摘要:

JUNE, 1963 THE ANALYST Vol. 88 No. 1047 PROCEEDINGS OF THE SOCIETY FOR ANALYTICAL CHEMISTRY NEW MEMBERS ORDINARY MEMBERS Norman Simpson Angus, B.Sc., Ph.D. (Q.U.B.) , M.Sc. (T.C.D.) ; Giacomo Bionda, L.C. (Torino) ; David Burns, B.Sc.(Lond.), F.R.I.C. ; Keith Angus Catto, jun., M.S.(Arkansas) ; Peter John Gilbert Dawson, L.R.I.C. ; David John Folkes, B.Sc. (Rristol) ; David John Nicholas Hossack, B.Sc.(Dunelm.), M.I.Bio1. ; Kenneth John Hunter, B.Sc., Ph.D.(Lond.) ; John Richard Jarratt, M.Sc. (Lond.), A.R.I.C. ; Robert Joseph Julietti, B.Sc.(Lond.) ; Cherry Maynard King, B.Sc.(Lond.) ; Laurance A. Knecht, B.S.(Illinois), Ph.D.(Minnesota) ; Peter James Long, B.Sc. (Southampton) ; Eric Leo McCafferty ; Harry Arthur Charles Montgomery, BSc. (Lond.), Ph.D. (Birm.) ; William J.Moore, B.S.(Albright) ; Michael Stephen Moss, M.Sc.(Manc.), F.R.I.C. ; Tomoyuki Mukoyama, Dr.Eng.(Nagoya) ; Alan Robert Oliver, B.Sc.(Leeds) ; Cyril Rainbow, R.Sc., Ph.D. (Birm.), F.R.I.C. ; Peter Thomas Sydney Sandon, A.R.I.C. ; Bryan John Simons, B.Sc.(Bristol), Gerald I. Spielholtz, B.S.(N.Y.), M.S.(Michigan), Ph.D.(Iowa) ; Brian Surfleet, B.Sc., Ph.D.(Sheff .) ; Anna Damita Szczepanowska, B.Sc.(Lond.) ; Barclay Thorpe Whitham, B.Sc. (Liv.), F.R.I.C., F.1nst .Pet. ; George Lawrence Willey, B.Sc., Ph.D., M.P.S. ; Wendy Diane Sylvia Wooldridge, B.Sc.(Lond.), A.R.C.S. JUNIOR MEMBERS Carolyn Joyce Ashby; John Alan Baker; Arnold Morris Dean; Donald Graham Devey; Donald Stuart Goodwin; Jane Elizabeth Grimsley, B.Sc.(Birm.) ; Nicholas Kamm ; Barbara Joy Loxton; P.Jumah Madati, B.Sc.(Lond.) ; George Wilkins. DEATHS Frank William Greaves Louis Francis McCallum Alec Duncan Mitchell. WE record with regret the deaths of NORTH OF ENGLAND SECTION AND BIOLOGICAL METHODS GROUP A JOINT Meeting of the North of England Section and the Biological Methods Group was held at 6 p.m. on Friday, April 5th, 1963, at the Lecture Theatre, Evans Medical Research Laboratories, Speke, Liverpool 24. The Chair was taken by the Chairman of the North of England Section, Mr. C. J. House, B.Sc., A.R.C.S., F.R.I.C. A lecture on “Pharmacological Studies of Habituation” was given by E. M. Glaser, M.C., Ph.D., M.D., M.R.C.P. The meeting was preceded at 2 p.m. by a tour of the Evans Medical Research Laboratories. SCOTTISH SECTION A JOINT Meeting of the Scottish Section with the Stirlingshire and District Section of the Royal Institute of Chemistry was held at 7.30 p.m.on Friday, March 29th, 1963, at the Lea Park Tea-room, Callendar Road, Falkirk. The Chair was taken by the Chairman of the Scottish Section, Dr. R. A. Chalmers, B.Sc. The following paper was presented and discussed: “The Forensic Chemist,” by F. G. Tryhorn, D .Sc., F.R.1 .C.412 PROCEEDINGS [Analyst, Vol. 88 WESTERN SECTION A JOINT Meeting of the Western Section and the South Wales Section of the Royal Institute of Chemistry was held at 6 p.m. on Friday, March 15th, 1963, in the New Science Building, Swansea College of Technology, Swansea. The Chair was taken by the Chairman of the South Wales Section, Mr. H. Evans, B.Sc., F.R.I.C. A lecture on “Atmospheric Pollution” was given by 13.T. Commins, M.Sc., Ph.D., A.R.I.C. WESTERN SECTION AND PHYSICAL METHODS GROUP A JOINT Meeting of the Western Section and the Physical Methods Group was held at 6.30 p.m. on Tuesday, March 26th, 1963, in the Department of Inorganic and Physical Chemistry, The University, Hristol. The Chair was taken by the Chairman of the Western Section, Dr. F. H. Pollard. The subject of the meeting was “Determination of Pesticide Residues” and the following papers were presented and discussed: Introductory Talk by D. T. Lewis, C.B., Ph.D., D.Sc., M.R.S.H., F.R.I.C. ; “Ordinate Scale Expansion Techniques in the Infrared Analysis of Pesticide Residues,” by Miss G. P. Mansfield, B.Sc., and D. F. Muggleton, B.Sc., Ph.D., A.R.I.C. ; “Mercury Residues in Plants,” by J .A. Pickard and J. T. Martin, Ph.D., D.Sc., F.R.I.C. ; “The Detection and Determination of Some Chlorinated Pesticide Residues in Crops, Soils and Animal Tissues by Gas - Liquid Chromatography,” by R. Goulden, F.R.I.C. (see summaries below). The meeting was preceded at 2.30 p.m. by a visit to Long Ashton Research Station. ORDINATE SCALE EXPAXSION TEXHNIQUES IK THE INFRARED ANALYSIS OF PESTICIDE RESIDUES DR. D. F. MUGGLETON said that the study of residues on growing plant material and harvested crops was now an integral part of the development of new pesticides. In consequence, satisfactory routine analytical methods were required for many com- pounds that might be present in fruit and similar material to the extent of only a few parts per million.Until quite recently infrared measurements were relatively insensitive and usually required several milligrams of sample. With pesticides, therefore, they were of value only for production control. Now, however, as the result of improvements in instru- ments and technique, infrared spectroscopy could be used satisfactorily for residue studies involving microgram amounts of a compound. Indeed, it had several ad- vantages- (1) Most compounds exhibited several absorption bands in the readily accessible infrared region from 2 to 16 1-1. These were available for both qualitative and quantitative studies so that complicated analytical procedures were avoided. (2) Because every compound had a characteristic infrared spectrum, metabolites could readily be detected.