

DOI:10.1039/AN97601BX043

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

1v SUMMARIES OF PAPERS I N THIS ISSUE November, 1976Summaries of Papers in this IssueAn Assessment of Various Types of Reference Electrode for Usein Continuous Potentiometric Analysis with ParticularApplication to Highly Pure WatersThe suitability of a number of reference electrodes for use in the continuouspotentiometric analysis of highly pure water has been assessed. The electrodeswere chosen to be as representative as possible of their class, but severalelectrodes of novel design were also .included. The electrodes were testedfor the constancy of their potentials over 100 d of continuous operation in10-6 moll-1 sodium hydroxide solution (a simple model boiler water) andfor the extent to which they were affected by ultraviolet light, particulatematter and changes in flow-rate and ionic strength.The times taken by theelectrodes to recover from changes in temperature, from an interruption in theflow of sample and from being immersed in buffer solutions were measuredand proved to be important characteristics in discriminating between theover-all performance of electrodes.It was concluded that conventional types of electrode, whether calomelor silver - silver chloride, with ceramic frit junctions and unsaturated potas-sium chloride electrolytes, were superior in almost every respect to the othertypes tested.D. MIDGLEY and K. TORRANCECentral Electricity Research Laboratories, ]Kelvin Avenue, Leatherhead, Surrey,KT22 7SE.Analyst, 1976, 101, 833-847.Factors Affecting the Limit of Detection of Sodium-responsiveGlass ElectrodesFactors affecting the limit of detection of sodium-responsive glass electrodeshave been examined and it is postulated that the ultimate limit is set bythe dissolution of alkali metal ions from the glass membrane itself.Experi-mental results are presented that are consistent with this hypothesis and fromwhich the following conclusions can be drawn: (a) measurements of very lowsodium concentrations are favoured by high linear flow-rates of solution pastthe membrane surface and by having as low a sample temperature as possible,consistent with acceptable response times ; (b) hydrogen-ion interference isnegligible provided that the pH is greater than 10.5; (c) interference of ionsfrom the alkaline additive is similarly negligible if an alkylamine such asdiethylamine is used.The limit of detection obtained with EIL GEA.33 electrodes in an EILflow cell with a flow-rate of 4mlmin-l was about 0.07 pg1-l of sodiuma t 20 "C, when the pH was adjusted to 11.0 with diethylamine.G.I. GOODFELLOW, D. MIDGLEY and H. M. WEBBERCentral Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey,KT22 7SE.Analyst, 1976, 101, 848-855November, 1976 SUMMARIES OF PAPERS IN THIS ISSUEColorimetric Method for the Determination of Arsenic(II1) in .Potable WaterVA colorinictric method for the detcrniination oi' arsenic(ll1) a t a concentrationof 0.002 mg 1-1 in potable water was developed. Arsenic(II1) reacts quanti-tatively with potassium iodate in the presence of sulphuric acid and releasesan equivalent amount of iodine, which imparts a pink colour to a carbontetrachloride extract, this colour being measured at 520 nm.The method wasused to differentiate between arsenic(II1) and arsenic(V). No interferencefrom arsenic(V), chloride, fluoride or nitrate was observed. Natural watersamples analysed by this technique showed trace amounts of arsenic(II1).SHINGARA S. SANDHUWater Laboratory, South Carolina State College, Orangeburg, South Carolina '291 17,USA.Analyst, 1976, 101, 856-859.Rapid Spectrophotometric Determination of Cobalt in High-speedSteels Based on the Formation of Tricarbonatocobaltate( 111)The influence of a number of experimental parameters on the determinationof cobalt in high-speed steels has been investigated, taking advantage ofthe property of the cobalt(II1) ion of forming very selectively a stablegreen-coloured complex compound with carbonate. As a result, a methodhas been devised that has been applied to the analysis of six certified standardsteel samples with cobalt contents ranging from 3 to 12%.From a statisticaltreatment of the results it is concluded that the average relative error of themethod lies within the range &1.37%, while its relative standard deviationis 0.67%, thereby indicating the absence of systematic errors.A. SANZ MEDEL, A. COB0 GUZMAN and J. A. PRREZ-BUSTAMANTEDepartamento de Quimica Analitica, Facultad de Ciencias Quimicas y CSIC, Uni-versidad Complutense, Ciudad Universitaria, Madrid-3, Spain.Analyst, 1976, 101, 860-866.Micro - determination of Pyrrole Derivatives withN-Bromosuccinimide in Acetic Acid MediumA simple micro-procedure for the determination of pyrrole derivatives withN-bromosuccinimide is described. A 2-10-mg sample dissolved in acetic acidis subjected t o reaction with a known excess of N-bromosuccinimide at roomtemperature and the excess of reagent is back-titrated iodimetrically. Themaximum error in the results is &1.44%.J. P. SHARMA, V. K. S. SHUKLA and A. K. DUBEYDepartment of Chemistry, University of Allahabad, Allahabad-2, India.Analyst, 1976, 101, 867-869

ISSN:0003-2654    

DOI:10.1039/AN97601FP085

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

...Vlll SUMMARIES OF PAPERS I N THIS ISSUE November, 1976*Losses of Trace Metals During the Ashing of Biological MaterialsLosses of chromium, iron, zinc and cadmium during ashing were studied usingtissues that contained endogenously incorporated radioisotopes. No losseson drying a t 110 "C and no volatility losses were detected for any of theelements a t temperatures below 600 "C. Chromium was lost from blood butnot from liver samples heated a t 700 "C. Losses as insoluble material on thecrucible surfaces were more significant. Up to 42% of the isotope wasretained on the dish after dissolution with acid, the amount depending on theashing temperature, crucible material and surface condition. No evidence forthe formation of volatile compounds from the endogenously incorporated iso-tope was found.S .R. KOIRTYOHANNEnvironmental Trace Substances Research Center and Department of Biochemistry,University of Missouri, Columbia, Mo. 65201, USA.and CAROLE A. HOPKINSDepartment of Biochemistry, University of Missouri, Columbia, Mo. 65201, USA.Analyst, 1976, 101, 870-875.Determination of Triamcinolone Acetonide in Cream and SuspensionFormulations by High- performance Liquid ChromatographyMethods are described for the determination of triamcinolone acetonideat concentrations of 0.025 and 0.1% uu/m in creams and 2 and 6 mg ml-l insuspensions. Triamcinolone acetonide extracted from these formulations isassayed by high-performance liquid chromatography, using a reversed-phaseseparation on a Spherisorb S10 ODS column with a solvent system of methanoland water.The main advantages of the method are the over-all speed of analysisand the degree of precision obtained.G.GORDON and P. R. WOODE. R. Squibb and Sons Ltd., Reeds Lane, Moreton, Merseyside.Prednisolone is used as an internal standard.Analyst, 1975, 101, 876-882.Quantitative Determination of Ethoxyquin in Apples byGas ChromatographyEthoxyquin is extracted from an apple homogenate with hexane and dis-tributed in an acidic aqueous phase. After treatment with sodium hydroxidesolution, ethoxyquin is taken up in hexane and mixed with heptafluorobutyricanhydride. The heptafluorobutyryl derivative so formed is analysed by gaschromatography using an electron-capture detector with tetrahydroquinolineas an internal standard.The lower limit of detection for ethoxyquin is0.02 ng.BJORN WINELLThe National Food Administration, Box 622, S-751 26 Uppsala, Sweden.Analyst, 1976, 101, 883-886November, 1976 SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Ozone by Gas ChromatographyAn improved gas-chromatographic method for determining ozone in thepresence of other gases has been devised. The column can be used a t ambienttemperature with negligible decomposition of ozone during the determination.The apparatus is calibrated unequivocally by thermally decomposing ozoneand measuring the resulting oxygen chromatographically. Mixtures of varyingproportions of ozone and oxygen from the anode of a perchloric acid electro-lysis cell were used for calibration, and in all instances the total amount ofanode gas produced was as expected from the Faraday equivalent.Calibrationusing iodimetric determination of ozone was unsatisfactory ; the iodimetricmethod appeared to overestimate ozone by about 30%.R. V. HOLLAND and P. W. BOARDCSIRO, Division of Food Research, P.O. Box 52, North Ryde, N.S.W. 2113,Australia.Analyst, 1976, 101, 887-891.ixAppraisal of the Determination of Molecular Size and Mass byMembrane DiffusionThe diffusion of dissolved substances through porous membranes, the poresof which approach molecular or macro-molecular dimensions, may providea simple and sensitive means of determining molecular size or mass and ofdetecting changes in molecular, especially macro-molecular, size that havebeen brought about by changing the solvent composition or temperature.The value of the method depends largely on the ability of such membranesto hinder to varying degrees the diffusion of different solute species havingsimilar molecular sizes or masses.An explicit mathematical formulation ofthe problem is undertaken with a view to assessing, in the light of experi-mental results published by other workers, the potential value of the method.PETER E. O’CONNORWater and Effluents Laboratory, Engineering School, University College, UpperMerrion Street, Dublin 2, Eire.and MICHAEL G. HARRINGTONBiochemistry Department, University College, Blackrock, County Dublin, Eire.Analyst, 1976, 101, 892-900.X-ray Analysis of Particulate Matter Collected in an EnvironmentalMonitoring DeviceX-ray induced photoelectron spectroscopy, energy dispersive X-ray fluores-cence spectroscopy and X-ray diffraction techniques are used in conjunctionto give routine qualitative and semi-quantitative elemental and chemicalanalyses of airborne particulate matter collected in an environmental moni-toring device.The advantages and disadvantages of each technique arediscussed and examples are given of typical analyses of samples of particulatematter weighing down to 100 pg. The results from the first two techniquesare complementary but X-ray diffraction does not often give the unequivocalstructural information expected.G. C. S . COLLINSIBM United Kingdom Laboratories Ltd., Hursley Park, Winchester, Hampshire,SO21 2JN.and D.NICHOLASFulmer Research Institute, Stoke Poges, Buckinghamshire.Analyst, 1976, 101, 901-911x SUMMARIES O F PAPERS IN THIS ISSUE November, 1976Determination of Arsenic( 111) and Total Arsenic by theSilver Diethyldithiocarbamate MethodCommunicationJ. AGGETT and A. C. ASPELLChemistry Department, University of Aul-,kland, Auckland, New Zealand.Analyst, 1976, 101, 912-913.ANALYTICAL SCIENCES MONOGRAPHNo. 1High-Precision Titrimetryby C. Woodward and H. N. RedmanImperial Chemical Industries L iniite d (Agricultural Division)BRIEF CONTENTS:lntro ductionVisual Titrations, with sections on Apparatus, Standard Substances and their preparation andInstrumental Methods, with sections on Photometric ,Titrations, Electrornetric Titrations andReferences to the literature of high-precision titrirnetry.assay, and Standard Solutions.Miscellaneous Techniques.Price f2.50Pp. viii+63 ISBN 0 85990 501 2Obtainable from :The Publications Sales Officer,THE CHEMICAL SOCIETY,Blackhorse Road, Letchwmth, Herts., SG6 1 H NMembers may buy personal copies at the special price of f2.00 provided they orderdirect and enclose remittanc