( 3 ) “Clean-up” requirements were no more rigorous than for other methods, and indeed might be less so. (4) The same basic equipment and technique were available for many compounds so that instrumentation and technician training were simplified. Unfortunately, samples containing a few micrograms of pesticide might give extremely weak absorption bands, whose accurate measurement by conventional methods was difficult. Spectral bands sufficiently intense for precise measurement were ob- tained if the amount of energy absorbed by the sample was increased by using long path-length cells or beam-condensing optics. A more flexible approach was offered by ordinate scale .expansion, which gave easily measured spectral bands, but required no large absorption of energy by the sample.I t merely involved adjusting the spectro- photometer to the most favourable operating conditions and then increasing electronically the recorder pen deflection for a given change in sample transmittance. Some modi- fication of the standard double-beam spectrophotometer was required, but the intro- duction of two precision potentiometers, one linked to the optical wedge and the otherJune, 19631 PROCEEDINGS 41 3 to the pen carriage, permitted appropriate expansion factors to be introduced at will. Factors of 5 x , 10 x and 20 x were convenient for many purposes. With the potassium bromide pressed-disc technique about 1 mg of phosphamidon, for example, was required for a full-scale spectrum, but absorption bands of comparable intensity were obtained with 50 pg and 20 x scale expansion.Thus 5 pg of this compound could be detected, and this represented less than 0.1 p.p.m. on 100 g of plant material, since, in general, recoveries were not quantitative. First, it permitted very small amounts of pesticide residue (about 0.2 p.p.m.) to be qualitatively identified and quantitatively measured on reasonably sized samples (about 100 g) of plant material. It might therefore increase the sensitivity of an established method and allow a reduction to be made in sample size. Secondly, it could be employed with samples in the liquid or solid state. The potassium bromide pressed-disc technique was very convenient for samples containing completely unknown amounts of residue and a wide range of expensive cells is not required.Finally, useful measurements could be made on relatively weak absorption bands that occurred at convenient points in the spectrum. Infrared spectroscopy was thus of considerable value for residue studies and the sensitivity attainable with ordinate scale expansion compared favourably with that of other techniques. Ordinate scale expansion had useful applications in residue analysis. MERCURY RESIDUES IN PLANTS DR. J . T. MARTIN said that inorganic mercury compounds were used for soil and root treatment and organic mercury compounds were used in seed dressings, orchard spravs and glasshouse aerosols for the control of fungal diseases. The uptake of mercury by roots, leaves and fruits after applications at commercial rates had been followed in relation to the efficiency of disease control and to possible risks to the consumers of treated crops.Analytical methods had been developed for the determination of mercury residues in plant material ( J . Sci. Food A g ~ i c . , 1960, 11, 374) and soil (to be published). The roots of lettuce and dwarf bean plants accumulated mercury from nutrient solution containing phenylmercuric acetate, but little translocation occurred to the foliage. Root treatment with calomel or mercuric chloride of cauliflower seedlings before planting led to absorption by the roots but the curds were uncontaminated. Carrots grown in soil treated with mercuric chloride contained up to 0.05 p.p.m. and roots from calomel-treated soil showed 0-02 p.p.m. of mercury when seed was sown immedi- ately after soil treatment ; delay in seeding eliminated contamination.Lettuces, dwarf beans, carrots, potatoes and turnips from field experiments involving calomel and mercuric oxide soil treatments showed mercury residues not exceeding 0.03 p.p.m. ,4pple leaves absorbed mercury, deposited as phenylmercuric acetate, within a few days. Mercury was found in young coffee and citrus lime leaves that emerged after spraying with phenylmercuric acetate, indicating translocation. Broad bean plants sprayed with phenylmercuric acetate at early flowering later showed 0.02 p.p.m. in the pods, 0.04 p.p.m. in the seeds and 0-07 p.p.m. of mercury in the roots. Applicationsof phenylmercuric acetate to the leaves of potato