ISSN:0003-2654    

DOI:10.1039/AN97601BP089

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

NOVEMBER 1976 The Analyst Vol. 101 No. 1208 An Assessment of Various Types of Reference Electrode for Use in Continuous Potentiometric Analysis with Particular Application to Highly Pure Waters D. Midgley and K. Torrance Central Electvicity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey, KT22 7SE The suitability of a number of reference electrodes for use in the continuous potentiometric analysis of highly pure water has been assessed. The electrodes were chosen to be as representative as possible of their class, but several electrodes of novel design were also included. The electrodes were tested for the constancy of their potentials over 100 d of continuous operation in mol 1-1 sodium hydroxide solution (a simple model boiler water) and for the extent to which they were affected by ultraviolet light, particulate matter and changes in flow-rate and ionic strength.The times taken by the electrodes to recover from changes in temperature, from an interruption in the flow of sample and from being immersed in buffer solutions were measured and proved to be important characteristics in discriminating between the over-all performance of electrodes. It was concluded that conventional types of electrode, whether calomel or silver - silver chloride, with ceramic frit junctions and unsaturated potas- sium chloride electrolytes, were superior in almost every respect to the other types tested. In almost every application of ion-selective electrodes, use is made of a reference electrode, Le., an electrode of supposedly constant potential, so that changes in the e.m.f.of an electro- chemical cell comprising sensing and reference electrodes can be ascribed solely to changes in the potential of the sensing electrode and consequently to changes in the activity of the species to which that electrode responds. The term “reference electrode” is generally a misnomer for a reference half-cell, consisting of the reference electrode itself, immersed in a solution of constant composition that makes contact with the test solution by some means of establishing a reproducible liquid junction between the two. We have adopted the con- ventional usage and call the half-cell an electrode, keeping the term reference element for the electrode contained by the half-cell. The book by Ives and Janzl is the prime source of information on the basic electrochemistry and construction of reference electrodes. Mattock2 has given an account of applications of reference electrodes and Covington3 has considered them in relation to ion-selective electrodes.Recently, new forms of the traditional reference electrodes have been developed, both com- mercially, for use in process contr01,*~~ and experimentally.6 Other less familiar types are used for special applications, e.g., the Thalamid electrode at high temperatures. So far, little attention has been paid to the specific problems that arise with reference electrodes in the continuous monitoring of water. The manner in which the reference electrode potential changes will affect both the precision and the accuracy of the analytical measure- ment, and a better knowledge of the behaviour of reference electrodes will benefit all potentio- metric methods of analysis, regardless of the species in question. We have investigated the properties of some novel or relatively little used electrodes in comparison with classical designs of electrodes, and with the particular application to the anslysis of boiler water, high-purity boiler feed-water and condensed steam in mind.Experimental Electrodes Tested The electrodes tested in this work (see Table I) were chosen to be as typical as possible of classes of electrodes rather than merely isolated examples of a manufacturer’s products. 833834 MIDGLEY AND TORRANCE: ASSESSMENT OF REFERENCE ELECTRODES Analyst, VOZ. 101 In view of the absence of published results, electrodes of traditional design, such as the three types of Pye electrode (similar electrodes are available from most manufacturers), were included for their own sake, as well as to provide a basis with which the less familiar types could be compared.Three types of reference element were tested: calomel, silver - silver chloride and Thalamid; the commonest (calomel) was tested in a number of configurations. Manufacturer and name Beckman Permaprobe EIL RK.28 .. Philips R.44D Pye 305 . . .. Pye 305-W .. . . Pye 360 . . .. RhiiEka . . . . REFERENCE ELECTRODES TESTED Reference element* C C C C C S C Filling solution Sat. KCl Sat. KC1 Junction Intended type? usel P I F I F/D L F I G I F I F E Code for electrodes tested Comments D, E B, C Sealed unit, Catalogue 1 M KNO, Type 39407, "solid state" NO.33-1320-610 F Fitted with reservoir of See text for details of construction 3.5 M KC1 Schott 9828 . . . . T 3.5 M KC1 F I I(, L Fitted with reservoir of *C = calomel; S = silver - silver chloride; T = Thidamid. t D = double-junction; F = ceramic-frit; G = ground-.glass sleeve; P = porous PTFE. SE = experimental; I = industrial; L = laboratory. R 6 M k a electrode RfiiiEka et aZ.6 described a reference electrode in which the reference element was formed by impregnating a graphite electrode with calomel or silver chloride. We made a similar electrode of the calomel type (G) as follows. A RGiiEka electrode body (Radiometer F 3012, V. A. Howe and Co. Ltd.) was modified by cutting a thread along the bottom 40 mm of its length.A groove, 1 mm wide by 1 mm deep, was then cut from the bottom of the electrode along 70 mm of its length. The graphite surface was impregnated with electrolytically prepared mercury( I) chloride (Hopkin and Williams Ltd.). An electrolyte reservoir was constructed from a PTFE tube (80 mrn long, wall thickness 4.5 mm) threaded internally for a distance of 30 mm from one end with a thread which matched that cut on the external surface of the electrode. The threaded end of this cylinder was sealed by screwing a PTFE base piece (8 mm long) into the internal thread. This base piece was drilled centrally along its length and a ceramic frit (Electronic Instruments Ltd.) was pressed through the slightly undersized hole, forming a satisfactory seal for the ceramic- frit junction.The reservoir was filled to a depth of 5 mm with a paste of potassium chloride crystals and screwed on to the modified electrode body. Apparatus Potentials were measured on a Corning 110 digital pH meter, reading to 0.1 mV. When used with the Central Data Acquisition and Processing System (CDAPS) data-logging system (see below), potentials were read to 0.01 mV in the &lo0 mV range and 0.1 mV in the & 1000 mV range. A cabinet, maintained at 25 j-- 1 "C by means of a 300-W fan heater controlled by a relay from a mercury contact thermometer, housed the electrodes, the flow cell and the reservoirs of electrolyte during the tests on long-term stability and effects of temperature, immersion in buffer solutions and stopping the flow of electrolyte.The flow cell was a Perspex cylinder (100 mm deep and 126 mm internal diameter) enclosed by a concentric water jacket (internal diameter 190 mm) through which water at 25 "C was circulated by a Churchill Laboratory Thennocirculator. The test solution was fed by gravity through a glass capillary at the rate of 8 ml min-1 to the flow cell from a constant-h.ead reservoir supplied from a 50-1 reservoirNovesn ber, 19 76 FOR USE I N CONTINUOUS POTENTIOMETRIC ANALYSIS 835 by a peristaltic pump. The effluent from the flow cell overflowed through a side-arm and was led to a drain. Fig. 1 shows the flow system schematically. The voltage drop across a 400-SZ resistor in series with a 13.5-kQ resistor connected to a 1.35-V Mallory cell (Type RM.42R) was used as the “test potential” to check the operation of the data-logging system.The common reference electrode was placed in the middle of the flow cell with the test electrodes in a circle of radius 45 mm around it. For the long-term evaluation of the electrodes and other experiments in which it was desired to monitor several electrodes virtually simultaneously, the Central Data Acquisi- tion and Processing System (CDAPS) at CERL was used. The electrodes were connected through a signal multiplexer to a Corning 110 pH meter, which was itself connected through a second multiplexer to the central digital voltmeter of the CDAPS system. The first multi- plexer served to switch each electrode in turn to the pH meter and the second transmitted the signal from the recorder output terminals of the meter to the CDAPS.The Corning 110 pH meter was particularly suitable, as it has a fast response time (the switches in the multi- plexer are closed for only 0.1 s) and a potentiometric recorder output that can be adjusted to serve as a unity gain amplifier. Technicon pump Constant-head ~ reservoir To drain - * Multiplexer- 2 pH meter To CDAPS 400 s2 Fig. 1. Schematic diagram of flow and electrical systems. The arrangement is shown schematically in Fig. 1. The scan through each set of readings, the eleven electrode potentials plus three other readings (“CDAPS zero,” “pH meter zero” and “test potential”), was completed in 1.5-3 s. The rate at which readings were taken could be adjusted in multiples of 1 niin. Each reading was printed on a Data Dynamics 390 Teletype Console and could also be punched on to 8-channel paper tape.“CDAPS zero” was a reading taken across two shorted terminals in the second multiplexer and was a check on the zero of the CDAPS voltmeter; no problems were encountered. ‘‘pH meter zero” was a reading taken across two shorted terminals in the first multiplexer and was a check on the zero of the pH meter - CDAPS combination. This reading, although836 MIDGLEY AND TORRANCE: ASSESSMENT OF REFERENCE ELECTRODES Analyst, VoZ. 101 generally larger and more variable than the CDAPS zero (indicating that the pH meter was the less stable part of the measuring system), was small and it was not necessary to adjust the instrument zero during the run. The standard deviations of successive readings 1 min apart were in the same range (0.01-0.02 mV) for the electrodes, the test potential and the pH meter zero, while the CDAPS zero had a standard deviation of 0.005 V.Before the tests started, the accuracy of the readings for each electrode was checked with a Solartron LM1420.2 digital voltmeter reading to 0.1 mV. ’The A1 000 mV range of the CDAPS and the Solartron meter showed no difference (i.e., -:0.1 mV) for the electrodes connected to that range (K, L). The difference in readings between the &lo0 mV range of the CDAPS and the Solartron meter was no greater than 0.05 mV for the remaining electrodes. Result s8 The Pye 305 electrode was initially chosen as the “master” electrode, against which the potentials of the electrodes under test were measured, as it was of the most widely used calomel type.Preliminary experiments on the effect of flow-rate did not suggest that a better choice was available and the same electrode was used in the long-term tests. When the observed effect could not be attributed unambiguously to the test electrode alone, another electrode was used as the “master” to check the first set of results. Effect of Flow-rate In order to investigate the effect of flow-rate on the reference electrodes, a number of systems were tried in which the sample solution was pumped through a flow cell and then run to waste. A common failing of all of these systems was the high level of electrical noise that was associated with the static electricity generated by the pumping. Efforts to reduce the noise to tolerable levels by earthing the sample solution or using stainless steel instead of plastic transmission tubing produced no significant improvement.Two arrangements that avoided these difficulties were a gravity-fed flow system and a closed-loop pump system; both were used to study the effects of flow-rate as described below. As linear rather than volume flow-rates determine any effects on the electrodes, the results obtained were relative, not absolute. A gravity-fed flow cell was constructed from a 2-1 Pyrex beaker by the addition of a small glass compartment with a capacity of about 15 nil connected to the beaker by a short glass tube. The electrode under test was placed in the side compartment, where the linear flow-rate was greatest, while the master electrode (M) was placed in the 2-1 beaker, where the linear flow was negligible by comparison.Ammonium chloride solution containing ammonia at a concentration of 1 pg d-l was delivered by a peristaltic pump to the constant-head reser- voir; the flow-rate was 40 -+ 5 ml min-l. The ]!inear flow-rates past the liquid junctions depended on the geometry of the electrodes and varied from about 14 cm min-l for electrodes B, C, F, I< and L to 36 cm min-l for electrode A. The cell e.m.f.s were recorded under both flowing and static conditions; the difference in performance was attributed to the effects of flow on the test electrode. The results of these tests are summarised in Table 11. The e.m.f.s of all of the electrodes except Thalanid K were more positive in flowing solutions than when the flow was stopped.Electrodes with a shift of less than 6 mV (equivalent to 0.1 pH unit) were calomel electrode A, solid-state electrode D, sealed electrode B and the two Thalamid electrodes (K and L). Those with a shift of less than 2.4 mV (equivalent to a 10% error in concentration in the determination of a univalent ion) were electrodes A, D and K. The electrodes with ground-glass sleeve junctions (P and Q) exhibited gross shifts. The noise levels of all of the electrodes increased in flowing solutions, but the Thalamid and ground-glass sleeve electrodes were particularly bad. Electrodes of the same type could differ significantly in behaviour as regards both millivolt shift and noise; one electrode of a pair could be better in one respect without being so in the other.Part of the noise might have arisen from streaming potentials at the inlet and outlet of the side compartment and in order to eliminate these variables, and also any adverse effect of the liquid junction of the master electrode, ;a closed-loop pump system was used. In these experiments, the reference electrode under test was placed in a 25-ml glass beaker containing approximately 10 ml of sodium fluoride solution containing 1 pg ml-1 of fluoride ions. The cell was completed by an Orion, Model 96-09, fluoride electrode. The latter wasNovember, 1976 FOR USE I N CONTINUOUS POTENTIOMETRIC ANALYSIS 837 TABLE I1 EFFECT OF FLOW-RATE ON ELECTRODES IN GRAVITY-FED FLOW CELL Mean e.m.f. vevsus 3 M calomel electrode M/mV Short-term noise/mV Electrode Static F l o w i n g * -ing* A + 0.9 + 1.7 0.1 0.6 P + 18.9 +49.0 0.1 2.0 + 34.5 + 73.0 0.1 6.0 + 14.4 + 19.6 0.2 0.5 Q B C + 16-20t + 30.0 0.1 0.3 D +9.1 +11.2 0.1 0.2 E + 26.3 + 32.8 0.3 0.9 F - 150.6 - 141.5 0.2 0.5 K + 837 + 836 2.0 6.0 L + 860 + 864 2.0 6.0 * Flow-rates were 40 f 6 ml min-1.t This electrode exhibited drift in potential. chosen because it could provide a reference potential that contained no component from a liquid junction. The fluoride solution was circulated by a Quickfit, Type 10 PP60, peristaltic pump at flow-rates that could be varied from 24 to 0 ml min-1. The ends of the transmission tube that carried the solution round the system were attached to glass tubes immersed in the beaker and held in a fixed position by a glass cross-piece.The effects of flow on a selection of reference electrodes at flow-rates of 24 and 11 ml min-l, together with their responses in static solution, are reported in Table 111. At 24mlmin-l, the linear flow-rate past the electrodes was about 4 cm min-l. TABLE I11 EFFECTS OF FLOW-RATE IN CLOSED-LOOP PUMP SYSTEM Mean e.m.f. vemus Orion fluoride electrode/mV Electrode M A P 8 C D G F K L Fldw-rate Flow-rate 24 ml min-' 11 ml min-1 + 53.1 - +54.3 - + 62.2 + 62.3 +64.5 + 64.4 + 66.8 - + 67.4 + 67.1 $62.1 + 62.2 + 104.0 + 105.0 + 885.8 + 885.7 + 885.4 + 885.4 - 99.5 - 98.6 Static +53.4 + 54.2 + 62.5 + 64.6 + 66.9 + 67.2 $61.8 + 106.0 +885.2 + 885.6 -97.0 Short-term noise/mV 11 ml min-l A r Flow-rate Flow-rate 24 ml min-l 0.1 0.1 - 0.2 0.1 0.2 0.1 0.1 - 0.1 <0.1 0.3 0.3 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.1 - 7 Static (0.1 t0.1 <0.1 (0.1 < 0.1 <0.1 <0.1 <0.1 tO.l (0.1 <0.1 The effects of flow-rate in this system were much less than those observed in the gravity-fed flow cell, the most probable explanation being the lower linear flow-rates of the solution in the 25-ml beaker.In general, there was little difference between the e.m.f.s recorded at 24 and 11 ml min-l, and these values were very close to the e.m.f. recorded in static solution. The noise levels recorded in both flowing and static solutions were lower than those recorded in the gravity-fed cell. As the noise levels in the static solution were considerably different from those in the other flow cell (Table II), it was concluded that the geometry of the cell and the relative positions of the electrodes have a major influence on the noise levels.The contribution from the different master electrode and the change of electrolyte can be dis- counted because experiments with a Thalamid electrode in the gravity-fed flow cell, using the Orion fluoride electrode and sodium fluoride solution, produced results similar to those reported in Table 11. The e.m.f. of the double-junction electrode F was about 160mV different from those of other saturated calomel electrodes. It was supposed that the potassium chloride in the inner838 MIDGLEY AND TORRANCE: ASSESSMENT OF REFERENCE ELECTRODES Analyst, Vd. 101 electrode compartment, which would not be refilled, had diffused into the solution in the outer sleeve. As there was no solid potassium chloride in the inner compartment, the chloride concentration would have been gradually depleted, with a consequent change in e.m.f, In Tables I1 and 111, noise is defined as the maximum amplitude of the irregular oscillation about the mean e.m.f.The frequency of this oscillation in the flowing solutions was about 20 min-1 in the gravity-fed flow cell and was dominated by the impulses of the pump in the closed-loop system. In static solutions the frequency was irregular. Effect of Head of Electrolyte The effect of the level of the electrolyte solution in the reference electrode above the liquid junction was investigated for those electrodes which had side-arms fitted to their bodies. These tests were carried out in flowing solution using the closed-loop pump system described above.The flow-rate was 24mlmin-l and the e.m.f.s were measured with the head of electrolyte adjusted to 300 and 500 mrn above the junction. The millivolt shifts on increasing the head of electrolyte were as follows: M, -0.2; A, -0.1; P, + O . l ; Q, +0.5; F, -0.3; K, -0.4; and L, -0.2. The noise on the signal was not more than 0.1 mV for all of the electrodes except the double-junction electrode F (0.5 mV) in both cases. The shifts were of little, if any, significance, although it may be rioted that the electrodes with ground-glass sleeves (P and Q) changed in the opposite sense to the others. In the subsequent long-term tests, the head of electrolyte was 300 mm above the liquid junctions. Long-term Tests As the aim was to test the suitability of the electrodes for continuously monitoring power station waters of high purity, the electrodes were run for 100 d on 10-5 moll-1 sodium hydroxide solution, which was chosen as a simple model boiler water.The stability of the electrodes was assessed in terms of drift (the cha:nge in e.m.f. with time) and the variability of the drift (the standard deviation of the drift rate). The electrodes were placed in the cylindrical l?erspex flow-cell with the master electrode in the middle. The test electrodes were grouped according to the convenience of siting the electrolyte reservoirs, which contained enough solution to last the test. Preliminary tests had shown that switching the positions of two electrodes or putting the master electrode in one of the peripheral positions made no difference to the potentials or the variance in the potentials and it was concluded that all positions in the flow cell were equivalent. Readings of potential were recorded every hour by the data-logging system.Readings of the electrode potentials were corrected by subtracting the pH meter zero at an early stage of data processing. The “test potential’’ was used to check the performance of the CDAPS, enabling data recorded during periods of malfunction to be eliminated automatically at the data-processing stage. For each electrode, the mean drifts over periods of 1 h, 1 d and 1 week were calculated, together with the corresponding standard deviations. The individual figures from which the mean drifts were calculated were the algebraic differences between each hourly reading and the corresponding reading 1 h, 1 d or 1 week previously, provided that this was a valid reading.Student’s t-test was used to check if the drift rates were significantly different from zero. Regression analysis was carried out in order to obtain the correlation between each pair of electrodes. A valid reading was defined as one that (a) was actually on the punched tape, i e . , had not been omitted because of a fault on the Teletype, (b) had not been recorded during a malfunction of the data-logging system (checked by the “test potential”) and (c) was not invalidated by being mispunched by the Teletype. A FORTRAN program was written to enable the calculation with the large amount of data to be performed on an IBM 370 computer.The results are presented in two sections: using data from the first 23 d (Table IV) and from thle entire period (Table V). The results were calculated in this way as a guide to whether changes occurred in the drift rate as the run progressed. The first period coincided with the life of the RGiiEka electrode, which went open-circuit after 23 d of operation and was thus exclud2d from the calculations over the full period. The readings for each electrode were correlated with those for all other electrodes. Pairs with correlation coefficients greater than 01.70 are listed in Table VI. It can be seen that in terms of drift there was little difference between the performanceNovember, 1976 FOR USE IN CONTINUOUS POTENTIOMETRIC ANALYSIS TABLE IV DRIFT RATES versus MASTER ELECTRODE M OVER 23-d PERIOD 839 Mean*/ s.d.t/ Mean*/ s.d.t/ Mean * / s.d.t/ Electrode mV h-l mV mV d-1 mV mV week-l mV B 0.02 0.29 0.09 1.33 -0.95; 1.04 C -0.01 0.24 - 0.075 0.51 -0.91: 0.54 D 0.05 3.42 0.39 5.77 -0.10 6.61 E -0.01 0.40 0.02 1.09 - 0.32: 1.08 G - 0.28 7.39 - 1.20 15.2 3.52: 15.8 H - 0.006 0.20 - 0.10: 0.39 -0.791 0.43 -0.003 0.17 - 0.0411 0.35 -0.51; 0.53 -0.02 0.50 -0.159 1.09 - 1.42: 1.96 K L -0.02 0.51 -0.155 1.13 - 1.44: 2.04 * Not significantly different from zero unless marked with symbols as below.t Standard deviation of a single difference (408 hourly points, 332 daily points, 235 weekly points). ; Significantly different from zero a t the 0.1% level. 5 Significantly different from zero a t the 1% level.11 Significantly different from zero at the 5% level. A 0.000 0.11 -0.02 0.22 - 0.20; 0.20 F -3.34 66.8 -0.91 104.9 -0.81 88.0 J of the sealed units (B and C) and silver - silver chloride electrodes (H and J), while the 3 M calomel electrode (A) was slightly better than both. The Thalamid (K and L) and solid- state electrodes (D and E) were less satisfactory and the double-junction (F) and Rfiiitka (G) electrodes behaved poorly. The standard deviations of silver - silver chloride electrode J were, over the whole period, superior to those of the other electrodes. The sealed electrodes (B and C) also performed well in this respect. TABLE V DRIFT RATES verws MASTER ELECTRODE M OVER 100 d Mean * / s.d.t/ Mean*/ s.d.t/ Mean*/ s.a.t/ Electrode mV h-l mV mV d-1 mV mV week-' mV A B C D E F H J I< L - 0.00 1 0.004 0.001 0.01 -0.002 - 0.69 0.007 - 0.003 0.003 -0.009 4.83 0.29 0.25 4.17 0.41 30.3 11.5 0.08 0.89 0.90 -0.02 - 0.04; -0.075 0.15 0.125 -0.36 - 0.06 - 0.06s 0.135 -0.195 5.09 1.05 0.53 4.86 1.37 45.1 12.1 0.21 1.63 1.56 -0.13 -0.475 - 0.535 0.19 0.638 - 1.59; - 0.45 - 0.455 - 1.379 0.805 5.42 1.02 0.70 6.51 2.32 33.8 12.8 0.40 4.77 2.28 * Not significantly different from zero unless marked with symbols as below.t Standard deviation of a single difference (1 989 hourly points, 1 796 daily points, 1 687 weekly points). Significantly different from zero at the 5% level. 