plants led to contamination of the tubers (0.15 p.p.m. of mercury in peel, 0.18 p.p.m.in flesh) and roots (1-2 p.p.m. of mercury). Each application of phenylmercuric acetate deposited up to 10 p.p.m. of mercury on apple leaves. The surface deposits varied in different parts of a tree, and fell to 2 to 3 p.p.m. between successive treatments; mercury within the leaves rose steadily to 3 p.p.m. Apple fruits from commercially sprayed orchards contained 0.05 p.p.m. of mercury distributed between the peel, flesh and core. Five applications of phenyl- mercuric acetate, under experimental conditions, gave 0.24 p.p.m. on whole fruits ; one-third of the mercury was located in the flesh. Mercury deposited on the surface of the fruits was largely held in the cuticle; much of the mercury in the flesh arose from translocation from the leaves.Mercury was detected in the fruitlets (0.4 p.p.m.) and young leaves (0.07 p.p.m.) of trees sprayed the previous year. The bark of trees treated for six consecutive years contained 4 p.p.m. of mercury. Naturally-occurring mercury in soils varied between 0-05 and 0.012 p.p.m. Soil from beneath sprayed apple trees contained 0.2 p.p.m. (untreated 0.06 p.p.m.), Soils414 PROCEEDINGS [Analyst, Vol. 88 treated with inorganic mercurials showed up to 2 p.p.m. of mercury (untreated0.05 p.p.m.). The results indicated that while absorption of mercury by roots was followed by little translocation to aerial parts, the penetration of mercury into leaves led to movement throughout the plant. The edible portions of crops treated under commercial conditions with mercurial fungicides contained less than 0.1 p.p.m.of mercury. THE DETECTION AND DETERMIX ATION OF SOME CHLORINATED PESTICIDE RESIDUES IN CROPS, SOILS AKD ANIMAL TISSUES BY GAS-LIQUID CHROMATOGRAPHY MR. R. GOULDEN outlined a rapid method for the identification and determination of traces of chlorinated pesticides in crops, soils and animal tissues by gas - liquid chromatography with an essentially non-polar column with electron capture ionisation detection. Only conventional G.L.C. equipment was required, and neither preliminary “clean-up” nor concentration of the extract solution of the pesticide was, in general, necessary. The procedure was usually applicable to chlorinated pesticide residues in concentrations down to 0.1 to 0-25 p.p.m., and a further increase in method sensitivity for a particular pesticide could generally be achieved by adjustment of sample volumes or operating conditions or by the introduction of an extract “clean-up” stage or by some combination of these changes.Interference present in extracts from grain samples could be resolved by the use of a polar G.L.C. column or removed by liquid - solid chromatography. The standard procedure devised required about 50 minutes for a single analysis and only 30 minutes for serial analyses. MIDLANDS SECTION AN Ordinary Meeting of the Section was lield at 7 p.m. on Thursday, March 14th, 1963, at the Wolverhampton and Staff ordshire College of Technology, Wulfruna Street, Wolver- hampton. The Chair was taken by the Vice-chairman of the Section, Mr. W. H. Stephenson, The following paper was presented and discussed: “Refractory Analysis,” by H.Bennett, A JOIXT Meeting of the Midlands Section with the Lea Valley Section of the Royal Institute of Chemistry was held at 6.30 p.m. on Wednesday, April 24th, 1963, at the College of Tech- nology, Park Square, Luton. The Chair was taken by the Chairman of the Lea Valley Section, Dr. J. Haslam, F.R.I.C. The following paper was presented and discussed : “The Examination of Questioned Documents,” by Professor C. L. Wilson, Ph.D., D.Sc., F.R.I.C., F.I.C.I. F.P.S., D.B.A4., F.R.I.C. B.Sc., A4.R.I.C. MICROCHEMISTRY GROUP THE thirty-ninth London Discussion Meeting of the Group was held a t 6.30 p.m. on Wednes- day, March 20th, 1963, at “The Feathers,” Tudor Street, London, E.C.4. The Chair was taken by the Chairman of the Group, Mr. I>. W. Wilson, M.Sc., F.R.I.C. A discussion on “Microchemical Problems in Air Pollution” was opened by T. Nash, M.A., B.Sc., A.R.I.C. BIOLOGICAL METHODS GROUP THE Group held a Symposium on “The Pharmacological Screening of New Drugs” at 10 a.m. on Wednesday, March 13th, 1963, in the Friends Meeting House, Euston Road, London, w.c.1. The Chairman of the morning session was Dr. F. Hartley, F.P.S., F.R.I.C., and the following papers were presented and discussed : Introduction by the Chairman ; “Organisation of Pharmacological Screening Tests,” by M. W. Parkes, BSc., Ph.D. ; “Cardiovascular Drugs,” by A. F. Green, B.A. ; “Anticonvulsant Drugs,” by C. Cashin, B.Pharm., M.P.S. The Chairman of the afternoon session was Dr. L. F. Wiggins, F.R.I.C., and the following papers were presented and discussed: “Analgesic Drugs,” by A. Macfarlane, B.Sc. ; “Anti- inflammatory Drugs,” by s. s. Adams, R.Pharm., Ph.D., F.P.S.; “Anti-allergic Drugs,” by W. G. Smith, B.Pharm., Ph.D., A.R.I.C., F.P.S. ; Closing Remarks by the Chairman.