5 Significantly different from zero at the 0.1% levcl. Comparing results from the 23-d and 100-d periods, for most electrodes there was little change, but the standard deviations of the 3 M calomel (A) and silver - silver chloride (H) electrodes deteriorated markedly, although the drift rates were little altered.The double- junction electrode (F) improved over the full period, as the high initial rate of change had less significance over the longer period. The high rate of drift for this electrode is probably due to the depletion of the chloride concentration in the inner electrode compartment (see Effect of Flow-rate). As the difference in concentration between the inner and outer com- partments decreases, the rate of drift should also decrease, as was observed over the period of the test. The Thalamid electrode K was remarkable for significantly reversing its direction of drift over the two periods. As might be expected, there was a fairly high correlation between the e.m.f.s of pairs of similar electrodes (B and C; H and J ; K and L).The number of significant correlations decreased as the test progressed and the standard deviations of some electrodes increased840 MIDGLEY AND TORRANCE: ASSESSMENT OF REFERENCE ELECTRODES Analyst, VoZ. 101 TABLE V1: CORRELATION BETWEEN ELECTRODES,* MEASURED AGAINST ELECTRODE M Correlation coefficient, Rt 2 3 G d r i o d C - A 0.75 f 0.014 - C - B 0.77 f. 0.04 0.945 f 0.005 C - H 0.90 f 0.03 - 0.78 f 0.04 0.972 f 0.003 0.72 f 0.06 - C - J C - K J - A J - B 0.74 f 0.04 - J - E 0.87 f 0.03 - J - H 0.88 f 0.03 0.916 f 0.007 J - K 0.78 f 0.05 - J - L H - K 0.75 f 0.05 - H - L 0.74 -J= 0.05 - K - L 0.994 f 0.001 - B - L - 0.84 f 0.01 * Only correlation coefficients with R > 0.7 or R < -0.7 are shown. t 95% confidence limits are given.Although the distribution of R is not Gaussian, at high values of R the error in quoting symmetrical figures is small and conservative values have been taken. Electrode combination C - E - -0.72 f 0.02 c - L 0.72 f 0.05 0.90 f 0.01 0.72 0.05 -0.75 f 0.02 - 0.90 f 0.01 - -0.75 0.02 H - A (A and H). Electrodes with high rates of drift had large standard deviations, but the converse was not necessarily true. The high correlation coefficients could be caused by the similar susceptibility of the different electrodes to the effects operating on them, or by the possibility that the variation attributed to the test electrodes may come predominantly from the master electrode.The fact that high correlation coefficients are obtained only between electrodes with good performances, even though they are of dfferent types (silver - silver chloride, Thalamid and calomel electrodes are represented) suggest that in these instances the master electrode may make a large contribution to the variation, whereas for the poorer electrodes its contribution is relatively small. In order to check this possibility, the results were re- calculated with respect to one of the test electrodes as the master electrode. Although both the silver - silver chloride (J) andt the sealed electrodes (C) were consistently good, the latter was preferred as the alternative master electrode because it had one fewer variable in its performance because, being saturated with potassium chloride, there could be no concentration of the filling solution by evaporation of water from the reservoirs.The high correlation between the two also made re-calculation on both bases redundant. The revised data were obtained by subtracting the e.m.f. recorded with electrode C from those observed with the other electrodes. A reading for the previous master electrode, 3 M calomel electrode M, was obtained by subtracting the reading of C from the pH meter zero reading. Once these corrections had been made, the calculations were made exactly as before. The effect of taking sealed unit C as the master electrode was to reduce slightly the rate of drift in most instances. The effect on the standard deviations was negligible, except that the similar electrode B was improved in this respect while electrodes A and J deteriorated.The number of significant correlations decreased sharply (Table VII) , especially when pairs of similar electrodes (M and A ; K and L) were discountcd. On the basis of the 100-d period of constant immersion, electrode C would have been a better choice as the master electrode, although in general this type of electrode is not as good as the 3 M calomel electrode. Effect of Stopping the Flow of Electrolyte The effects of stopping the flow of electrolyte through the flow cell for periods of 1 and 16 h were determined. A period of 1 h was considered to be representative of short stoppages that occur as a result of routine maintenance, e.g., replacement of an electrode or standardisa-November, 1976 FOR USE I N CONTINUOUS POTENTIOMETRIC ANALYSIS TABLE VII CORRELATION BETWEEN ELECTRODES,* MEASURED AGAINST ELECTRODE C Correlation coefficient, Rt 2 3 e i o d Electrode combination M-A 0.982 f 0.004 - n3 - J 0.82 f 0.04 - 0.84 -f: 0.03 - 0.989 -f: 0.002 - A - J 0.931 & 0.006 M - E - K - L * Only correlation coefficients with R > 0.7 or R < -0.7 are given.t 95% confidence limits are given. Although the distribution of R is not Gaussian, at high values of R the error in quoting symmetrical figures is small and conservative values have been taken. 841 tion of pH systems by buffer solutions. The period of 16 h was included to give an indication of the behaviour of the electrodes following a longer stoppage, such as might be caused by plant failure, pump failure or a blockage of sample lines.The results are shown in Fig. 2. Within 1 h of re-starting after a stoppage of 1 h, the e.m.f.s. of all of the electrodes had recovered to within 2.4 mV (equivalent to a 10% error in concentration for a univalent ion) of their original values. The longest recovery times were recorded for the solid-state electrodes (D and E). - stop flow &-- Start flow Time - Fig. 2. Effect of stopping flow for 16 h. In general, the electrodes took longer to recover from the 16-h stoppage, but only the solid-state electrodes (D and E) took longer than 2 h to reach e.m.f.s that were within 2.4 mV of their original values. The recovery of these electrodes was very slow, approximately 6-7 h being required to re-establish equilibrium.Effect of Temperature Temperature studies on reference electrodes have mainly been of two types : non-isothermal measurements against a similar electrode at a fixed temperature and isothennal measurements842 MIDGLEY AND TORRANCE: ASSESSMENT OF REFERENCE ELECTRODES Analyst, Vd. 101 against a hydrogen electrode. Of more significance for industrial analysis is the question of how quickly the electrode returns to its original potential after being at other than its normal operating temperature. This effect was tested by measuring the e.m.f.s of the electrodes when they were returned to the flow cell after being equilibrated for 2 h, in batches of five, in a separate vessel containing 10-5 mol 1-1 sodium hydroxide solution a t 15 or 37 "C. - 25 "C c Time - Fig.3. Recovery of electrodes after decrease in temperature. 25OC 37°C > E l h L E F B J mvf l h LI Time - Fig. 4. Recovery of electrodes after increase in temperature. The master electrode remained in the flow cell, which was maintained at 25 "C throughout. The results are presented in Figs. 3 and 4. Table VIII shows the times required for the electrodes to recover to within 2.4 and 6.0 mV of their previous values at 25 "C. These TABLE VIII RECOVERY OF ELECTRODES FROM EXPOSURE TO DIFFERENT TEMPERATURES Time to recover to within 2.4 mV of previous valuelmin Time to recover to within 6.0 mV of previous valuelmin - 15 "C 37 "C * * 120 A 168 50 360 <9 100 B 72 t C 24 150 9 ( 6 72 96 D 132 t <7 < 12 t < 12 E t F <7 36 <6 <9 (6 L <7 t <7 < 12 Electrode 7 * * * * HKJ * The e.m.f. varied less than 2.4 mV from the equilibrium value measured before f The equilibrium e.m.f.was more than 2.4 mV different from the value measured exposure to change in temperature. before exposure to change in temperature.November, 1976 FOR USE IN CONTINUOUS POTENTIOMETRIC ANALYSIS 843 limits are equivalent to errors of 10% and 0.1 pX unit, respectively, for a univalent ion. Handling the electrodes could not itself produce a comparable effect (see Effect of Buffer Solutions). In general, the calomel electrodes took longer to recover from their exposure at 37 "C than at 15 "C. The solid-state (D and E) and sealed electrodes (B and C) were unsatisfactory in that they could take more than 1 h to recover to within 2.4 mV of their original potentials from 15 "C and, at the higher temperature, did not generally achieve this recovery in a period of 16 h.The sealed electrodes (B and C) showed both positive and negative fluctuations before they returned to steady potentials. The Thalamid electrodes (K and L) also took several hours to regain their original potentials, but were within the limits of 2.4 and 6.0 mV much more rapidly than either the solid-state (D and E) or sealed electrodes (B and C). The performance of the double-junction electrode (F) was comparable to that of the Thalamid electrodes. The electrodes least affected by the changes in temperature were the silver - silver chloride electrodes (H and J), which had a maximum displacement of about 1 mV before returning to their previous equilibrium values, while the worst affected were the elec- trodes filled with saturated potassium chloride solution, possibly because slow dissolution and precipitation of crystals of potassium chloride increased the recovery times and the change in chloride concentration as the solid dissolved or precipitated increased the e.m.f.change beyond that caused by changes in the standard potential of the electrode and in the activity coefficients. Effect of Ionic Strength Each electrode was placed in turn in an EIL Perspex flow cell (Cat. No. 24-8990-240) and changes in its e.m.f. response (veysuus the 3 M calomel electrode 11.1) were recorded when solutions of various strengths of potassium chloride and moll-l) were pumped through the cell at 3.9 ml min-1.The results are summarised in Table IX. TABLE IX EFFECT OF IONIC STRENGTH IN FLOWING SOLUTIONS Cell potential*/mt' A r 7 Electrode 10-9 moll-' KC1 lo-* moll-' KCl 1 0-6 mol 1-1 KC1 A +0.1 +0.1 + 0.6 B + 0.6 +1.1 - 0.9 C 0.0 -0.2 -0.1 D + 3.4 + 4.0 + 6.0 E + 8.4 + 12.7 1-2.3 F +3.4 + 4.2 + 1.8 H + 0.8 +0.7 -0.1 + 0.7 +0.6 - 0.2 + 6.3 + 6.3 + 5.2 J K L + 10.7 + 12.7 + 9.3 * All the results were normalised with respect to the mol I-' solution. With the exception of one of the sealed electrodes (C), a distinct pattern was observed. The e.m.f. increased on changing from mol l-l, either stayed the same or increased further on transfer to 10-4 mol 1-1 solution and then decreased on changing to rnol 1-1 solution. For electrodes that showed only small increases in the first two stages, the final potentials were smaller than the initial potentials, whereas the electrodes with large effects still showed increased final potentials.The effect was not dependent on the order in which the solutions were run. When the tests were repeated, the same trend was observed, although the magnitude of the changes was slightly different. A similar trend was observed when the electrodes were placed in beakers containing ammonium chloride solutions at concentra- tions between 5 x and 5 x moll-l. In order to check that the effects could reasonably be attributed to the test electrodes rather than the master electrode, the latter was tested against a saturated calomel electrode with a ceramic-frit junction (EIL RJ.23) in both the flow cell and the beaker.Almost no effect was observed, whereas when solid-state electrode E was tested against the RJ.23 electrode it showed a difference of 11 mV between and moll-I potassium chloride to844 MIDGLEY AND TORRANCE : ASSESSMENT OF F:EFERENCE ELECTRODES Analyst, VoZ. 101 solutions; this difference is of the same order as that found in the initial tests (Table IX). It was therefore concluded that the changes in e.m.f.s could be attributed to the electrodes themselves rather than to the master electrode. A concentration-dependent streaming potential in flow cells has been demonstrated by Van den Winkel et a1.’ A general effect such as they describe would apply to all the electrodes and could not account for the difference between them. The effects cannot be related simply to the type of liquid junction formed, as electrodes with saturated (B, C, D and E) and unsaturated (A, H, J, K and L) potassium chloride solutions appear among the most and least affected, as do electrodes with flowing (A, H, ,J, K and L) and non-flowing (B, C, D and E) reference solutions.Effect of Buffer Solutions The commonest application of reference electrodes in a power station is in the measurement of pH where the electrodes are periodically withdrawn from the flowing test solution of low ionic strength and immersed in a static buffer solution of relatively high ionic strength. From the deviations in the e.m.f.s of some of the: electrodes as a result of stopping the flow and changing the ionic strength, it was decided to check the effect of residence in the buffer solutions.The conditions were the same as for the long-term test except that readings were taken every 15 min. When steady readings had been established, the electrodes were withdrawn in batches of five and placed in a beaker of buffer solution. After 10 min, the electrodes were removed from the buffer solution, rinsed with distilled water and replaced in the flow cell. The potentials were recorded over the next 24 h. Three different buffer solutions were used: 0.05 mol 1-1 potassium hydrogen phthalate (pH 4.01), 0.025 moll-’ potassium di- hydrogen orthophosphate - 0.025 mol 1-1 disodiurn hydrogen orthophosphate (pH 6.86) and 0.01 moll-1 sodium tetraborate (pH 9.18). The temperature of the buffer solutions and wash water was the same as that of the flow cell (25 “C).The times taken for each electrode to return to a potential (a) 6mV from that before the buffer, equivalent to a shift of 0.1 pH unit, and (b) 2.4 mV from that before the buffer, equivalent to a 10% error in concentration of hydrogen ions, are summarised in Table X. TABLE :X RECOVERY OF ELECTRODES AFTER IMMERSION IN BUFFER SOLUTIONS Time to return to within 6 mV of previous Time to return to within 2.4 mV of previous value/min value/min I A \ f A \ Electrode Phthalate Phosphate Borax Phthalate Phosphate Borax A * * B 15 * C 10 15 15 16 30 60 15 15-30 E F * * * * * * * * * * 45 t 15 t t 16 t * * D t s * * * * * * 120 15-30 120 180 30-45 180 HKJ L 90 30-45 120 270 180 t * The potential did not exceed the stated limits. t The electrode took up a new steady potential outside the stated limit.It can be seen that the 3 M calomel (A) and silver - silver chloride electrodes (H and J) were satisfactory in each instance and that the sealed (B and C) and double-junction elec- trodes (F), although inferior, were generally satisfactory. Neither the solid-state (D and E) nor the Thalamid electrodes (K and L) were satisfactory. A “blank” experiment in which the electrodes were removed, rinsed with distilled water and returned to the flow cell produced little disturbance in the potentials, showing that the deviations must have arisen from immersion in the buffer solutions. Borax and phosphate buffers produced deviations of opposite sign to those caused by phthalate buffer. It is possible that these effects were caused ’by the absorption of salts from the buffer solution by the liquid junction, which released them on return to the flow cell.ElectrodesNovember, 1976 FOR USE I N CONTINUOUS POTENTIOMETRIC ANALYSIS 845 such as the 3 M calomel and silver - silver chloride electrodes (H and J), in which there is a mass flow of reference electrolyte across the junction, were less badly affected than those such as the solid-state (D and E) and sealed electrodes (B and C), which have no head of electrolyte solution. In the former instance, uptake of buffer salts by the junction would be minimised, while in the latter, interdiffusion between the buffer solution and the reference electrolyte solution would occur. The difference in sign between the effects of different buffers could be caused by the different sign of the sum of products C ciui (where ci and ui are the concentration and mobility, respectively, of the ith chemical species) which appears in the equation for the liquid-junction potential.The sign would be affected by the pH and concentration of absorbed salts. i Effect of Ultraviolet Light The effect of ultraviolet light on electrodes A-L was investigated by exposing them to a 15-min period of radiation from a 30-W Hanovia lamp at a distance of 500mm. The electrodes were immersed during this test in mol 1-1 potassium chloride solution in a glass beaker, and their e.m.f.s were measured versus the 3 M calomel electrode (M) protected from the radiation by a coat of black paint. In no instance was there a significant change of e.m.f.during or after irradiation. Effect of Immersion in Aqueous Suspensions of Magnetite The most common form of particulate matter encountered in power station waters is magnetite. On start-up, concentrations greater than 100 pg ml-l may occur. The effect on the reference electrodes was tested by measuring their potentials in 100-ml portions of 10-5 mol 1-1 sodium hydroxide solution, to which 0, 2.5, 25 and 250 pg ml-1 of about 1 pm long needle-shaped particles of magnetite had been added. Much of the magnetite aggregated and could not be kept in suspension. The potentials observed, measured against electrode M, are shown in Table XI. TABLE XI EFFECT OF IMMERSION IN AQUEOUS SUSPENSIONS OF MAGNETITE Values given are millivolt readings. Magnetite concentration/pg ml-l A I 7 Electrode 0 2.5 25 260 A B C D E F H - J K L -2.5 3.9 0.8 7.3 29.0 37.5 41.9 160.4 - 838 825 - 2.5 3.9 0.8 8.9 28.9 37.5 41.9 160.5 - 837 825 -2.5 3.8 0.8 7.1 31.6 37.6 42.0 159.7 - 837 825 -2.2 2 5 0.6 4.8 22.6 160.9 37.5 41.9 834 824 The solid-state electrodes (D and E) were badly affected at the highest concentration of magnetite even though it had aggregated, and the sealed (B and C) and Thalamid electrodes (K and L) also showed some effect at this concentration.The other electrodes were not affected over the period of time (several minutes) that they were exposed to the magnetite. Discussion The testing of reference electrodes is hampered by the lack of a single, unambiguous basis for comparison. The results can often be represented only as “the extent to which a test electrods is affecttd differently from a standard electrode.” The results in this work are of this nature, except for those on the effect of temperature, ultraviolet light and residence in buffer solutions, but care has been taken to provide firstly, the best reasonable choice of standard elxtrod2, and secondly, an alternative basis for comparison whenever possible.Experimentally, it was found that in those tests which were unambiguous, the same electrodes846 MIDGLEY AND TORRANCE : ASSESSMENT OF REFERENCE ELECTRODES Analyst, VoZ. 101 emerged as “good” and “bad” as from the other experiments. It was not practicable to include more than two electrodes of one type, but the results show that similar electrodes generally behaved alike and in a distinctive way from other types.For these reasons, we are confident of the conclusions drawn from this work, while realising that the magnitudes of the effects reported cannot always be precisely attributed to the electrode under test. The results show that reference electrodes can be the source of a number of errors in potentiometric analysis, as they can be affected by temperature, the ionic strength of the solution and flow-rate. Changes in temperature or a cessation of flow can cause errors for a considerable time after the disturbance itself has ceased. The errors present after the immersion of certain types of electrodes in buffer solutions can be both large and prolonged, negating the purpose of standardisation itself. A comprehensive survey of the results shows that the following electrodes performed well in nearly all of the tests: 3 M calomel (A), silver - silver chloride (H and J) and the sealed units (B and C), although the last type was badly affected by temperature changes, possibly because of slow dissolution and precipitation of crystals in the saturated potassium chloride filling solution.The solid-state, double-junction and RiiiiEka electrodes were markedly inferior in per- formance to those above, although qualifications should be made. The advantages claimed8 for the solid-state electrode in “dirty” process streams were not tested, as that was not a relevant requirement, but as these electrodes were more affected by magnetite than the others even those claims must be doubted.The double-junction electrode (F) was not designed specifically for long-term continuous operation and its use here can be considered to be experimental. The RGiiEka electrode was a purely experimental arrangement and as such was not strictly comparable to the commercial electrodes tested. When the electrode was dismantled at the end of the test, it was found that the potassium chloride paste had dried out and for long-term use the reservoir woiild need to be enlarged. The present poor results over 23 d should not be regarded as a condemnation of the system itself. The Thalamid electrodes had performances intermediate between those of the preceding two groups; in no instance did they have a better performance than the calomel or silver - silver chloride electrodes.The comparison between similar electrodes with ceramic-frit and ground-glass sleeve junctions (A, P and Q), although iavourable to the former in flowing solutions, does not allow a definite preference to be made for all circumstances, e.g., the ground-glass sleeve is particularly suited to the manual analysis of samples when there is a danger of contamina- tion of the j u n ~ t i o n , ~ as it can be readily flushed and re-made. Silver - silver chloride and calomel electrodes of similar construction had generally similar performances, although the former were less affected by temperature. As shown by the several types of calomel electrode tested, the construction of the complete electrode has more influence on the performance of an electrode than has the type of reference element, at least under the conditions of our experiments. Conclusions The difference in performance between the various forms of reference electrode tested shows that the choice can be a major factor in optimising a method of potentiornetric analysis. The results of this work show that reference electrodes are much more prone to variation when used for the analysis of pure waters than may have generally been realised, especially with regard to the effect of concentrated solutions and of buffer solutions that may be used for st andardisation. Conventional reference electrodes incorporating a ceramic-frit and calomel or silver - silver chloride elements in an unsaturated potassium dhloride solution are superior in almost every aspect of performance to the more recent developments in this field, which offer only the advantage of convenience. This work was performed at the Central Electricity Research Laboratories and is published by permission of the Central Electricity Generating Board. References 1. Ives, D. J . G., and Janz, G. J., “Reference Electrodes, Theory and Practice,” Academic Press, New York and London, 1961.November, 1976 FOR USE I N CONTINUOUS POTENTIOMETRIC ANALYSIS 847 Mattock, G., “pH Measurement and Titration,” Heywood, London, 1961. Covington, A. K., in Durst, R. A., Editor, “Ion-selective Electrodes,” National Bureau of Standards Special Publication No. 314, US Department of Commerce, Washington, DC, 1969. Light, T. S., “An Improved Reference Electrode for Process Measurement,” 16th National Sym- posium, Analysis Instrumentation Division, Instrument Society of America, Pittsburgh, Pa., May 25-27th, 1970. Neti, R. M., and Jones, R. H., “Performance and Application of a New Reference Electrode for Process Potentiometric Measurements,” 16th National Symposium, Analysis Instrumentation Division, Instrument Society of America, Pittsburgh, Pa., May 25-27th, 1970. RbiiEka, J., Lamm, C. G., and Tjell, J. C., Analytica Chim. Acta, 1972, 62, 16. Van den Winkel, P., Mertens, J., and Massart, D. L., Analyt. Chem., 1974, 46, 1765. Beckman Bull., No. 7232, Beckman Instruments Inc., Fullerton, Calif., 1972. Goodfellow, G. I., and Webber, H. M., Analyst, 1972, 97, 95. 2. 3. 4. 5. 6. 7. 8. 9. Received April I d , 1976 Accepted July 12th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100833

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

848 Ana(yst, November, 1976, Vol. 101, $9. 848-855 Factors Affecting the Limit of Detection of Sodium- responsive Glass Electrodes G. I. Goodfellow, D. Midgley and H. M. Webber Central Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey, K T22 7SE Factors affecting the limit of detection of sodium-responsive glass electrodes have been examined and i t is postulated that the ultimate limit is set by the dissolution of alkali metal ions from the glass membrane itself. Experi- mental results are presented that are consistent with this hypothesis and from which the following conclusions can be drawn: (a) measurements of very low sodium concentrations are favoured by high linear flow-rates of solution past the membrane surface and by having as low a sample temperature as possible, consistent with acceptable response times; (b) hydrogen-ion interference is negligible provided that the pH is greater than 10.5; (c) interference of ions from the alkaline additive is similarly negligible if an alkylamine such as diethylamine is used.The limit of detection obtained with EIL GEA.33 electrodes in an EIL flow cell with a flow-rate of 4 ml min-1 was about 0.07 pg 1-l of sodium at 20 OC, when the pH was adjusted t o 11.0 with diethylamine. The use of sodium-responsive glass electrodes for determining sodium concentrations of about 2 pg 1-1 and above in power station waters has been described by Hawthorn and Ray,l Webber and Wilson2 and Diggens et aL3 In all of' these determinations, electrodes manufac- tured by Electronic Instruments Ltd.were used, and the pH of the final solution was adjusted to about 10.5 with ammonia vapour in order to minimise the interference from hydrogen ions. Although it is possible to measure sodium concentrations of less than 1 pg 1-1 using this technique, measurements on highly pure waters in different power stations consistently gave values no lower than 0.3-0.5 pg F, even when radioactive tracer studies indicated much lower levels. Eckfeldt and Proctor4 reported the use of alternative alkaline additives with electrodes manufactured by Leeds and Northrup Co. They found that these electrodes gave a significantly smaller response to various alkylamines than to ammonia in solutions at the same pH, and they recommended the use of dimethylamine or diisopropylamine.Reagent consumption is less with dimethylamine and the use of this reagent with the Leeds and Northrup sodium ion analyser was described by Galetti and Spear.5 When diethylamine was used instead of ammonia with EIL electrodes in CEGB power stations, the lowest con- centrations measured decreased to about 0.1 pg l-l, still significantly above the level indicated by radioactive tracer measurements. Recently, Eckfeldt and Proctor6 have shown that increasing the linear flow-rate past the electrode decreased the measured sodium concentration of a stream of de-ionised water. They suggest that a t low flow-rates, a build-up of alkali metal ions (in their case lithium) occurs at the membrane surface, giving a falsely high sodium reading. As the flow-rate is increased, these ions are removed from the elect]-ode surface and the apparent sodium con- cent ration therefore decreases.The intention of this paper is to show that of the various factors that may affect the limit of detection, that arising from dissolution of alkali metal ions from the membrane is dominant. The other factors considered were as follows : hydrogen-ion interference, interference by the ionic form of the alkaline additive, interference by potassium ions from the reference electrode, temperature and flow-rate. The preliminary parts of the work were concerned with selection of the most suitable electrodes from different manufacturers for discriminative measurements of sodium at levels below 1 pg 1 -l. Experiment a1 Reagents The following alkaline additives were used : ammonia, 35% solution, AnalaR ; dimethyl- amine, isopropylamine, diethylamine, triethylamine and diisopropylamine, general-purposeGOODFELLOW, MIDGLEY AND WEBBER 849 reagents ; piperidine, purified general-purpose reagent ; and cyclohexylamine , laboratory- reagent grade.Hydrated antimony pentoxide (HAP) was a selective ion retention medium (Carlo Erba, Milan). The water used for the preliminary work was purified by passing water from a stainless-steel still through a twin-column mixed-bed deionisation unit. The sodium content of this water varied between approximately 0.5 and 2 pg 1-1, as measured by EIL sodium-responsive glass electrodes. “Sodium-free” water, henceforward termed HAP-water, was produced by treating de- ionised water with hydrated antimony pentoxide.’ An amount of this material suspended in water was packed into a polyethylene tube of 2 cm i d .to a height of 10 cm using a sintered polyethylene disc as a support. When deionised water was passed through the column the effluent was slightly cloudy owing to HAP of very small (virtually colloidal) particle size. This residual HAP was effectively removed by passing the water through a hollow-fibre ultra- filtration unit. The complete purification system is shown in Fig. 1; water containing the suspension was pumped through the hollow-fibre bundles. Operation of the back-pressure valve raised the trans-membrane pressure and water was forced through the fibre walls. Part of this filtrate was fed to the sodium monitor while the remainder (together with the unfiltered concentrate) was returned to the reservoir.The ultrafiltration unit had to be operated continuously for about 1 week in order to reduce the contamination by sodium from the components of the system to an acceptable level. L - \ Hollow fibre . ultra- Peristaltic 7 f I I trati on Pump ’ column HAP To sodium Filtrate return / \ Back-pressure Concentrate va I ve return 7 Fig. 1. Ultrafiltration apparatus for producing HAP-water. Apparatus Electronic llistruments Ltd. laboratory sodium monitors, Model 8980, were used rm a tests unless otherwise stated, each with a calomel reference electrode (EIL RJ.23) and a saturated potassium chloride salt bridge. For certain tests, the reference half-cell was connected via a salt bridge containing diethylamine solution to eliminate the possibility of leakage of potassium ions into the sample stream.Potentials were measured with either an Orion 801 digital pH meter, a Corning-EEL 110 digital pH meter or a Pye 290 pH meter. The following sodium-responsive glass electrodes were used : EIL GEA.33, Leeds and Northrup 117201, Electrofact N.V. OG572 and Beckman Instruments 39278. A Grant water circulator, Model LC 10, was used to control the temperature of the sample entering the electrode compartments of the monitors. An Amicon Ltd., Model DH 4, cartridge adaptor with an HIDPlO Diafiber hollow-fibre cartridge was used in the preparation of HAP-water. Procedure Deionised water or standard sodium solutions were passed continuously through the flow850 GOODFELLOW et a,?.: FACTORS AFFECTING THE LIMIT OF Analyst, VoZ. 101 cells of the sodium monitors, normally at a flow-rate of 4 ml min-l. The pH of these solutions was adjusted to between 10.3 and 11.5 by the addition of the vapour of ammonia or various volatile amines entrained in a stream of air (flowrate 4 ml min-l). Cyclohexylamine and dimethylamine were added directly to the solutions. The responses of a number of sodium- responsive glass electrodes were noted as the alkaline additive, temperature or flow-rate was changed. The electrodes were first calibrated at sodium concentrations (10-200 pg 1-l) at which a linear relationship was obtained between the e.rr1.f. and the logarithm of the concentration. These calibrations were extrapolated when determining the apparent sodium concentrations in very dilute solutions.Results Comparison of Different Makes of Sodium-responsive Electrodes Two or three electrodes of each type were tested simultaneously while at the same time an EIL GEA.33 electrode, the performance of which was known,lv3 was run on the same solutions and with the same alkaline additive, thus acting as a referee electrode against which the others could be compared. Leeds and Northrup 117201 electrodes Dimethylamine was used as the alkaline additi~e,~ although the manufacturers have since changed their recommendation to diisopropylamine. Although the responses of three of these electrodes were linear over a wide range of sodium concentrations (10 000-10 pg F), they did not give the theoretical slope. The measured sensitivities of the electrodes, 56.6, 53.8 and 54.3 mV for a decade change in concentration, were, however, within the manufacturer's specification of 56 mV & 5%.In a comparative test against the EIL referee electrode, which indicated a sodium concentration of 1.99 pg1-I in a sample of deionised water, the three Leeds and Northrup electrodes indicated 2.00, 2.06 and 2.03 pg 1-l. In a further test with a purer sample of deionised water, both the EIL electrode and one of the Leeds and Northrup electrodes indicated a sodium concentration of 0.5 pg 1-l. Electrofact 06572 electrodes As these electrodes suffer from ammonium-ion interference in ammonia solutions at pH 10.5, equivalent to a sodium concentration of about 50 pg l-l, the manufacturer recommends cyclohexylamine as the alkaline additive.Standard sodium solutions were prepared contain- ing 0.1% V/V of cyclohexylamine, which gave a pH of about 10.3. When the responses to these solutions of the EIL referee electrode and three Electrofact electrodes were measured, the latter gave results in good agreement with the EIL electrode at concentrations between 2 000 and 10 pg l-l, but in deionised water in which the EIL electrode indicated 0.65 pg 1-1 the Electrofact electrodes gave 1.55, 1.66 and 1.15 v g l-l, showing a possible interference from cyclohexylammonium ions equivalent to a sodium concentration of 0.5-1 .O pg 1-l. Beckman 39278 electrodes Beckmans quotes results showing a Nernstian response down to a sodium concentration of 2 300 pug l-l, but states that the electrode can respond non-linearly to sodium concentrations of less than 23Opg1-l. This effect was confirmed: with ammonia as the alkaline additive, the response between 200 and 20 pg 1-1 was only 50.5 mV and from 20 to 2 pg 1-1 only 48 mV instead of the theoretical value of 58.3 mV at 210 "C.In view of the improved sensitivity of other electrodes when alkylamines are used as alkaline additives, the ammonia was replaced with 0.1% V/V of cyclohexylamine. The sensitivity increased (52.6 mV between 200 and 20 pgl-l) but was still markedly inferior to that of the EIL electrode. When the latter indicated a sodium concentration of 0.65 pg 1-1 in deionised water, two Beckman electrodes gave 1.4 and 1.9 pg 1-l. From the difference, we infer that Beckman electrodes are more susceptible to interference from the ionic form of the alkaline additive than EIL electrodes, to the extent that they are not useful for measuirements at levels below 20 pg 1-l.From the above results, it was apparent that EIL and Leeds and Northrup electrodes had generally similar characteristics and were to be preferred to Electrofact and Beckman electrodes for measurements of concentrations below 1 pg 1-1. Further work was concen- trated on the former pair and particularly on the EIL electrode.November, 19’16 DETECTION OF SODIUM-RESPONSIVE GLASS ELECTRODES Estimation of Hydrogen-ion Interference 851 When the pH of deionised water was adjusted to 10.3 by the addition of ammonia vapour, two EIL GEA.33 electrodes indicated apparent sodium concentrations of 1.45 and 1.98 pg 1-1.When the measurements were repeated with the pH reduced to 8.5 by decreasing the amount of ammonia added, the indicated sodium concentrations were 2.01 and 2.90 pg 1-I, respectively; thus, at pH 8.5 the hydrogen-ion interference is equivalent to a sodium concentration between 0.56 and 0.92 pg 1-l. At pH 10.5, the hydrogen-ion concentration is 1% of that at pH 8.5 and its effect should therefore be less than 0.01 pg 1-l. Changes in pH between 10.5 and 11.5 that can occur with different alkaline additives could therefore be considered to have an insignificant effect on the determination of sodium concentrations in the range 0-1-1.0 pg 1-l. Electrode Response with HAP-water in the Presence of Various Amines A number of batches of HAP-water were analysed for sodium with four simultaneously operating EIL GEA.33 electrodes.Each batch was analysed using ammonia, diethylamine and chisopropylamine as the alkaline additive. The indicated sodium concentrations were in the range 0.6-1.5 pg 1-1 when ammonia was used and 0.2-0.4 pg 1-1 less with either of the alkyl- amines. Such decreases are similar to those found by Eckfeldt and Proctor4 for Leeds and Northrup electrodes. The lowest result obtained was 0.4pgl-l, which could be the actual concentration of sodium in the water or the interference effect from various sources, or a combination of both factors. To resolve the different contributions, it was necessary to have purer water than that from the mixed-bed deioniser, and HAP-water, prepared as described above, was therefore used.When diethylamine was the alkaline additive, HAP- water had an apparent sodium concentration of about 0.1 pg l-l, which is close to the minimum values found from electrode measurements in a number of power stations, and which was repeated with many different batches oi HAP-water. When the ultrafiltration step of the HAP-water preparation was omitted, very fine particles of HAP were carried by the water through the pump system and into the flow cell. The apparent sodium content in these circumstances was typically about 0.01 pg 1-1 (with addition of diethylamine), with a lowest indicated value of 0.005 p~ 1-l. Values of 0.01-0.02 pg. 1-1 were obtained when other amincs were used (dimethylamine, isopropylamine, diisopropylamine, triethylamine, piperidine).The decrease in the apparent sodium concentration in the presence of HAP particles could have been caused by absorption of sodium ions or of interfering species. The pH was unchanged by the presence of HAP particles and the contribution of any hydrogen- ion interference would be constant ; moreover, the hydrogen-ion interference calculated above could account for only a fraction of the observed decrease. The constancy of the pH also shows that there was no change in the alkylammonium-ion concentration, nor therefore in any possible interference. In contrast to the above results, when unfiltered HAP-water was dosed with ammonia, the apparent sodium concentration was 0.2 pg 1-l compared with 0.3 pg 1-1 in filtered HAP-water, i.e., the apparent sodium content was reduced by approxi- mately 0.1 pg 1-1 whatever alkaline additive was used, but the difference between readings in water dosed with ammonia and with alkylamines was maintained.The presence of HAP particles therefore has little, if any, effect on the interferences from the above sources. Anomalous potentials are sometimes observed in colloidal solutionss but the results with the ammonia-dosed water in the presence of HAP particles show that this effect cannot account for the corresponding results with the alkylamine additives. As none of the above causes can explain the results in the colloidal HAP solutions, sources of sodium or interfering alkali metal ions were considered. (i) To check that potassium ions from the reference electrode were not diffusing back against the flow of solution, the reference electrode was joined to the flowing solution through a salt bridge consisting of diethylamine solution.As the apparent sodium concentration of filtered HAP-water remained at 0.1 pg l-l, this source of error could be eliminated. (ii) Contamination of the purified water by desorp- tion of alkali metal ions from the walls of the flow system was statistically unlikely in view of the frequency with which the value 0.1 pg 1-1 has been found as the lowest observed concentration by different investigators using different electrodes in different locations. (iii) Alkali metal ions could have been dissolving from the glass membrane itself. This possibility is further considered below.852 GOODFELLOW et d. : FACTORS AFFECTING THE LIMIT OF Analyst, T/'Ol.101 Investigation of Solubility Effects from the Glass Membrane Efect of temperature Using a water-jacketed flow cell at a constant temperature, the EIL electrode system was initially equilibrated with filtered HAP-water flowing at 4 ml min-l. Known additions of sodium were made to the HAP-water by continuously metering controlled amounts of a 26 pg 1-1 solution into the water stream, using different sized pump tubes, with individually calibrated flow-rates, fitted on a Technicon Mark I proportioning pump. The yH of the solution was adjusted to 10.6-11.0 with diethylamine solution. The equilibrium e.m.f. at each level of addition was measured. This test was carried out once at each of four different temperatures with three different sodium additions at each temperature.The concentration of sodium in the HAP-water was determined by Gran's methodlo (Fig. 2) and the results, given in Table I, show that the apparent concentration of sodium in this water increased from 0.05 pg 1-1 at 8 "C to 0.42 pg 1-1 at 45 "C. The results were obtained from increases in the sodium concentration; the equilibrium times required increased from 30 min at 45 "C to 60 min at 8 "C, and an even longer time was required for a subsequent blank solution to give equilibrium readings. At 2 "C the response time was so long that it was not practical to obtain a true equilibrium reading for the HAP-water, but the lowest potential attained indicated a sodium concentration of 0.05 pg 1-1 (calculated by extrapolation of the Nernst slope, not by Gran's method).15 10 I- s 2 X 24 . 5 0 Sodium added/;dg I-' Fig. 2. Gran plots at various temperatures: A, 8 "C; B, 21 "C; C, 32 "C; D, 45 "C. Tests with Leeds and Northrup electrodes showed similar behaviour to that of EIL elec- trodes. When the solution temperature was decreased from 43 to 10 O C , the indicated sodium concentration was reduced from 0.09 to 0.045 pg 1-l. E f e c t of $ow-rate In order to vary the agitation of the solution at the membrane's surface, the flow-rate past an EIL electrode was varied by pumping the solutions with a Technicon Mark I pro- portioning pump with pump tubes of appropriate sizes, instead of the peristaltic pumps used in the EIL monitors. The air carrying the diethylamine vapour was pumped at the same rate as the solution.The results in Table I1 show that higher flow-rates gave lower apparentNovember, 1976 DETECTION OF SODIUM-RESPONSIVE GLASS ELECTRODES TABLE I EFFECT OF TEMPERATURE ON INDICATED SODIUM CONCENTRATION 853 Sodium Temperaturel'C added/pg 1-1 8 0 0.11 0.28 1.10 0.11 0.29 1.10 0.11 0.29 1.10 0.09 0.27 1.13 21 0 32 0 46 0 Calculated* sodium content E.m.f./mV of HAP-waterlpg 1-1 -400.0 0.05 -375.1 - 360.1 - 326.0 -414.6 0.07 - 396.0 - 378.3 -344.8 -420.3 0.12 -407.0 - 395.6 - 360.9 -415.7 0.42 -410.4 -402.0 - 379.9 * Calculated by the method of Gran.lo sodium concentrations, regardless of temperature, which may be attributed to the more rapid removal of the ions dissolving from the glass in the faster stream of solution. The ratio of the indicated sodium concentration in the slow stream to that in the fast stream is greater in the HAP-water than in the deionised water because of the higher sodium content of the latter, The effect cannot be attributed to streaming potentials, because with solutions of concentration 20 pg 1-1 or greater, at which the contribution of any alkali metal ions dissolved from the membrane would be negligible, the measured potentials were independent of flow- rate.Qualitatively similar results have been observed by Eckfeldt and Proctors with Lee& and Northrup electrodes. TABLE I1 EFFECT OF FLOW-RATE AND TEMPERATURE ON INDICATED SODIUM CONCENTRATION Flow-rate/ Sample Ternperature/'C ml min-l Deionised water . . .. 40 0.8 8.0 16 0.8 8.0 HAP-water . . .. 40 0.8 8.0 12 0.8 8.0 Indicated sodium concentration/pg 1-1 1.23 0.59 0.81 0.71 0.66 0.27 0.27 0.10 The following test was carried out in order to check whether any alkali metal ions dissolved from the glass membrane contributed to the apparent sodium concentration through their presence only in a film adjacent to the membrane or whether they were widely dispersed in the bulk of the solution.Four electrodes were arranged in a series of separate compartments in a specially made Perspex flow cell. A reference electrode common to all of the glass electrodes was placed in the last compartment. Deionised water dosed with diethylamine vapour was pumped through the cell at 4mlmin-l and the potential of each electrode measured. Within experimental error, there was no increase in the apparent concentration along the series, indicating that dissolved alkali metal ions occur in significant concentrations only in a small volume of water adjacent to the membrane. Discussion For certain investigations carried out in power stations, e.g., for measurement of steam854 GOODFELLOW et al.: FACTORS AFFECTING THE LIMIT OF Analyst, VoZ. 101 carry-over, it is necessary to determine sodium in water at concentrations in the range 0.01-0.1 pg 1-I. While a number of commercial electrode systems are available, there is now considerable evidence that these systems indicate sodium concentrations that become increasingly inaccurate below true sodium concentrations of 1 pg 1-1 and that they have almost no sensitivity below about 0.1 pg 1-l. Chemical factors such as hydrogen-ion interference, interference from the ionic form of the alkaline additive and alkali metal ions dissolved from the glass membrane can contribute to the background potential that will limit the measurement of very low concentrations of sodium.Experimental evidence in this work suggests that the first two factors together account for no more than 0.005-0.01 pg 1-1 of apparent sodium (observed in the presence of colloidal HAP particles) when diethylamine is used as the alkaline additive. The hydrogen- ion interference was found to be equivalent to less than 0.003-0.01 pg 1-1 at pH 10.5-11.0. An experimental assessment of the diethylamnionj um-ion interference was not practicable, as it was too small to be significant at an accurately known sodium concentration.Rechnitz and Kuglerll obtained a selectivity coefficient of 1.4 x low4 for diethylammonium ions over ammonium ions with a glass electrode with a poor selectivity for sodium (the Beckman 39137 cation-sensitive electrode). Even if it is recognised that their value would probably give an over-estimate of the diethylammonium interference with the EIL electrode, the interference calculated on this basis in a solution dosed to pH 11 with diethylamine is only 2 x pg l-l, i.e., diethylammonium-ion interference is almost certainly negligible at any sodium concen- tration measured so far. The very similar results obtained when other alkylamines were used suggest either that any differences in the selectivities for the various alkylammonium ions have a negligible effect on the e.m.f.observed and that the contribution of the inter- ference by these ions is therefore insignificant or that the electrode is equally selective for all the alkylammonium ions. Work with other electrodes4J1 does not support the latter proposi- tion. Moreover, while the selectivities for hydrogen or alkylammonium ions could change with temperature, such changes are expected to be small12 and there is no reason why they should be affected by the flow-rate. The contribution of these interferences to the apparent sodium concentration near the limit of detection is therefore inferred to be negligible. The experiments on the effect of flow-rate showed that alkali metal ions dissolved from the membrane had a detectable concentration only in a thin layer of solution close to the membrane and determination by another method of the concentration of alkali metal ions to which the electrode was responding was considered impracticable.The evidence for the dissolution mechanism therefore rests on the effect of physical factors such as temperature and flow-rate, neither of which should have affected the level of impurities in the water before it reached the flow cell, and the latter should also make no difference t o the interference by species originally in the solution. The dissolution of alkali metal ions from glass is well known, although most studies have been made at high temperatures,13 and a similar effect has been observed with pH-sensitive glass electrodes in highly pure water.2 The solubility would increase with temperature and therefore raise the level of sodium recorded by the electrode while a higher flow-rate would remove the dissolution products from the region of the membrane more quickly, causing a reduction in the apparent sodium concentration, as reported by Eckfeldt and Proctor.6 It is the 1inea.r rather than the volume flow-rate that is significant for these changes.Eckfeldt14 has pointed out that the hydrogen-ion concen- tration obtained with a given dosage of alkylamine increases with temperature because of changes in the autoprotolysis constant of water and the dissociation constant of the alkyl- ammonium ion. The change in pH between 21 and 45 "C in our experiments would not, by our calculation, increase the apparent sodium content by more than 0.05 pg 1-1 compared with the observed difference of 0.35 pg 1-l. Neither impurities in the water nor interferences by hydrogen or alkylammonium ions can satisfactorily explain the results described here and we conclude that dissolution of alkali metal ions from the glass membrane is the major factor in the deviations of EIL and Leeds and Northrup electrodes from the theoretical response at concentrations below 1 pg 1-l.As both the EIL and Leeds and Northrup electrodes have membranes composed of lithium aluminosilicate glass, the alkali metal ion dissolving from the membrane will be lithium. From the potentials of the EIL electrode in separate sodium and lithium solutions of the same concentration, the selectivity of the electrode for lithium over sodium, KNaLI, is 0.2.The selectivity of the Leeds and Northrup electrode for lithium ions is "only slightly less thanNovember, 1976 DETECTION OF SODIUM-RESPONSIVE GLASS ELECTRODES 855 that for sodium ions.’’6 Although the two glasses showed the same trends, and indicated very similar sodium concentrations a t temperatures of 10-20 “C, the apparent sodium content of HAP-water increased much more for the EIL electrode than for the Leeds and Northrup electrode as the temperature increased. Both the solubility and the lithium selectivity of lithium aluminosilicate glasses used for determining sodium will influence the limit of detec- tion. One glass may therefore be more soluble than another, but still have the lower limit of detection because it is more selective for sodium. As selectivities change little with tem- perature,12 solubility should become relatively more important as the temperature increases.In this light, the EIL electrode appears to be more soluble and less lithium selective than the Leeds and Northrup electrode. The compositions of the glasses tested were not known to us, except in qualitative terms, and we could not attempt to correlate the results with the structure of theglasses. Of the two electrodes that were not suitable for measurements below 1 pgl-l, one, the Beckman 39278, also had a lithium aluminosilicate glass while the other, the Electrofact 06572, had a sodium aluminosilicate glass and thus even a qualitative assessment of the influence of composition was not possible. Conclusions In practical terms, the effect of dissolution of alkali metal ions on the limit of response may be reduced by having a high linear flow-rate past the electrode and a low solution temperature, but the latter has to be reconciled with significant increases in response time.Of the electrodes tested in this work, the EIL GEA.33 and Leeds and Northrup 117201 appeared to be equally the most favourable for determining sodium concentrations below 1 pg l-l, provided that the temperature was about 20-25 “C, when the blank due to dissolution of alkali metal ions from the glass membrane was about 0.07 pg 1-I for both types. Individual electrodes of each type may, however, have different limits of detection. by permission of the Central Electricity Generating Board. This work was performed at the Central Electricity Research Laboratories and is published 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Hawthorn, D., and Ray, N. J., Analyst, 1968, 93, 158. Webber, H. M., and Wilson, A. L., Analyst, 1969, 94, 209. Diggens, A. A., Parker, K., and Webber, H. M., Analyst, 1972, 97, 198. Eckfeldt, E. L., and Proctor, W. E., Analyt. Chem., 1971, 43, 332. Galetti, B. J., and Spear, J. F., Proc. Am. Pwr Conf., 1970, 32, 817. Eckfeldt, E. L., and Proctor, W. E., Analyt. Chem., 1975, 47, 2307. Girardi, F., and Sabbioni, E., J . Radioanalyt. Chem., 1968, 1, 169. Beckman Instructions 81 155-C, Beckman Instruments Inc., Fullerton, Calif., 1971. Bates, R. G., “Determination of pH, Theory and Practice,” Second Edition, Wiley-Interscience, New York and London, 1973. Gran, G., Analyst, 1952, 77, 661. Rechnitz, G. A., and Kugler, G., 2. Analyt. Chem., 1965, 210, 174. Eisenman, G., Adv. Analyt. Chem. Instrum., 1965, 4, 213. Eitel, W., “Silicate Science,” Volume 11, Academic Press, New York and London, 1965. Eckfeldt, E. L., Analyt. Chem., 1975, 47, 2309. Received June 17th, 1976 Accepted July 7th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100848