ISSN:0003-2654    

DOI:10.1039/AN9638800411

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

June, 19631 ANALYTICAL METHODS COMMITTEE 41 5 Analytical Methods Committee REPORT PREPARED BY THE CHLORINE I?: ORGANIC COMPOUNDS SUB-COMMITTEE The Semi-micro-determination of Chlorine in Agricultural Technical Organic Chemicals and their Formulations THE Analytical Methods Committee has received the following report from its Chlorine in Organic Compounds Sub-committee. The Report has been approved by the Analytical Methods Committee and its publication has been authorised by the Council. REPORT The work of the Sub-committee resulted from a request from the Pesticides Analysis Advisory Committee of the Ministry of Agriculture, Fisheries and Food (PAC), who had been allocated by the Collaborative Pesticides Analytical Committee of Europe (CPAC) the task of preparing methods for determining compounds containing chlorine in organic combination in pesticides. Because PAC considered this problem to be one of general analytical chemistry rather than one specific to pesticides, the Analytical Methods Committee was asked to undertake this investigation.A sub-committee was therefore set up under the chairmanship of Professor R. Belcher, and its membership was: Mr. J. H. Dunn, Mr. K. Gardner, Mr. R. Goulden, Mr. G. Ingram (who resigned during the course of the work) and Dr. A. M. G. Macdonald, with Dr. C. H. Tinker as Secretary; Mr. C. A. Johnson joined the Sub-committee later. The Sub-Com- mittee’s terms of reference were “to prepare methods for the determination of organically bound chlorine, having special reference to commercial preparations such as pesticides.” INTRODUCTION It was considered by the Sub-committee that any method recommended should, if possible, be applicable to both volatile and non-volatile compounds.The work resolved itself into two parts: (i) the selection of a method for destroying the organic matter and bringing the chlorine into solution in a suitable form and (ii) the development of a method for determining the chlorine in the solution. Consideration was given to several methods for decomposing the sample, including the flask-combustion,l Stepanow,2 peroxide-bomb3 and combustion-train procedure^.^ A semi-micro method was considered desirable, because the results obtained with a micro method are likely to be adversely affected by heterogeneity of the sample, whereas a macro method was to be avoided, if possible, as the preliminary destruction of organic matter is likely to be less efficient with the large sample required.The ultimate choice was between the flask-combustion and Stepanow methods. The former was preferred rather than the latter, which, although easier to carry out in all laboratories because it does not require any special apparatus, is not suitable for all types of compounds, particularly pentachlorophenol. The flask-combustion method is easy to manipulate and more universally applicable, although during the work of the Sub-committee it was shown to have limitations, borne out in the published chemical literature, particularly when the chlorine was present in the organic matter in low concentrations and when inorganic fillers largely predominated in formulations.The decision to use the flask-combustion method was justified by the collaborative trials, and the method finally recommended in the Appendix (see p. 417) is based on this technique. It should be noted that this method determines total organic and ionisable chlorine, and if ionisable chlorine is present in the material a correction must be made.416 ANALYTICAL METHODS COMMITTEE : SEMI-MICRO- [Alzdyst, VOl. 88 The determination of the hydrochloric acid in the solution resulting from the decom- position required investigation. Pot entiomet ric tit rat ion with standard silver nitrate solution was considered preferable, but not all laboratories have the necessary apparatus. The method involving the addition of mercuric oxycyanide to the neutralised solution and titration of the liberated alkali with standard acid is sensitive, but required study to establish the best method of determining the end-point.Direct alkalimetry proved applicable only in special instances. After completion of the collaborative work, several members of the Sub-Committee examined Cheng’s method,K as modified by White,s which involves the use of a solution of mercuric nitrate for a direct titration in aqueous ethanol, and found it satisfactory in their hands. Several met hods are feasible. EXPEHTMESTAL A considerable amount of experimental work was undertaken, as a result of which improvements were made to the technique originally proposed. The need for special pre- cautions a t certain stages was observed and clifficulties that arose in the work by individual members of the Sub-committee were noted. The method finally recommended has been drafted in the light of this experience. Points needing special emphasis are summarised below.