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

856 Analyst, November, 1976, Vol. 101, pp. 856-859 Colorimetric Method for the Determination of Arsenic(ll1) in Potable Water Shingara S. Sandhu Water Laboratory, South Carolina State College, Orangeburg, South Carolina 291 17, USA A colorimetric method for the determination of arsenic(II1) at a concentration of 0.002 mg 1-1 in potable water was developed. Arsenic(II1) reacts quanti- tatively with potassium iodate in the presence of sulphuric acid and releases an equivalent amount of iodine, which imparts a pink colour to a carbon tetrachloride extract, this colour being measured a t 520 nm. The method was used to differentiate between arsenic(II1) and arsenic(V) . No interference from arsenic(V), chloride, fluoride or nitrate was observed. Natural water samples analysed by this technique showed trace amounts of arsenic(II1).Arsenic is widely distributed in the human environment1 and has attracted considerable attention as a public health hazard. The legal limit for its presence in potable water in the USA has been set at 0.01 mg 1-1 and water supplies containing 0.05 mg 1-1 of arsenic are rejected.2 The levels of arsenic found in US surface waters range from less than 0.01 to more than 0.1 mg l-l., Two forms of arsenic exist in the human environment, pentavalent and trivalent. The biological activities of arsenic in these valence states differ markedly4 but most methods5-7 currently used for the determination of arsenic at low concentrations in water are total elemental analysis procedures. Although some work has been reported on the determination of various chemical forms of arsenic,* the methodl suggested needs elaborate equipment.This paper presents a simple analytical method for the quantitative evaluation of arsenic(II1) in water. The proposed colorimetric method offers a number of advantages over the presently available standard method, using silver diethyldithi~carbamate.~ The colorimetric procedure is sensitive and simple and the results are accurately reproducible at microgram per litre concentration levels. More important, it differentiates between the two valence states of arsenic and will yield a satisfactory interpretation of the potential hazard related to the use of potable water supplies that contain high arsenic concentrations. The procedure does not require any elaborate equipment. A Spectronic 20 spectrophotometer, or similar instrument, can be used for routine determinations.Arsenic( 111) is a reducing agent and reacts quantitatively with i ~ d a t e , ~ releasing an equivalent amount of iodine in the presence of sulphuric acid: 5As033- + 210,- + 2Hf -+ I, + 5As04,- + H20 (E" = 0.64 V) . . (1) The liberated iodine was extracted immediately with carbon tetrachloride. The partition coefficient (K) of 85.5 for the distribution of iodine between carbon tetrachloride and aqueous phaseslO was employed to determine the experimental conditions for the transfer of iodine from the aqueous to the organic phase. Assuming that iodine exists in the same state in the aqueous and organic phases, equation (2) was derived for the total amount of iodine liberated by arsenic(II1) in a one-step (single, liquid - liquid) extraction: - * (2) n, = n2(v1 ++) 0 ... .. where n, is the amount of total iodine liberated {millimoles per litre), n2 is the amount of iodine in the organic phase (millimoles per litre), Vl is the volume of carbon tetrachloride used for extraction, V , is the final volume of the aqueous phase and K is the partition coefficient. Using the stoicheiometry of equation (1) and substituting the value of n2, based on Beer's law, equation (2) becomes Amount of As,+ (mg 1-1) == .. (3) where A is the absorbance at 520 nm and V , is t:he volume of sample used.SANDHU 857 Experiment a1 Reagents and Apparatus Standard precautions for trace analysis were taken throughout the measurements.Ana- lytical-reagent grade chemicals were used, except for carbon tetrachloride, which was of Nanograde quality. A series of solutions containing 0.002-100.0 mg 1-1 of arsenic(II1) were prepared from sodium arsenite. Arsenic(V) standards were prepared from sodium arsenate. A number of standard solutions containing 0.10-10.0 mg 1-1 of arsenic(II1) as well as 0.50- 20.0 mg 1-1 of arsenic(V) were prepared. A few natural waters from drinking fountains, lakes and rivers were also sampled and analysed within 3 h of their collection. Method Standard solutions (50ml) of arsenic(II1) containing 1, 2, 5, 10 and 20mg1-1 of metal cations were transferred in triplicate into 125-ml separating funnels. A pre-determined volume of 18 N sulphuric acid was added to each to adjust the hydrogen-ion concentration to 3 N.The contents of each separating funnel were treated with an excess of potassium iodate (about 200mg) and 1 O m l of carbon tetrachloride. The amount of potassium iodate used in this method could be varied in accordance with the concentration of arsenic(II1) expected in the various aliquots. Fifty millilitres of the 100 mg l-l, 100 ml of the 0.5 mg 1-1 and 150 m.I of the 0.1 mg 1-1 arsenic(II1) standard solutions were treated in a similar manner except that 50, 5 and 3 ml of carbon tetrachloride, respectively, were used for extraction of liberated iodine. Standard solutions containing less than 0.1 mg 1-1 of arsenic(II1) and natural water samples were placed in a Rotavapor, Model 4358-S50, flash evaporator and the volume was reduced from 700 to 20 ml.In a few instances, when found necessary, concentrated natural waters were filtered through Whatman No. 42 filter-paper to remove the insoluble constituents. The reduced volume in each instance was transferred quantitatively into a 125-ml separating funnel and treated as described above for the reaction of arsenic(II1) with potassium iodate. The liberated iodine was extracted with 3 ml of carbon tetrachloride. The contents of each separating funnel were mixed by shaking manually for about 2 min and allowed to stand for 15 min for complete separation of the organic and aqueous phases. The iodine - carbon tetrachloride extracts were directly drained into a matched set of 1.27-cm cuvettes (Bausch and Lomb, Cat. No. 33-29-27) and the absorbance was read on a Spectronic 20 spectrophoto- meter at 520 nm.The total arsenic in the reduced volume of natural water samples and standard solutions was determined with an atomic-absorption spectrophotometer (Perkin-Elmer, Model 306) under standard conditions, using an argon - hydrogen flame and a three-slot burner head. Arsenic(V) was obtained by the difference between total arsenic and arsenic(II1). A number of solutions were prepared for the study of possible interference by arsenic(V), chloride, fluoride and nitrate (which invariably occur in natural waters). The concentration of each of these ions in the solutions was maintained at 100.0 mg 1-1 while the concentration of arsenic(II1) was varied from 0.002 to 10.0 mg 1-l. The prepared standards were treated with iodate as described previously and liberated iodine was extracted with carbon tetra- chloride for colorimetric determination. Results and Discussion The wavelength of maximum absorption for iodine in carbon tetrachloride is slightly below 520 nm.ll The molar absorption coefficient of 726.31 1 mol-l cm-l was obtained at this wavelength.This value of the absorption coefficient was used in equation (3) for calcu- lating the concentration of iodine transferred from the aqueous to the organic phase. The amount of arsenic(II1) in the unknown sample can be obtained either from the graph of absorbance of iodine in carbon tetrachloride veysus the concentration of arsenic(II1) in the standards or by use of equation (3). In this study, the use of equation (3) was found con- venient for precise and rapid calculations of arsenic(II1) concentrations.The results obtained by the suggested colorimetric method were compared with those obtained through the use of an atomic-absorption spectrophotometer. The results in Table I show that the recovery of arsenic(II1) from the standard solutions was quantitative. The average recovery The recovery of arsenic(II1) from standard solutions is shown in Table I.858 SANDHU : COLORIMETRIC METHOD FOR THE AnaJyst, VoL 101 TABLE I RECOVERY OF ARSENIC( 111) BY THE COLORIMETRIC METHOD COMPARED WITH THEORETICAL VALUES AND ATOMIC-ABSORPTION SPECTROPHOTOMETRIC RESULTS Concentration of As(III)/mg 1-1 Theoretical Determined* Recovery, % - 0.005 0.0053 106 0.01 0.0096 96 0.05 0.052 104.0 0.1 0.09 90.0 0.5 0.52 104.0 1.0 0.98 98.0 2.0 2.11 105.0 5.0 5.22 104.0 10.0 10.09 100.9 20.0 21.75 108.7 100.0 103.0 103.0 Mean ... . 102.1 Standard deviation 4.5 Standard deviation, 5.3 2.4 4.3 8.4 4.5 1.2 3.8 3.2 1.4 4.8 3.4 % Recovery by AAS, % - 103.0 104.0 101.0 101.0 99.0 101.0 101.6 1.76 * Mean of three determinations. of arsenic(II1) by the colorimetric procedure was about the same as that by the atomic- absorption spectrophotometric method, except that the former showed a higher standard deviation. The precision and accuracy of the suggested colorimetric technique can be improved considerably by the use of a spectrophotometer of better quality. The analytical method was applied to the differentiation of arsenic(II1) and arsenic(V). The results in Table I1 indicate that the suggested technique determined arsenic(II1) and effectively separated it from arsenic(V).TABLE I1 COMPARISON OF RECOVERY OF As(II1) AND As(V) All concentrations expressed as mg 1-I. Theoretical content of As (I I I) 0.002 0.005 0.10 0.50 1.00 6.00 10.00 Arsenic( 111) r 3 Theoretical Experimental 0.10 0.09 0.50 0.61 1.00 1.05 2.00 1.95 5.00 5.10 10.00 10.07 Total arsenic 0.58 1.52 3.15 6.92 15.33 30.35 Arsenic(V) Theoretical Recovered 0.5 0.49 1 .oo 1.01 2.00 2.10 5.00 5.97 10.00 10.23 20.00 20.28 & TABLE III EFFECT OF VARIOUS IONS ON THE RECOVERY OF ARSENIC(III) BY THE COLORIMETRIC METHOD All concentrations expressed as mg 1-1 Experimental recovery in the presence of 100 mg 1-1 of added ions 7 Arsenic(V) 0.0018 0.0047 0.085 0.48 0.94 6.04 10.12 Chloride 0.0021 0.0049 0.11 0.51 0.96 4.96 10.05 Fluoride 0.0021 0.0052 0.093 0.53 1.02 4.97 10.10 Nitrate 0.0019 0.0048 0.105 0.47 0.98 5.10 10.05 AU fou'r ions 0.0021 0.0048 0.096 0.50 0.96 4.96 9.90 Recovery in the presence of all four ions yo 105 96 96 100 96 99 99 Average .. . . 98.3November, 1976 DETERMINATION OF ARSENIC(III) IN POTABLE WATER 859 The presence of arsenic(V), chloride, fluoride and nitrate, added separately or together to the standard arsenic(II1) solutions, did not have any noticeable interfering effect on the recovery of arsenic(II1). The results presented in Table I11 show that the average recovery of arsenic(II1) in the presence of all four ions, each at 100.0 mg 1-1 concentration, was 98.3%. The method was used for the analysis of natural water samples and the results are given in Table IV.The results indicate that natural waters can be analysed without prior treatment other than reducing their volume by flash evaporation and filtering out the insoluble con- stituents. The drinking waters from the fountains had less arsenic(II1) than the river waters. Arsenic(V) was detected in all of the natural water samples, although the concentrations were below the mandatory limit. Arsenic(II1) was found in 80% of the natural waters analysed. TABLE IV ARSENIC CONTENTS OF POTABLE AND NATURAL WATERS Sample Drinking water fountain, Drinking water fountain, Drinking water fountain, Drinking water fountain, Lake Murry Dam, near Congaree River, Columbia Santee River, Highway42 * ND = not detected. Claflin College South Carolina State College U.S.C., Columbia Voorhees College power house Total arsenic/ mg 1-1 0.0076 0.0080 0.0096 0.0068 0.0083 0.0154 0.0248 Arsenic( 111) - Concentra- Percentage tion/mg 1-1 of total <0.0020 < 26.3 0.0020 25.0 0.0023 27.0 ND* - ND* - 0.0065 42.2 0.0039 15.8 Arsenic (V) & Concentra- Percentage tionlmg 1-1 of total > 0.0056 > 73.7 0.0060 76.0 0.0072 73.0 0.0068 100.0 0.0083 100.0 0.0089 57.8 0.0209 84.2 There are a number of reducing substances, such as thiocyanate and thiosulphate, and cations, such as antimony(III), tin(II), mercury(1) and iron(II), that react quantitatively with iodate, releasing i ~ d i n e .~ These substances will therefore interfere in the application of this technique to the determination of arsenic(II1) in highly polluted waters into which industrial waste is discharged.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Lee, D. H. K., “Metallic Contaminants and Human Health,” Academic Press, New York, 1972, Public Health Service, “Drinking Water Standards,” Publication No. 956, US Government Printing Office, Washington, DC, 1962. Subcommittee on Air and Water Pollution: Water Pollution, 1970 (Part 5). Hearing Before the Subcommittee on Air and Water Pollution of the Committee on Public Works, United States Senate Ninety-first Congress, Second Session US. Government Printing Office, Washington, Hearing Before the Subcommittee on Air and Water Pollution of the Committee on Public Works, United States Senate Ninety-first Congress, Second Session. US Government Printing Office, Washington, American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Standard Methods for the Examination of Water and Wastewater,” Thir- teenth Edition, American Public Health Association, New York, 1971, pp. 662-664. Caldwell, J. S., Lishka, R., and McFarren, E. F., J . Am. Wat. W k s Ass., 1973, 65, 731. Tam, K. C., Envir. Sci. Technol., 1974, 8, 734. Braman, R. S., and Foreback, C. C.. Science, N.Y., 1973, 182, 1247. Skoog, D. A., and West, D. M., “Fundamentals of Analytical Chemistry,” Second Edition, Holt, Rinehart and Winston, Inc., New York, 1969, pp. 438-445. Partington, J. R., “General and Inorganic Chemistry,” Third Edition, Macmillan, London, 1951, Drago, R. S., “Physical Methods in Inorganic Chemistry,” Van Nostrand Reinhold. New York, Received May 5th. 1976 Accepted June 25th, 1976 pp. 158-160. DC, 1970, pp. 1890-1913. Subcommittee on Air and Water Pollution: Water Pollution, 1970 (Part 4). DC, 1970, pp. 1380-1407. pp. 57-60. 1965, pp. 150-168.

ISSN:0003-2654    

DOI:10.1039/AN9760100856

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

860 Analyst, November, 1976, Vol. 101, $9. 860-866 Rapid Spectrophotometric Determination of Cobalt in High-speed Steels Based on the Formation of Tricarbonatocobaltate( I I I)* A. Sanz Medel, A. Cob0 Guzman and J. A. Perez-Bustamante Departamento de Quimica Analitica, Facultad de Ciencias Qarimicas y CSIC, Universidad Complutense, Ciudad Universitaria, Madrid-3, Spain The influence of a number of experimental parameters on the determination of cobalt in high-speed steels has been investigated, taking advantage of the property of the cobalt(II1) ion of forming very selectively a stable green-coloured complex compound with carbonate. As a result, a method has been devised that has been applied to the analysis of six certified standard steel samples with cobalt contents ranging from 3 to 12%.From a statistical treatment of the results it is concluded that the average relative error of the method lies within the range f 1.37 yo, while its relakive standard deviation is 0.67%, thereby indicating the absence of systematic errors. The formation of an intense, green-coloured, soluble species by cobalt(I1) with hydrogen peroxide in an aqueous alkaline hydrogen carbonaie medium was reported by Field as early as 1862.l This reaction is remarkably selective arid sensitive, although surprisingly it is not mentioned in most text-books dealing with qualitative inorganic analysis. Several workers have investigated the chemical composition of the green complex and arrived at conflicting conclusions, although it now seems safe to assume that the formation of the tricarbonato- cobaltate(II1) anion [CO(CO,),]~- accounts for the green complex species, on the basis of spectrophotometric predictions,2 the isolation of solid c o r n p ~ u n d s ~ ~ ~ and carrying out ion- exchange experiment^.^ The formation of the tricarbonatocobaltate(II1) complex ion has been made use of so far in order to determine a number of cobalt(II1) complex compounds by means of redox titri- metric methods that are based on the oxidation of excess of iron(I1) ions or iodide, titrating the excess of iron(I1) with permanganate6 or iodine set free i~dimetrically.~,* So far as we know, the use of this green complex as the basis of a spectrophotometric procedure to determine cobalt has been reported only by A y r e ~ .~ In the present work we have investigated the suitability of this complex for use in the spectrophotometric determination of cobalt in high-speed steels.Experiment a1 Reagents Cobalt(l1) stock solutions (1-10 mg ml-l). These solutions were freshly prepared by dis- solving cobalt(I1) nitrate hexahydrate in water. The stock solutions were titrated com- plexometrically with EDTA, using xylenol orange or murexide as indicators, or alternatively by use of a gravimetric method using quinolin-8-01 as precipitant. Sodium hydrogen carbonate solution, approximately 2 M. Hydrogen peroxide, 3% solution. This solution was prepared by suitable dilution of con- centrated Merck Perhydrol. Instrumentation and Equipment Spectrophotometer. Beckman, Model DU, single-beam. Spectrophotometer cells.TSL, glass, 10-mm light path. Centrifuge. Calibrated $asks, 10 ml. Capable of running at 3 500 rev rnin-l. * Communication presented to the XVIIth Biennial Meeting of the Spanish Royal Society of Physics and Chemistry, Alicante, Spain, September 29th-October 4th, 1975.SANZ MEDEL, COBO GUZMAN AND P~REZ-BUSTAMANTE 861 Investigation of the Conditions of Formation of the Green Tricarbonatocobaltate( 111) Complex The formation of the complex has been carried out by the addition of excess of saturated sodium hydrogen carbonate solution to a cobalt(I1) solution, followed by the addition of hydrogen peroxide solution. The slightly alkaline medium might give rise to the precipitation of rose-coloured cobalt(I1) hydroxide, which will dissolve readily upon the addition of hydro- gen peroxide during formation of the green [Co(CO,),]” species.In some instances the subsequent precipitation of brown cobalt (111) hydroxide from the green solutions was noticed, which has resulted in a number of serious experimental problems. In view of these facts, a detailed experimental study has been carried out in order to establish the best conditions for the formation of the green complex, considering the influence of different parameters and operating conditions. Summing up the results obtained, we have concluded that a quantitative and stable solution of the green cobalt(II1) complex can be obtained when 7 ml of saturated sodium hydrogen carbonate solution are poured over the optimum amount (3 mg) of cobalt(II), followed by the addition of 0.5 ml of Perhydrol (1 + 10) solution, diluting the solution thereafter with water to 10 ml and throughly mixing the contents of the calibrated flask.Increasing the amounts of either sodium hydrogen carbonate or hydrogen peroxide over the stated values did not result in any noticeable effect. The green solutions obtained in this way remained stable for at least 24 h [no sign of the formation of the undesirable brown colloid or suspension of cobalt(II1) hydroxide could be detected], exhibiting two absorption maxima at 440 and 640 nm, with molar absorptivities of 195 and 176 cm2 mmol-l, respectively. The wavelength of 640 nm has been selected for the spectrophotometric section of the determination, despite its lesser sensitivity, because of the greater reproducibility of readings at this wavelength and the lesser extent of interferences compared with that at 440 nm.After suitably establishing the optimum formation conditions the green cobalt (111) complex solutions have been shown to conform to Beer’s law for cobalt concentrations within the range 0.04-0.4 mg rnl-l at 640 nm. Application of the Method to the Determination of Cobalt in Steel As the [CO(CO,),]~- complex species is best formed at pH 8.8 & 0.1 (in a saturated sodium hydrogen carbonate aqueous medium), the precipitation of the less soluble hydroxides of metals accompanying cobalt in the steels can be expected. On the other hand, knowing the general composition of high-speed steels (iron, cobalt, tungsten, chromium, manganese, vanadium and molybdenum), it is of great interest to ascertain the effect of their ions on the green cobalt (111) complex species considered, at the concentration levels corresponding to their normal occurrence in this type of steel. Accordingly, a study was carried out to investigate the influence of different ions on the separation process of cobalt as well as on the final spectrophotometric determination.Separation of Cobalt from Iron As much cobalt was shown to remain adsorbed, as a result of co-precipitation, on the bulky, gel-like precipitate of iron(II1) hydroxide, even after repeated leaching and centri- fugation of the solid using saturated sodium hydrogen carbonate solutions containing hydrogen peroxide, suitable steps were taken to separate the iron and cobalt. Two different separation schemes were investigated in order to carry out the separation; the first was based upon physical precipitation, while the second was an attempt at the removal of interferences through suitable masking. After long and exhaustive experimentation, the first approach had to be abandoned as it led to a non-quantitative recovery of cobalt with poor reproducibility. The separation scheme was based on the total precipitation of iron(II1) hydroxide and partial precipitation of cobalt(I1) hydroxide in the alkaline solution created by the addition of excess of sodium hydrogen carbonate following the neutralisation of any free acidity remaining after dis- solution of the sample. The combined precipitate, in a medium of excess of hydrogen carbonate, was then treated with hydrogen peroxide in an attempt to effect the quantitative transformation of cobalt(I1) [present both as solid basic carbonate and free cobalt(I1) ions] into [Co(CO,) ,I3- while the iron(II1) hydroxide precipitate remained undissolved.862 Analyst, Vol.101 The procedure proved to be tedious, long and unreliable as the extraction of cobalt never attained a quantitative level, exhibiting additionally an undesirable lack of reproducibility, depending on the experimental conditions. In view of these disadvantages, a second approach was attempted, taking care to avoid any precipitation of iron(II1) hydroxide and attempting to eliminate the interference caused by iron by means of suitable masking of the iron(II1) ion. The effect of the addition of tartrate and fluoride ions to the dissolved steel sample prior to the adhtion of hydrogen carbonate was investigated in order to avoid the undesirable precipitation of a significant amount of iron(II1) hydroxide.The presence of tartrate was shown to bring about decolorisation of the green [CO(CO,),]~- complex. However, the addition of increasing amounts of fluoride, in the form of potassium fluoride dihydrate, to accomplish the effective masking of iron(II1) did prove to be very effective as the pre- cipitation of iron(II1) hydroxide could be completely circumvented, while the absorbance of the tricarbonatocobaltate(II1) complex proved to be almost insensitive to the presence of even 20% of potassium fluoride dihydrate. These facts induced us to carry out a more systematic investigation of the masking of iron(II1) with fluoride ions, using synthetic mixtures containing 1 : 20 ratios of cobalt(I1) to iron(II1).On the basis of the results obtained, the following conclusions were drawn: firstly, that a large excess of fluoride has to be added over the theoretical stoicheiometric amount needed to form the hexafluoroferrate(II1) ion, [ FeF,J3-, soluble complex in order to ensure that no precipitation of iron(II1) hydroxide takes place upon the addition of sodium hydrogen carbonate solution ; secondly, a sufficient excess of solid potassium fluoride dihydrate must be used to ensure that the formation of a white precipitate takes place, it being assumed that the white precipitate corresponds to the salt potassium pentafluoroferrate(II1) mono- hydrate, K,[FeF,].H,O, according to the conclusions of Cox and Sharpelo; thirdly, a large excess of sodium hydrogen carbonate is needed in order to accomplish the quantitative formation of the [CO(CO,),]~ complex; and fourthly, when formation of the complex occurs in the presence of the white precipitate the absorbance measurements should be made 5-10 min after its formation (Le., the addition clf hydrogen peroxide) because a gradual decomposition of the complex has been shown to take place, presumably induced by the slow transformation of the white precipitate of K,[FeF,].H,O into iron(II1) hydroxide in the slightly alkaline medium created by the excess of sodium hydrogen carbonate.The kinetics of the decoloration phenomenon have been shown to be faster when dealing with actual steel samples than with synthetic solutions, thereby leading to a worsening of the reproducibility of the absorbance measurements.In view of these conclusions, we have finally modified the operating conditions, proceeding in such a way that the white precipitate formed on the addition of solid potassium fluoride dihydrate in acidic medium is separated by centrifugation from the supernatant solution, which reacts to form the [CO(CO,),]~- complex following the addition of sodium hydrogen carbonate and hydrogen peroxide (see Standard Procedure below). Proceeding in this way results in a very satisfactory final stabilisation of the colour of the complex. A last important observation has been made in connection with the cobalt(I1) solutions (resulting from the acidic dissolution of the steel samples), which, upon lengthy contact with the white precipitate of K,[FeF,].H,O, have been shown to co-precipitate on the solid phase to some extent, as can be seen from the rosy colour that appears on the initially white precipitate after standing for a long time.It is for this reason that we recommend that after centri- fuging the white mixed precipitate it should be properly washed, as soon as possible, with an aqueous solution containing potassium fluoride, sodium hydrogen carbonate and hydrogen peroxide, collecting the washings with the supernatant solution from the initial filtration of the cobalt(I1) solution. SANZ MEDEL et al. : RAPID SPECTROPHOTOMETRIC Effect of Other Ions on the Absorbance of the Green Tricarbonatocobaltate(II1) Complex Because high-speed steels usually contain, in addlition to iron, various amounts of different elements that are present in comparable amounts with cobalt, due consideration must be given to their interference in the absorbance of the green cobaIt(II1) complex.The elements considered in this connection were chromium, vanadium, molybdenum,November, 1976 DETERMINATION OF COBALT IN HIGH-SPEED STEELS 863 manganese, copper and tungsten. Preliminary experiments indicated that 1-2 mg of vana- dium [as vanadium(V) oxide dissolved in saturated sodium hydrogen carbonate solution] , molybdenum [as sodium tetraoxomolybdate(III)] , manganese [as manganese( 111) sulphate] or copper [as copper(I1) sulphate] did not interfere to any noticeable extent in the determination of 2 mg of cobalt.The influence of tungsten was not investigated as it is precipitated at the initial steel dissolution step as WO,.nH,O, which can be readily filtered off. The interferences deriving from the presence of chromium can be of more concern, depending on the ratio of chromium to cobalt. Under the conditions of steel dissolution (using aqua regia) and subsequent treatment (with an alkaline hydrogen carbonate medium) a considerable amount of chromium can exist as the yellow-coloured chromate (CrO,”) ion together with the solution of the final green cobalt(II1) complex. An examination of the corresponding absorption spectra showed that the molar absorptivity of the chromate ion is negligible above 470 nm so that it presumably poses no special problem at the wavelength of 640 nm used to measure the absorbance of the green cobalt (111) complex.A detailed experimental investi- gation of the influence of chromate on the determination of cobalt , carried out on a high-speed steel sample of low chromium content, indicated that at the 2 mg of cobalt level no noticeable interference is produced by the presence of up to 6 mg of chromium (added a s potassium chromate), while a slight increase in the absorbance (at 640 nm) of the green cobalt(II1) complex can be observed on increasing the cobalt to chromium ratio to 1 : 10. When maxi- mum accuracy is required in this type of analysis, the otherwise unimportant chromate interference can be cancelled out by using suitable chromate blanks, provided that the chrom- ium content of the steel is known.If it is not known, it is always possible to determine the blank value of the particular steel sample, proceeding as described under Standard Procedure, below, for the dissolution step but omitting the addition of sodium hydrogen carbonate [in order to prevent the formation of the cobalt(II1) complex], adding instead 0.1 N sodium hydroxide solution until a pH of 8.8 &- 0.2 is reached. Nickel(I1) ions were shown to interfere very seriously with the method, when present in amounts comparable with the amount of cobalt present in the steel samples, because of the complementary absorbance of the green nickel(I1) ions ; however, no attempt has been made to overcome this important interference as no nickel was present in the high-speed steel samples investigated. As a result, it can certainly be concluded that the proposed method is unsuitable for the analysis of nickel-containing steels, unless due attention is paid to the elimination of the interference brought about by nickel( 11) ions. Standard Procedure Dissolution step A number of steel samples (0.5-2 g, containing 20-130 mg of cobalt) are accurately weighed and dissolved in 40 ml of 1 + 1 diluted hydrochloric acid with mild heating.Once the evolution of hydrogen has ceased, concentrated nitric acid is added carefully dropwise (a few drops in all) until no further formation of a yellow froth is observed. The solution is then evaporated until its bulk has been reduced to about 10 ml and 15 ml of aqua regia are added to ensure complete oxidation of the sample.The solution is next evaporated on a hot-plate or sand-bath until it appears syrupy and the precipitate eventually formed digested by means of 40 ml of hot 1 + 10 dilute hydrochloric acid. The solution is filtered (blue ribbon paper) into a 100-ml calibrated flask; by this means any tungsten present is filtered off as tungstic acid. Washing of the retained tungstic acid is carried out a few times using hot 1 + 10 dilute hydrochloric acid, collecting the washings together with the original filtrate, and subsequently diluting to the mark with water. Spectrophotometric determination of cobalt Aliquots (1-2 ml containing 0.4-2.5 mg of cobalt) of the acidic steel solution are pipetted into centrifuge tubes and excess of solid potassium fluoride dihydrate is added until the forma- tion of a white precipitate occurs (the dissolution of the salt can be assisted by stirring with a small glass rod).The suspension is then centrifuged and the supernatant liquor pbetted off into a 10-ml calibrated flask, where it is neutralised (until no further carbon dioxide evolution is observed) by means of careful, small additions of solid sodium hydrogen carbonate,864 Analyst, Vol. 101 adding finally a 50-100-mg excess of hydrogen carbonate. Then 1.0 ml of 3% hydrogen peroxide is added and the contents of the flask are swirled gently to effect development of the green colour. The centrifuged precipitate is washed twice, first with 3 ml and then with 1 ml of a solution obtained by mixing 10ml of saturated sodium hydrogen carbonate solution and 2ml of 3% hydrogen peroxide, dissolving in this mixture 2 g of potassium fluoride dihydrate. The two washing solutions, resulting from dispersion of the precipitate and centrifugation, are collected into the 10-ml flask together with the.initial filtrate, 50-100mg of solid sodium hydrogen carbonate are added and the volume made up to the mark with saturated sodium hydrogen carbonate solution. The absorbance measurements are carried out at 640 nm, 15-20 min after the addition of the hydrogen peroxide, using distilled water as the reference blank and taking care to centri- fuge the solutions before pouring them into the cells in order to eliminate any gaseous bubbles and/or solid sodium hydrogen carbonate in suspension.SANZ MEDEL et al. : RAPID SPECTROPHOTOMETRIC Calibration graph A stock, pure cobalt(I1) solution was prepared by dissolving 3.005 7 g of cobalt(I1) nitrate hexahydrate in water and diluting the solution to 250 ml. The exact cobalt content of the solution was established gravimetrically by precipitation of the cobalt with quinolin-8-01.11 Under the standard conditions described in the pi-eceding paragraphs a straight line was obtained for the calibration graph up to 5 mg of cobalt, the graph exhibiting a very satis- factory reproducibility and being insensitive to the amount of fluoride added (to effect the masking of iron). This linearity corresponded to absorbance values of up to 1.3 units. The calibration graph was adjusted by the least-squares method and was shown to conform to the expression hence Absorbance = [0.265 7 x amount of cobalt (mg per 10 ml)] + 0.003 Amount of cobalt (mg per 10 ml) = 3.762 x absorbance - 0.012 4 The optimum range of the calibration graph was from 0 to 3 mg of cobalt, as for higher concentrations the solutions were found to exhibit a tendency to form a brownish black suspension or precipitate in time, which we assume to be cobalt(II1) hydroxide.Results and Discussion The standard method of analysis as described above was applied to six certified high-speed steel samples, the compositions of which are given in Table I. The results obtained, together with a statistical evaluation, are recorded in Table 11. From a study of this table it can safely be concluded that the new method exhibits very satisfactory characteristics for the purposes outlined; this conclusion is based on the following facts.TABLE I CERTIFIED ANALYTICAL CHARACTERISTICS OF THE HIGH-SPEED STEEL SAMPLES INVESTIGATED Elemental composition, % Steel t- I* 0.183 I I t 0.85 1111 0.755 IVg 0.75 type c P Si Mn Cu Cr V Mo W Traces 0.235 0.419 - 4.753 1.325 0.590 15.88 0.021 0.33 0.30 - 5.03 1.57 0.520 19.61 0.016 0.316 0.290 - 4.223 0.077 0.953 18.63 - 0.53 0.48 - 7.79 3.04 1.50 2.80 - 0.21 0.12 0.12 2.72 1.50 4.61 5.7 - 0.14 0.15 0.059 2.12 2.11 0.07 13.0 1 co 4.935 5.67 4.934 2.9 7.8 11.8 * Artillerie Laboratory, Madrid, Spain. t British Chemical Standards. $ Instituto del Hierro y del Acero (CENIM), Spain. €j National Bureau of Standards, Washington, DC, USA.November, 1976 DETERMINATION OF COBALT I N HIGH-SPEED STEELS 865 The accuracy of the method, considered on the basis of the calculated average relative error, which was shown to fluctuate within the range &1.37% for the six mean values (corresponding to the six steel samples investigated), can be considered reasonable for the type of method involved.The precision of the method, expressed in terms of its relative standard deviation, which is 0.67 yo, shows equally good characteristics. The relative standard deviation was calculated as follows. The errors of the 18 determinations shown in Table I1 were obtained by subtracting the average values in column 5 from the values in column 4, expressing each of these as a percentage error and obtaining the standard deviation of these 18 percentage errors. Maximum precision is obtained when adhering very closely to the optimum measuring time (15-20 min).Steel I I I I1 I1 I1 I11 I11 I11 IVt IVt IVt V V V VI VI VI TABLE I1 RESULTS OBTAINED IN THE ANALYSIS OF THE STEEL SAMPLES BY MEANS OF THE PROPOSED SPECTROPHOTOMETRIC METHOD Sample*/ g 1.036 4 1.425 6 1.488 0 1.616 2 1.034 6 1.491 6 1.889 1 1.908 3 1.652 5 1.380 0 1.869 8 1.380 0 1.803 6 1.568 9 1.729 7 0.729 7 0.923 6 1.032 3 Absorbance 0.274 f 0.005 0.366 f 0.004 0.391 f 0.004 0.497 f 0.002 0.313 f 0.004 0.458 f 0.003 0.507 & 0.005 0.495 f 0.003 0.435 f 0.004 0.436 f 0.003 0.291 f 0.005 0.223 f 0.005 0.374 & 0.004 0.326 f 0.003 0.357 f 0.004 0.454 f 0.003 0.575 f 0.004 0.644 f 0.003 ( 4 3 ) Amount of cobalt, yo 4.91 f 0.09 4.79 f 0.05 4.94 f 0.05 5.74 f 0.03 5.63 f 0.07 5.73 f 0.04 5.01 f 0.05 4.85 f 0.03 4.91 f 0.04 2.94 f 0.02 2.89 f 0.05 2.99 f 0.07 7.73 & 0.08 7.74 f 0.07 7.69 f 0.08 11.62 & 0.08 11.64 f 0.08 11.67 f 0.05 (4s) Amount of cobalt, % - Found (average) 4.88 5.70 4.92 2.94 7.72 11.64 Certified er, % 4.93 - 1.01 5.67 + 0.63 4.93 -0.20 2.9 + 1.37 7.8 - 1.02 11.8 - 1.35 * The filtered acidic solutions were diluted to 100 ml, taking five aliquots from each calibrated flask for analysis.The aliquot size was generally 2.00 ml, except for samples V, for which they were 1 .OO ml. 7 The reported absorbance values have been corrected for the corresponding steel blanks, prepared as indicated in the text to account for the small level of interference brought about by the high absolute and relative (to cobalt) chromium content.As regards the accuracy of the method (see Table 11) it is important to bear in mind the following points. When the cobalt content is very low and the chromium to cobalt ratio is high, i t is advisable to carry out the evaluation of the spectrophotometric steel blank as indicated above and for the reasons discussed. When the cobalt content is very high, it should be noted that correspondingly smaller amounts of samples should be weighed or pipetted from the 100-ml calibrated flask in order to remain within the optimum absorbance range (0.30-0.60) and to avoid the higher concentration region where a tendency has been observed for the green cobalt(II1) complex in solution to decompose, giving rise to turbidity, suspensions or precipitates of cobalt (111) hydroxide. The results obtained under these conditions show a slight negative bias associated with a lower reproducibility. Finally, we consider that when the pertinent precautions are strictly adhered to, the method can be considered to be a very simple, fast, selective, precise, reasonably accurate and very cheap means of carrying out the determination of cobalt in high-speed steels a t the semi- micro level.866 1. 2. 3. 4. 6 . 6. 7. 8. 9. 10. 11. SANZ MEDEL, COBO GUZMAN AND PBREZ-BUSTAMANTE References Field, F., J. Chem. Soc., 1862, 14, 61. Mori, M., and Shibata, M., J . Chem. Soc. Japan, Pure Chem. Sect., 1954, 75, 1044. McCutcheon, T. P., and Schuele, W. J., J. Am. Chew. Soc., 1953, 75, 1845. Mori, M., Shibata, M., Kyuno, E., and Adachi, T., Bull. Chem. Soc. Japan, 1956, 29, 883. Hummelstedt, L., Epstein, M., Huggare, T.-L., and Relander, J., Acta Acad. Abo. Ser. B, Mat. Job, A., Ann. Chim. Phys., 1900, 20, 214. Metzl, G., 2. Analyt. Chem., 1914, 53, 637. Laitinen, H. A., and Burdett, L. W., Analyt. Chem., 1951, 23, 1268. Ayres, G. H., Rep. New Engl. Ass. Chem. Teachers, 1941, 42, 143. Cox, B., and Sharpe, A. G., J . Chem. Soc., 1954, 179:s. Erdey, L., “Gravimetric Analysis, ” Volume 2, Perga:mon Press, Oxford, 1966. Phys., 1965, 25, 1. Received November 21st, 1976 Amended April 30th, 1976 Accepted June 18th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100860