(2) Although a 500-ml flask is adequate for the decomposition of many materials likely to be encountered, a 1-litre flask is specified, as this will be suitable for all the materials for which the method is recommended, including miscible-oil products. (ii) A cylindrical platinum-gauze sheath, rather than a spiral, is recommended as this provides a better catalytic action to assist combustion and also holds the sample more securely. The suggested dimensions of the combustion apparatus are such that the possibility of the flame impinging on a cold glass surface, causing soot formation accompanied by incomplete decomposition, is minimised. (iii) An absorbent solution consisting of water and hydrogen peroxide is recom- mended in order to preclude the formation of any hypochlorite.Moreover, the peroxide facilitates removal of the carbon dioxide resulting from the combustion of certain materials ; this carbon dioxide, if present, interferes with the oxycyanide method for determining the chloride in the absorbent solution. (iv) The presence of small amounts of soot after combustion does not appear to affect results obtained with the potentiometric finish. Soot does, however, obscure the end-point in the oxycyanide finish, and its formation should therefore be avoided. (v) With liquid materials, the sample should be weighed into a suitable,hard capsule containing ashless filter-paper floc as an absorbent.Some types of hard gelatin capsules contain a significant amount of chloride, and before a particular batch of capsules is accepted as satisfactory it should be examined to establish that its chloride content is low. Methylcellulose capsules contribute little chloride to the blank and do not give rise to acidic combustion products that would invalidate an alkalimetric finish ; they are therefore preferable. (vi) The recommended method is suitable for pure organic chlorine-containing substances and for many formulations, provided that the chlorine content exceeds about 5 per cent. However, low results have been reported for one highly chlorinated material.For routine purposes, satisfactory results may be obtained with materials containing as little as about 1 per cent. of chlorine. Attempts to apply the method to formulations containing about 0.5 per cent. of chlorine in a largely inorganic base met with varied success, and, although a technique might be developed that would be suitable for routine application in a particular laboratory, the flask-combustion method cannot be recommended as a standard procedure for this type of formulation. (vii) Although the possibility of explosion in carrying out a combustion by the flask-combustion method on a known type of compound is remote, it has been thought wisest to recommend the use of a safety screen and protective gloves or other suitable safety devices.June, 19631 DETERMINATIOK OF CHLORIKE IN AGRICULTURAL CHEMICALS 417 RESULTS The results of a final collaborative test, in which the method recommended in the Appen- dix was used, and which involved some members of the Sub-committee and other workers in this field who were not familiar with the technique, are recorded in Table I.TABLE I Laboratory x B c I) E F RESULTS OF FINAL COLLABORATIVE TEST Chlorobenzoic acid (Reference Standard) P.C.1’. miscible oil (chlorine = 22-650/,*) (chlorine = 9.84y0t) -- Weight of Chlorine Type of finish sample, found, Potentiometric 24.0 22.3’ 26.8 21.5: Potentiometric 26.5 22.6 30.5 22.7 Potentiometric 27.9 22.6 9 -7 .5 32.4 -- 30.7 22.5 40.9 22.s Osyc yanide 28.6 23.1 36.6 22.9 mg % 0 s jvc yanide 30.0 -- 30.5 Ox!-c>-anide 35.0 52.4 T-----7 IVeight of Chlorine sample, found, mg /O 36.9 9.13 40.7 9.41 45-3 9.39 49.8 9.40 45.6 9.49 5 0 4 9.49 60.0 9.45 61.6 9-32 99.4 9.4s 94.6 94;s O / 25.9 11.2 40.2 10.8 Dieldrin 50% water- dispersible powder (chlorine = 28.0%t) \ireight of Chlorink sample, found, mg /O 17-7 28.9 25.5 28.4 29-6 28.9 32.7 28.8 25.7 28.3 26.2 28.1 31.2 27.8 31.5 28.4 51.8 28.3 45-3 28.3 22.0 30.4 24.6 29.6 01 * Theoretical figure, assuming 100 per cent.purity. 1. Expected figure, based on formula of preparation. $ Own M.A.K. grade material used. 9 A little soot formed on combustion of both oil samples. An analysis of variance of the results of this collaborative test is shown in Table 11. TABLE I1 ASALYSIS OF VAKIAXCE OF RESULTS FROM FINAL COLLABORATIVE TEST Including results of Iaaboratory I; 7--- -7 Degrees of Source of variation Mean square freedom HetLveen samples .. . . .. . . 1128.76 5 Samples Y laboratories interaction . . 0.23 10 Between laboratories . . . . . . 8.26 5 Residual . . . . . . . . 0.07 18 Ratio (2)/(3) . . . . . . . . 35.9 Highly significant Excluding results of Laboratory I? Degrees of Mean square freedom 944-67 2 0.1 2 4 0-15 8 0.05 15 1 Not significant The analysis indicates a significant difference in the results obtained between the laboratories, but this difference is removed if the results from Laboratory F are excluded. The mean squares for (samples x laboratories) interaction of 0-23 and 0.15 show that the standard deviation of a single observation, irrespective of the laboratory in which the determination was made or of the chlorine level, is of the order of 0.5 to 0.4. Appendix RECOMMEXDED METHOD FOR THE SEMI-MICRO-DETERMINATION OF CHLORINE I?; ORGANIC SUBSTANCES PRINCIPLE OF METHOD- The method is based on the decomposition of the material by combustion in oxygen, with subsequent absorption of the combustion products in neutral peroxide solution and41 8 ANALYTICAL METHODS COMMITTEE : SEMI-MICRO- [Analyst, Vol.88 then determination of the hydrochloric acid formed by (2) potentiometric titration with silver nitrate solution, (ii) the use of mercuric oxycyanide with a comparative end-point procedure or (iii) in certain instances, direct alkalimetry. APPLICABILITY- The method is applicable to the determination of chlorine in organic insecticides, fungi- cides, herbicides, etc., in the form of technical