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Anahst, November, 1976, Vol. 101, $9. 867-869 867 Micro-determination of Pyrrole Derivatives with A/-Bromosuccinimide in Acetic Acid Medium J. P. Sharma," V. K. S. Shuklat and A. K. Dubey Department of Chemistry, University of Allahabad, Allahabad-2, India A simple micro-procedure for the determination of pyrrole derivatives with N-bromosuccinimide is described. A 2-1 O-mg sample dissolved in acetic acid is subjected to reaction with a known excess of N-bromosuccinimide a t room temperature and the excess of reagent is back-titrated iodimetrically. The maximum error in the results is &1.44%. It has been shown that N-bromosu~cinimide~-~ is stable, reactive and convenient for the routine micro-determination of various organic compounds, their reaction with N-bromo- succinimide being at least 25 times faster than that with br~mine.~ It has been used6,' as a brominating agent for phenol and its derivatives and as a highly selective 0xidant.8,~ There appears to be no report in the literature concerning the reaction between N-bromosuccinimide and pyrrole derivatives in polar media. A rapid and convenient method is described for the determination of pyrrole derivatives on the micro-scale using this reagent.Stoicheiometry of the Reaction The stoicheiometry of the reaction with pyrrole derivatives was determined by dissolving a sample of 2-10 mg in glacial acetic acid and allowing it to react with a measured excess of N-bromosuccinimide solution. The reaction was allowed to proceed for 15 min at room temperature, after which the excess of reagent was determined iodimetrically.The results obtained are shown in Table I. TABLE I DETERMINATION OF THE STOICHEIOMETRY OF THE REACTION Mass of sample Compound takenlmg Indole .. .. .. 1.976 3.952 Carbazole . . .. .. 1.884 3.768 Isatin .. .. .. 3.658 5.487 Indol-3-ylacetic acid . . 2.018 6.054 Indol-3-ylpropionic acid . . 2.812 5.624 6.03 Indol-3-ylbutyric acid . . 2.01 * NBS = N-bromosuccinimide. Number of moles of NBS* consumed 3.106 3.115 2.994 3.001 1.002 1.052 3.102 2.993 3.001 2.991 3.000 3.103 per mole of the compound Experimental Reagents N-Bromosuccinimide, approximately 0.02 M . An amount of 0.356 g of N-bromosuccinimide was weighed out and dissolved in the minimum amount of hot distilled water. The solution was then diluted to 100 ml with cold distilled water in a calibrated flask and was standardised iodimetrically.10 The solution should be freshly prepared before use.Sample solwtions. Stock solutions of each sample were prepared by dissolving an accu- * Present address : Burnsides Research Laboratory, University of Illinois, Urbana, Ill., USA. t Present address : Roskilde Universitetscenter, Hus 161, P.O. Box 260, 4000 Roskilde, Denmark.868 SHARMA et a,!. : MICRO-DETEKMII'?ATION OF PYRROLE Analyst, VOZ. I01 rately weighed amount of analytical-reagent grade material in glacial acetic acid in a 100-ml calibrated flask. Different aliquots of the solutions were used in the determinations. Acetic acid, glacial. AnalaR. Potassium iodide solution, 15% m/V. Baker analysed reagent. Starch solzttion, 1% m/V.Sodium tlziosulphate solution, 0.02 N. Prepared by dissolving 4.964 g of sodium thio- sulphate (AnalaR) in 1 1 of distilled water and standardised before use. Procedure An aliquot containing 2-10mg of the sample was placed in a 100-ml flask and 2ml of glacial acetic acid were added, followed by 10 ml of A'-bromosuccinimide solution. The flask was stoppered and shaken thoroughly and the reaction was allowed to proceed for 15 min at room temperature. The stopper was then washed with 5 ml of distilled water and 5 ml of potassium iodide solution were added to the flask. The liberated iodine was titrated with standard sodium thiosulphate solution, using starch as indicator. A blank experiment was run under identical conditions. All of the determinations were carried out in triplicate. Calculation Relative molecular mass of sample 50 0 6 z g of sample 1 ml of 0.02 M NBS = where n = number of moles of N-bromosuccinimide consumed per mole of the sample.Results and Discussion The results of the determination of several pyrrole derivatives are given in Table 11. The results are moderately precise and the maximum error is about &1.44yo. The mean and standard deviations have also been calculated and are presented in Table 11. Acetic a ~ i d l , ~ has been used as solvent medium for N-bromosuccinimide in many determinations and N-bromosuccinimide in a polar medium can act as a good source of molecular bromine. TABLE I1 MICRO-DETERMINATION OF PYRROLE DERIVATIVES WITH N-BROMOSUCCINIMIDE I N ACETIC ACID ME;DIUM Compound Indole .. .. Carbazole . . .. Isatin . . .. Indol-3-ylacetic acid Indol-3-ylpropionic acid Indol-3-ylbutyric acid . . Sample masslmg Taken Recovered 1.976 1.957 2.964 2.980 3.952 4.009 5.928 6.051 1.884 1.866 3.768 3.757 5.652 5.598 7.536 7.469 1.829 1.839 3.658 3.678 5.487 5.440 7.316 7.279 2.018 2.033 4.036 4.014 6.054 5.995 8.072 8.081 2.812 2.815 5.624 5.574 7.030 6.980 8.436 8.389 2.01 2.01 4.02 4.02 6.03 6.07 8.04 8.04 Error, % + 0.53 + 1.44 +0.39 - 0.96 - 0.95 - 0.29 -0.95 - 0.88 + 0.54 + 0.54 - 0.87 - 0.50 +0.73 -0.11 - 0.97 +0.11 +0.10 - 0.89 -0.71 -0.55 0.00 0.00 $0.99 0.00 Mean deviation * 0.02 *o.oo f 0.00 kO.00 f0.02 f 0.01 f 0.01 50.02 f 0.03 f 0.03 5 0.03 -10.03 50.01 f 0.02 f 0.02 & 0.02 & 0.03 & 0.02 *O.Ol & 0.03 & 0.03 & 0.03 k0.06 & 0.06 Standard deviation 0.33 0.01 0.06 0.09 0.01 0.01 0.01 0.38 0.54 0.54 0.53 0.53 0.10 0.36 0.36 0.36 0.44 0.39 0.42 0.02 0.42 0.42 0.48 0.48November, 1976 DERIVATIVES WITH N-BROMOSUCCINIMIDE IN ACETIC ACID 869 The effect of the concentration of N-bromosuccinimide has also been studied and it was found that a 0.02 M concentration was the best, as higher concentrations of the reagent lead to high results, perhaps caused by loss of bromine.The reaction of N-bromosuccinimide with small amounts of pyrrole derivatives is not instantaneous and it was found that 15 min were required for completion of the reaction. The advantage of the present method over others lies in the uniform reaction temperature and time, which are suitable for all of the compounds studied. The authors express their thanks to Professor R. D. Tiwari for his interest in the investi- gation and supply of materials. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. References Barakat, M. Z., El-Wahab, M. F. A., and El-Sadr, M. M., Analyt. Chem., 1966, 27, 536. Barakat, M. Z., and Shaker, M., Analyst, 1963, 88. 59. Mathur, N. K., and Narang. C. K., “Determination of Organic Compounds with N-Bromosuccin- Shukla, V. K. S., Pande, U. C., and Sharma, J . P., Mikrochim. Actu, 1972, 522. Ross, S. D., Finkelstein, M. F., and Petersen, R. C., J. Am. Chem. Soc., 1968, 80, 4327. Trischler, F., and Szivos, K., Magy. Kem. Foly., 1966, 72, 203. Trischler, F., and Szivos, K., Mugy. Kem. Foly., 1966, 72, 322. Filler, R., Chem. Rev., 1963, 62, 21. Tiwari, R. D., and Pande, U. C., Analyst, 1969, 94, 813. Barakat, M. Z., and El-Wahab, M. F. A., Analyt. Chem., 1964, 26, 1973. imide and Allied Reagents,” Academic Press, London, 1976. Received March Sth, 1976 Accepted July 9th 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100867

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

870 Awlyst, November, 1976, VoL. 101, pp. 870- 875 Losses of Trace Metals During the Ashing of Biological Materials* S. R. Koirtyohann Environmental Trace Substances Research Center and Department of Biochemistry, University of Missouri, Columbia, Mo. 65201, USA and Carole A. Hopkinst Department of Biochemistry, University of Missouri, Columbia, Mo. 65201, USA Losses of chromium, iron, zinc and cadmium during ashing were studied using tissues that contained endogenously incorporated radio;isotopes. No losses on drying at 110 "C and no volatility losses were detected for any of the elements at temperatures below 600 "C. Chromium was lost from blood but not from liver samples heated a t 700 "C. Losses as insoluble material on the crucible surfaces were more significant. Up to 42% of the isotope was retained on the dish after dissolution with acid, the amount depending on the ashing temperature, crucible material and surface condition.No evidence for the formation of volatile compounds from the endogenously incorporated iso- tope was found. Suspected losses of trace metals during ashing has been a subject for debate among analysts for many years. Some of the difficdties in measuring small losses have been overcome through the use of radioactive tracer techniques, and the work of Gorsuch1p2 is probably the most definitive so far reported on this topic. Even with tracers, however, legitimate questions can be raised concerning differences in chemical form between added and endogenous trace metals. This concern was given added support by the work of Strohal et U Z ., ~ who reported losses of cerium, cobalt, manganese, protactinium, ruthenium and zinc from molluscs. In their experiments, the organisms were grown in wa.ter containing the radioactive isotopes of the elements in question, and the tracers were incorporated into the tissues by normal or near-normal biological processes. Significant losses of several of the elements were found after heating at temperatures as low as 110 "C. It was suggested that organometallic com- pounds in the tissues were being vaporised. The possibility that a similar mechanism for the loss of trace metals might occur in other organism!; led to the experiments reported here. Experimental Reagents All radioisotopes were obtained from International Chemical and Nuclear Corp., Chemical Radioisotope Division, Irvine, Calif., USA.The isotopes used, with the chemical forms and specific activities, are given in Table I. All other reagents were of analytical-reagent grade. Ternary acid was made in the laboratory by combining concentrated nitric, perchloric and sulphuric acids in the proportions 10 + 4 + 2. TABLE I ISOTOPE DATA Isotope Chemical form Specific activity/mCi mg-1 W r CrCI, in 0.6 N HCl > 100 5QFe FeCI, in 0.5 N HCl* 25 65Zn ZiiC1, in 0.1 N HC1 > 10 looCd See text * 59Fe was complexed with citrate for administration. * Contribution from the Missouri Agricultural Experiment Station. t Present address : Veterans Administration Hospital, Columbia, Mo. 66201, USA.KOIRTYOHANN AND HOPKINS 871 Apparatus All radioactivity was measured on a Radiation Instrument Development Laboratory, Model 34-12, 400-channel analyser equipped with a Harshaw 10-cm sodium iodide (thallium) scintillation crystal.The counts were summed over the gamma peak for the isotope measured. Crucibles to be counted were placed about 10 cm from the crystal and were carefully posi- tioned so as to ensure a similar geometry for all counts. A Cenco-Cooley electric furnace (Catalogue No. 13633D) was used for dry ashing. Power was supplied to the furnace through a variable transformer for temperature control. At the setting used for most of the work, the temperature of the furnace increased from room temperature to 200 "C in 2.5 h, 300 "C after 4 h, 400 "C after 7 h and 500 "C after 12 h. A LabConCo six-element micro-Kj eldahl digestion rack and 100-ml Kjeldahl flasks were used for wet ashing.Platinum, porcelain and Vycor (96% silica) crucibles were used for dry ashing. In some experiments, an attempt was made to collect the vapour emitted during wet ashing, and in this instance Kjeldahl flasks were fitted with 19/38 standard taper ground-glass connec- tions so as to permit the use of a bent glass tube to carry the vapour from the flask into an ice - water trap. Procedure The animals used in this research were young, male white rats, each weighing approximately 200 g. Radioactive tracers were administered to the animals in normal saline solution by means of an oesophageal - gastric catheter or intraperitoneal injection. Half of the animals in each group were sacrificed 72 h and the other half 1 week after administration of the isotope and samples of blood and liver were collected.If the samples were not to be processed immediately, they were frozen for storage. The liver samples were homogenised in a ground- glass tissue blender prior to removal of a sample for ashing. For dry ashing experiments, about 2g of tissue were placed in a crucible and the radioactivity of the wet sample was determined. The samples were then dried overnight (16 h) at 110 "C and re-counted. The dried samples were then placed in the cold furnace and the temperature was increased slowly to 500 "C as indicated above. The total time in the furnace was 16 h. If the ashing behaviour at a higher temperature was to be studied, the material was counted again and the samples were returned to the furnace for an additional 16 h.The initial temperature for the second ashing was 500 "C and the final temperature 700 "C. After the final ashing, whether at 500 or 700 O C , the ash was dissolved in a solution containing 0.2 ml of concentrated nitric acid and 1.8 ml of water in order more nearly to duplicate the counting geometry of the original wet sample. The test for retention of the isotope on the crucible was carried out by heating the ash for 15 min with 3 ml of 6 N hydrochloric acid. A 10-ml volume of water was then added and the heating continued for 10 min. The solution was transferred and the crucible thoroughly rinsed with water. Any material that remained on the surface was measured by an additional count on the "empty" crucible.A standard was prepared for each metal studied by drying an aliquot of the radioisotope in a platinum crucible. This standard was counted daily and was used to correct for decay and also for possible changes in the counting geometry of the samples or variation in the counter response. Great care was taken to maintain the same counting geometry a t every step in the procedure, as small changes in the geometry result in a significant error in the results, even when dealing with a crystal as large as that used in these experiments. In some instances, the amount of radioisotope retained in the tissues was rather small, and counting times of up to 30min were used in order to improve the precision. Daily background counts were also taken. Counting errors based on the total counts under the photo-peak and on the corresponding background were calculated by using standard equa- t i o n ~ .~ Calculations of the ashing losses involved the ratio of the counts before and after treatment, and the methods of Chase and Rabinowitz5 were used to estimate the error in this ratio. Samples to be wet ashed were prepared and weighed into crucibles (platinum only) in the same manner as for dry ashing. The fresh samples were counted, dried overnight at 110 "C and re-counted. The dried samples were transferred into 100-ml Kjeldahl flasks with872 KOIRTYOHANN AND HOPKINS : LOSSES OF TRACE METALS Analyst, VoZ. 101 about 10 ml of concentrated nitric acid. The contents of the flasks were boiled until all material had dissolved, then 5ml of ternary acid were added and the samples were ashed until the remaining solution was colourless.The necks of the flasks were fitted with bent glass tubes, the open ends of which were immersed in beakers containing 30 ml of cold (0 "C) distilled water as a trap for any volatilised substances. After ashing was complete, the samples were transferred from the flasks back into the original platinum crucibles, and both the ashed samples and the water traps were counted. Zinc Eight rats received 10 pCi of zinc-65 by intraperitoneal injection and eight by oesophageal - gastric catheter. Two control animals received 110 radioisotope. Recoveries of zinc from liver tissue and blood were investigated for samples dry ashed a t 500 and 700 "C. Recoveries from wet-ashed samples were also examined.Retentions on the surface of the ashing vessel were studied for crucibles of the following types: platinum, new and etched porcelain and new and etched Vycor. The zinc-65 photo-peak occurs at 1.11 MeV and was integrated by summing the counts in the peak channel and the ten channels on each side of the peak. Ivon Six animals were each given intraperitoneal injections of 10 pCi of iron-59. Two animals used as controls received no treatment. The iron was injected in a saline solution with sufficient citric acid present to complex all of the iron xnd render it more readily metabolised. Recoveries of iron from blood and liver tissue after dry ashing at 500 and 700 "C were. studied, and also recoveries after wet ashing. Retentions on the surface of Vycor, porcelain and platinum were studied.The iron-59 photo-peaks occur at 1 100 and 1290 keV and were integrated by summing the peak channels and five channels on each side of each peak. Chromium Four animals were used for the purpose of obtaining isotopically labelled tissue. Each was given 100 pCi of chromium-51 by intraperitoneal injection. (A previous experiment had shown that lower doses were excreted too rapidly for sufficient take-up by the organs studied in this work.) Losses of chromium from liver tissue and blood were studied for samples that were dry ashed at 500 and 700 "C and also for samples that were wet ashed. Platinum crucibles were used exclusively as vessels for dry ashing, so the only retention studies made were on platinum at 700 "C.The chromium-51 photo-peak occurs at 320 keV and was integrated by summing the counts in the peak channel and the ten channels on each side of the peak. Cadmium Although no animals were given radioactive cadmium in this work, we were fortunate to obtain a sample of kidney and a sample of liver from a rat that had been given cadmium-109 by intramuscular injection about 1 week prior to sacrifice. The samples were left over from previous experiments, but they still contained readily measurable activity because of the long half-life of cadmium-109. The results are included here in spite of the fact that the conditions of the original experiment cannot be specified in detail. These tissues were dry ashed at 500 and 600°C. One sample of liver homogenate was examined for loss after a total ashing time of 109 h at 600 "C.Retention was tested on plati- num dishes at 500 and 600 "C and on new and etched Vycor at 500 "C. The cadmium-109 photo-peak occurs at 87 keV and was integrated by summing the peak channel, the four channels on the low energy side of the peak and the ten channels towards higher energies. Results and Discussion The assumption was made that the isotopes administered to the animals were metabolised in a normal or nearly normal way in the time period $allotted between administration of theNovember, 1976 DURING THE ASHING OF BIOLOGICAL MATERIALS 873 isotope and sacrifice of the animal. The validity of this assumption was not tested experi- mentally, and there is the possibility that the isotopes were metabolised in an abnormal manner.However, no differences were found in the behaviour of the isotopes from tissues collected 72 h after administration compared with those collected 1 week after administration. Isotopes administered in this way are much more likely to behave as normal tissue com- ponents than those added to samples after co1lection.l Zinc The results from a large number of experiments involving dry ashing of liver tissues and blood containing zinc-65 are summarised in Table 11. No significant losses were encountered on drying at 110 "C, on ashing overnight at 500 "C or even on ashing at 700 "C, a temperature well above that normally recommended for biological samples. The only losses occurred by retention on the dish. Porcelain, a material that is not usually recommended for use in trace analysis, gave the greatest retention.The condition of the surface and the temperature are important factors affecting retention. The loss of 6.7% on to the surface of etched Vycor after ashing blood at 500 "C is probably the most significant finding to the practising analyst. Vycor crucibles are frequently recommended for ashing, and they nearly always become etched after repeated use. The possibility of retention as acid-insoluble compounds on the crucible surface must be kept in mind. TABLE I1 RECOVERY OF ZINC-65 FROM LIVER AND BLOOD SAMPLES AFTER DRY ASHING Avesage recovery after No. of Crucible drying at Sample samples material 110 "C, % Liver 21 Platinum 100.0 8 Etched porcelain - 16 Platinum - 8 Etched porcelain - Average Average Average retention on recovery after recovery after the crucible at ashing at ashing at indicated 600 "C, % 700 "C, % temperature, % 99.9 - - 98.9 - - - 101.4 0.26 (700 "C) - 98.7 5.1 (700 "C) Blood 3 Platinum 103.4 97.6 - 0.94 (600 "C) I) New porcelain 104.2 100.9 - 1.63 (600 "C) 42.6 (600°C) 3 New Vycor 96.7 97.8 - 1.29 (500 "C) 6.7 (600 "C) 99.0 16.7 (700 "C) 4 Etched porcelain 100.6 100.2 - 4 Etched Vycor 103.6 104.8 - 4 New porcelain - - It should be emphasised that recoveries reported in other columns do not indicate the presence or absence of retention on the dish as the ash was counted in the dish, and all activity was included without regard to solubility behaviour.Wet-ashing experiments gave an average recovery of 100.7% for zinc-65 (nine samples) with no appreciable amounts of activity recovered from the traps.Iron The results from the dry ashing of blood and liver samples containing iron-59 are summarised in Table 111. There is no indication of volatility loss from heating at 110,500 or 700 "C. Some loss by retention on the dish was encountered if blood samples were heated to 700 "C in platinum or at 500 "C in etched porcelain. The apparent recoveries from blood ashed in porcelain and Vycor, which are significantly above lOOyo, were probably due to differences in counting geometry. Chromium The results from dry ashing blood and liver containing chromium-51 are summarised in Table IV. There was considerable variation in the chromium retained in the tissues of in- dividual animals and therefore some samples with low activity gave larger counting errors.Of the nine liver samples used for ashing at 500 "C, two gave errors in excess of 10% in the Wet ashing gave complete recovery of iron-59 from both blood and liver.874 KOIRTYOHANN AND HOPKINS: LOSSES OF TRACE METALS Analyst, 'vd. 101 TABLE I11 RECOVERIES OF IRON-59 FROM LIVER AND IBLOOD SAMPLES AFTER DRY ASHING No. of Sample samples Liver 11 2 2 Blood 15 4 2 2 Average Average Average Average retention on recovery after recovery after recovery after the crucible at Crucible drying at ashing a t ashing at indicated material 110"C, % 500 "C, % 700 "C, % temperature, % Platinum 101.0 100.0 - 0.02 (600 "C) Vycor - 99.5 - 0.06 (500 "C) Porcelain - 100.8 - 0.16 (500 "C) Platinum 98.7 99.6 - 1.08 (500 "C) Platinum - - 99.6 8.55 (700 "C) Vycor - 106.0 I 0.0 (500 "C) 2.8 (6OOOC) Etched porcelain - 110.6 - ratio between fresh and ashed material.If these two are excluded, the average recovery for drying and for ashing at 500 "C is improved. LO!;s of chromium-51 from blood was not serious on drying or on ashing at 500 "C, but more than half was lost on heating at 700 "C. A small amount of retention from both liver and blood was seen on the platinum dish. Large variations between individual dishes were encountered, with a total range of 0-12% retained. There was no obvious reason for the variation. TABLE IV RECOVERY OF CHROMIUM-61 FROM LIVER AND BLOOD SAMPLES AFTER DRY ASHING IN PLATINUM DISHES Average recovery Average recovery Average recovery Average retention on No.of after drying at after ashing at after ashing at the dish at indicated Sample samples 110 "C. % 600 "C, 700 "C, yo temperature, % Liver 9 97.1 93.9 97.8 4.4 (700 "C) 7* 99.81 97.7* 97.2* 1.8* (700 "C) Blood 5 95.0 96.0 48.7 4.1 (700 "C) * Average recovery if two samples with counting error in excess of 10% are excluded. Wet ashing gave average recoveries of chromium-51 of 99.5% for liver and 99.1% for blood. Recently, Jones et aLg reported no loss of chromiunn-51 from brewers' yeast from dry ashing at temperatures up to 800 "C. Cadmium The recoveries of cadmium-109 from liver and kidney samples are summarised in Table V. No evidence of volatility loss was seen at 500 or 600 "C under the ashing conditions used, even if the heating period was greatly extended.A significant retention on an etched Vycor surface was observed. Conclusions The results gave no evidence of loss of endogenously incorporated trace metals from tissue samples at low temperatures similar to those reported by Strohal et aZ.3 The only evidence of volatility loss was for chromium from blood when heated at 700 "C. If the radioactive tracers were incorporated into tissues in any volatile form they were decomposed below the temperature at which they would be vaporised by the treatments used. The fact that all samples were placed in a cold furnace and the temperature was increased slowly over a period of several hours may be significant in this respect. Loss of trace metals by retention in an acid-insoluble form in the ashing vessel was en- countered much more frequently.The container material, the condition of the surface, the type of tissue and the ashing temperature are the important variables that affect this loss; the surface condition is the most difficult to control.November, 1976 DURING THE ASHING OF BIOLOGICAL MATERIALS TABLE V RECOVERIES OF C A D M I U M - ~ ~ ~ FROM KIDNEY AND LIVER TISSUES AFTER DRY ASHING 875 Aver age recovery after No. of Crucible drying a t Sample samples material 110 “C, yo Liver 7 Platinum 99.1 4 Etched Vycor - 2 New Vycor - 6 Platinum - Kidney 2 Platinum 98.9 Average recovery after ashing a t 600 “C, % 98.0 102.2 99.6 96.6 - Average Average retention on recovery after the crucible a t ashing at indicated 600 “C, % temperature, % - (0.2 (500 “C) - 9.4* (500 “C) - 0.24 (600 “C) 98.4t 0.12 (600 “C) - (0.2 (600 “C) * Range 1.4-14.4% retention. t One sample from this group was held a t 600 “C for a total of 109 h ; the recovery was 97.2%. The authors are grateful to Dr. E. R. Graham for the use of the counting equipment and for advice on counting procedures, to Dr. and Mrs. H. M. Perry for supplying the tissues containing cadmium-109, to Dr. B. L. O’Dell for assistance with the animal experiments and to Dr. E. E. Pickett for numerous helpful suggestions. References 1. 2. 3. 4. 5. 6. Gorsuch, T. T., Analyst, 1959, 84, 136. Gorsuch, T. T., “The Destruction of Organic Matter,” Pergamon Press, New York, 1970. Strohal, P., LuliC, S., and JelisavW, O., Analyst, 1969, 94, 678. Friedlander, G., and Kennedy, J. W., “Nuclear and Radiochemistry,” John Wiley & Sons, New Chase, G. D., and Rabinowitz, J. L., “Principles of Radioisotope Methodology,” Burgess Publishing Jones, G. B., Buckley, R. A., and Chandler, C . S . , Analytica Chim. Acta, 1975, 80, 389. York, 1949. Co., Minneapolis, Minn., 1968. Received April 22nd, 1976 Accepted May 17th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100870

出版商:RSC    

年代:1976

数据来源: RSC