chemicals and formulations.The method is not suitable, however, for formulated products containing low levels of chlorine (about 0.5 per cent.) in bases consisting largely of inorganic filler. RANGE- For chlorine contents higher than about 5 per cent. (see Note 1, p. 420). SPECIAL APPARATUS- Combustion apparatus-This consists of a round flat-bottomed 1-litre flask with a B2P ground-glass neck into which fits a glass stopper of the design shown in Fig. 1 (see Note 2). To the glass extension rod, 3 mm in diameter, is sealed a 30-mm length of platinum wire, 1 mm in diameter, terminating in a 15-mm x 20-mm piece of 36-mesh (or coarser) platinum gauze formed into a cylindrical sheath. The lower end of the sheath should be located centrally about 4 to 5cm from the base of the flask.For potentiometric titration-If the potentiometric finish is used, a cell system is required consisting of a glass reference electrode and a silver indicator electrode coupled to an electronic voltmeter or high-impedance pH/mV meter, together with a means for mechanically stirring the titration solution. REAGENTS- Absorbent solzttion-Mix 20 ml of distilled water and 1 ml of 100-volume hydrogen peroxide solution (27 per cent. w/w of H202) (microanalytical-reagent grade) directly in the combustion flask, and exactly neutralise with 0.02 N sodium hydroxide, with screened methyl red solution as indicator. Screened methyl red indicator solution-Dissolve 0.125 g of methyl red (free acid ; indicator grade) in 50 ml of 90 per cent.ethanol. Dissolve 0.083 g of methylene blue (redox indicator grade) in 50 ml of 90 per cent. ethanol. Store the solutions separately, and mix equal volumes for use as required; the mixed solution keeps for 1 week. Oxygen. The additional reagents described below are required if the potentiometric finish is used : Sulphuric acid, 5 N-Add slowly and with stirring 1 volume of sulphuric acid, sp.gr. 1.84, Silver nitrate, 0.02 3-Accurately prepared or standardised. to 6 volumes of distilled water. f 40 n Fig. 1. Appara glass rod 15mmx20mm piece of 36-mesh platinum gauze .tus for flask-combustion method (all dimensions in mm)June, 19631 DETERMINATION OF CHLORINE IN AGRICULTURAL CHEMICALS 419 The additional reagents described below are required if the oxycyanide finish is used: Sodium hydroxide, 0.02 N.Mercuric oxycyanide solution, saturated-Shake until dissolved 20 g of mercuric oxy- Filter the Sulphuric acid, 0.02 N. Standard chZoride solution, 0.02 N-Prepare from analytical-reagent grade sodium chloride cyanide with 1 litre of recently boiled and cooled distilled water. solution, and store in an amber-coloured glass bottle. Do not heat. previously dried at 250" to 350" C for 1 hour. PROCED u RE YREPARATIOS OF SAMPLE FOR COMBUSTION- to be taken for each analysis is indicated below: Prepare a representative sample of the product for analysis. The weight of sample Approximate chlorine Amount of Type of sample content, sample needed, mg o / /O Technical chemicals . . . . 50 15 to 25 Emulsifiable concentrates .. .. 10 35 to 50 Concentrated powders . . . . 25 25 to 35 Diluted powders . . . . .. 5 50 to 100 Solid samples-With use of a long-handled weighing tube, weigh the sample on to a piece of Whatman KO. 44 filter-paper; use a 3-cm square for samples weighing less than 50 mg and a 4-cm square for samples weighing 50 to 150 mg. Wrap the sample in the paper, and fix the package securely in the platinum-gauze sheath (see Note 3). Handle the paper as little as possible; preferably use rubber finger stalls. Liquid samples-With use of a capillary dropper, weigh the sample into a hard capsule of gelatin* or, preferably, methylcellulose,t supported on the balance pan in a small glass or metal thimble. Any loss of unburned liquid on combustion of the sample can be prevented by packing the lower half of the capsule with Whatman ashless filter-paper floc before the sample is introduced; about 15 mg of floc is adequate for the absorption of 50 mg of liquid sample.Close the capsule, and fix securely in the platinum-gauze sheath. COMBUSTION OF SAMPLE- Insert a filter-paper fuse, 3 mm x 30 mm, into the gauze sheath, set light to it, and quickly insert the stopper into the combustion flask, which contains the absorbent solution and which has previously been flushed out completely with oxygen from a cylinder. Carefully invert the flask so that the absorbent solution forms a seal round the stopper, which should be held firmly in place while the combustion proceeds (see Sote 4); protect the exposed hand in a stout canvas glove and interpose a safety screen between the flask and the face.When combustion is complete, shake the flask intermittently for 10 to 15 minutes to ensure complete absorption of the combustion products. Rinse the stopper, wire and gauze with a little distilled water, collecting the rinsings in the combustion flask, boil the mixed solution and rinsings gently for 1 to 2 minutes to destroy most of the peroxide and expel carbon dioxide, and cool to room temperature. DETERMINATION OF CHLORIDE- silver nitrate solution or by the mercuric oxycyanide comparative procedure. Determine the chloride in the solution either by potentiometric titration with standard Potentiometric Titration Transfer the boiled and cooled solution from the combustion of the sample quantitatively to a lipless tall-form 100-ml beaker (see Note 5), dilute the solution to about 40ml with distilled water, and add 20ml of 5 N sulphuric acid.Stir the solution mechanically, and titrate potentiometrically with 0.02 N silver nitrate added from a 10-ml microburette; use the silver - glass electrode pair. Plot the curve of e.m.f. against volume of silver nitrate * Suitable capsules (size No. 4) may be obtained from Messrs. Parke, Davis & Co. t Suitable capsules (size S o . 4) may be obtained from A. Gallenkamp & Co. Ltd.420 ANALYTICAL METHODS COMMITTEE : SEMI-MICRO- [Analyst, Vol. 88 solution added, determine the end-point, and calculate the chlorine content of the sample (see Note 6). 1 ml of 0.02 N silver nitrate = 0.709 mg of chlorine. Mercuric Oxycyanide Comparative Procedure To the boiled and cooled solution from the combustion of the sample (see Note 5 ) add screened methyl red indicator solution, and neutralise with 0.02 N sodium hydroxide, matching the colour with that of a control solution consisting of an equal volume of neutralised freshly boiled and cooled distilled water containing the same amount of screened indicator solution in a similar flask (see Note 7).Add 20 ml, or more if necessary (see Note 8), of saturated mercuric oxycyanide solution to the test solution, and titrate the liberated alkali with 0.02 K sulphuric acid until the colour again matches the neutral shade of the indicator in the control solution. To the control solution add the same volumes of saturated mercuric oxj7cyanide solution and 0.02 N sulphuric acid as were added to the test solution, and titrate the mixture with 0.02 N standard chloride solution until the colour matches that of the test solution; at the end-point, the volumes and the temperatures of both test and control solutions must be the same.Calculate the chlorine content of the sample from the volume of standard chloride solution added (see Note 9). 1 ml of 0.02 N standard chloride solution = 0.709 mg of chlorine. Alternative Direct Alkalimetric Finish If a gelatin capsule has not been used and the sample contains no other interfering elements or components, eg., nitrogen, sulphur or alkalis, the hydrochloric acid in the boiled and cooled solution from the combustion of the sample may be titrated directly with 0.02 N sodium hydroxide and screened methyl red indicator solution.BLANK TESTS- Carry out blank determinations on appropriate amounts of filter-paper floc and on the Significant blank values are to be expected on some batches of gelatin capsules. capsules. NOTES 1. For some routine purposes, when high precision is not essential, the method may be applied to materials containing less than 5 per cent. of chlorine. When greater precision is required with such materials, the Stepanow procedure (as given in “Speci- fications for Pesticides,” World Health Organisation, 1956) is suitable. 2. For the combustion of solid samples, a 500-ml conical iodine flask is satisfactory. Care must be taken to avoid soot formation. 3. Normally, the sample can be wrapped in the middle of the paper, which is folded into three and then doubled over.For the heavier samples, when much inorganic matter is present, better combustion may be achieved by spreading the sample carefully over the paper before it is folded, so that interleaving of the sample and paper takes place. For samples that burn only with difficulty, the use of filter-paper pre-treated with a strong solution of potassium nitrate and then dried may be of advantage. 4. The flame of the burning sample should not impinge on the glass, as unburned carbon may be produced and in such a determination low results may occur. Should soot formation occur, the distance between the gauze and the bottom of the flask may be increased. Soot formation may also occur if the flask has not been adequately flushed out with oxygen.6 . If unburned carbon or insoluble alkaline residue (from inorganic filler) is present, this should be filtered off, preferably by the use of a sintered-glass funnel. 6. Initially, a full plot of e.m.f. against silver nitrate titre will be necessary to establish the e.m.f. of the titration end-point. Thereafter, however, it is possible t o titrate directly to the pre-determined e.m.f. and read off the end-point titre from the burette. The end-point e.m.f. should be re-determined whenever the electrodes are cleaned.June, 19631 DETERMINATION OF CHLORINE IN AGRICULTURAL CHEMICALS 421 7. If the material contains an alkaline filler, the solution after filtration (see Note 5 above) will be alkaline and must be neutralised with acid. 8. An excess of oxycyanide solution must be used to allow the reaction to go to completion. For 0 to 12 mg of chloride, use 20 ml of saturated mercuric oxycyanide solution and for 12 to 20mg use 30ml. 9. Experience may show that on this scale of working the chlorine content of the sample, calculated from the direct sulphuric acid titration of the alkali produced by addition of oxycyanide to the test solution, is identical with that obtained by the comparative procedure given above. Confirmation of this fact should allow the com- parative titration procedure to be replaced by a simpler direct oxycyanide titration method. REFERENCES 1. Mikl, O., and Pech, J . , Chem. Listy, 1952, 46, 382; 1953, 47, 904. 2. Stepanow, A., Bey., 1906, 39, 4056. 3. Pam, S. W., J. Amer. Chem. SOG., 1903, 30, 767. 4. Belcher, R., and Ingram, G., AnaZ. Chim. Acta, 1952, 7, 319. 5. Cheng, F. W., Microchem. J . , 1959, 3, 537. 6. White, D. C., Mikrochim. Acta, 1961, 449. Schoniger, W., Mikrochim. Acta, 1955, 123; 1956, 869.

ISSN:0003-2654    

DOI:10.1039/AN9638800415

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

422 ANALYTICAL METHODS COMMITTEE : NITROGEN FACTORS FOR BEEF [Analyst, Vol. 88 Analytical Methods Committee REPORT PREPARED RY THE MEAT PRODUCTS SUB-COM,I;\/IITTEE Nitrogen Factors for Beef THE Analytical Methods Committee has received the following Report from its Meat Products Sub-committee. The Report has been approved by the Analytical Methods Committee and its publication has been authorised by the Council. REPORT The Meat Products Sub-committee of the Analytical Methods Committee responsible for the preparation of this Report was constituted as follows: Dr. S. M. Herschdoerfer (Chairman), Mr. S. Rack, Mr. P. 0. Dennis, Mr. J. R. Fraser, Dr. A. J. Kidney, Mr. T. McLach- lan, Dr. R. A. Lawrie, Dr. A. McM. Taylor and Mr. E. F. Williams (deputy Mr. H. C. Hornsey) with Dr.C. H. Tinker as Secretary. In its report on “Nitrogen Factors for Pork,”l the Sub-committee expressed its view that Stubbs and More’s method2 for determining the raw fresh meat content of manufactured products was the best one currently available. There was, however, evidence that the nitrogen factors hitherto used were not altogether valid at present. In 1952 the Analytical Methods Committee3 recommended a factor of 3.4 for beef. The Sub-committee came to the conclusion that too few reliable results had been published2*4~5 to permit a critical review of the factor. The Sub-committee was fortunate in again securing the collaboration of various meat-product manufacturers and meat-research organisations (see p. 423) all over the world in determining the nitrogen content of fresh beef meat in accordance with its requirements.Figures quoted in the literature relating to individual muscles are not included, as they were not considered All the results collected by the Sub-committee are shown in Fig. 1. Plate or flank- ---/ Clod and sticking diaphragm of brisket Fig. 2. Diagram showing factory cuts of beef used in collaborative testsNitrogen Content % Samples 3-0 3.2 3.4 3-6 3.8 4.0 4.2 4 4 4.6 m c u t 1 I . [I I . I -. I I I Description Age I . ~ . ...... . . I . BRISKET I FLANK 1 . I I THIN FLANK FOREQUARTER FLANK STICKING STICKING AND CLOD CLOD SHIN SHIN CLOD, STICKING AND SHIN RUMP SKIRT BUTTOCK TOPSIDE ROUND SILVERSIDE 51 R LO1 N FORE-RIBS BACK-RIBS RIBS BLADE CHUCK BLADE AND RIBS MISCELLANEOUS WHOLE CARCASE (Caiculared) (Actual) Home killed Home killed Argentine chilled Guernsey Jersey Friesian Shorthorn Guernsey Australian Australian Australian Hungarian Friesian/Shorthorn Argentine (fresh) Australian.Aberdeen Angus Australian. Hereford Australian. Aberdeen Angus Australian. Unknown. “Overfat” Home killed Argentine chilled Guernsey jersey Friesian Shorthorn Guernsey Friesian /Shorthorn Australian Australian Australian Argentine (fresh) Australian. Aberdeen Angus Australian. Unknown. “Overfat” Australian Australian Australian Australian. Aberdeen Angus Australian. Hereford Australian. Aberdeen Angus Australian. Unknown. “Overfat” Hungarian Argentine (fresh) Home killed Argentine chilled Friesian Australian Australian Australian Hungarian Australian. Aberdeen Angus Australian.Hereford Australian. Aberdeen Angus Australian. Unknown. “Overfat” Home killed Home killed Argentine chilled Australian Boneless Australian Australian Australian Hereford /Jersey Hereford /Jersey Red Danish Dairy Hungariar, Argenti ne chi I led Friesian Australian. Aberdeen Angus Australian. Hereford Australian. Aberdeen Angus Australian. Unknown. “Overfat” Guernsey jersey Shorthorn Guernsey Home killed Friesian /Shorthorn Argentine (fresh) Home killed Home killed Home killed Sex Source Laboratory D F Laboratory E Laboratory F 4Q y r F Laboratory H 64 F !-boratory H 10 F Laboratory F 4 F Laboratory H 6 F Laboratory H Laboratory I If Laboratory I Q-8 Laboratory I Laboratory J 5 F Laboratory H Laboratory G M Laboratory I la M Laboratory I M Laboratory I M Laboratory I Laboratory D Laboratory F 42 F Laboratory H 6; F Laboratory H 10 F Laboratory F 4 F Laboratory H 6 F Laboratory H 5 F Laboratory H ; Laboratory I I ; Laboratory I 2; Laboratory I { M Laboratory I M Laboratory I Laboratory I I! Laboratory I 2$ Laboratory I I \ M Laboratory I j M Laboratory I M Laboratory I M Laboratory I 11-8 Laboratory j Laboratory G L Laboratory G Laboratory D Laboratory E 10: F LaboratoryF Laboratory I I Laboratory I 2; Laboratory I I:-8 Laboratory J I $ M Laboratory I M Laboratory I i M Laboratory I M Laboratory I Laboratory D Laboratory D Laboratory D Laboratory F ; Laboratory I 1; Laboratory I 24 Laboratory I ? M Laboratory K I; Laboratory K I; M Laboratory K if-8 Laboratory J Laboratory F 10 F Laboratory F 1; M Laboratory I i M Laboratory I ?, M Laboratory I M Laboratory I 41 F Laboratory H 7 F Laboratory H F Laboratory H 6 F Laboratory H Laboratory E 5 F Laboratory H Laboratory G Laboratory D Laboratory E Laboratory D No.- Home killed Home killed Home killed Herefordjjersey Hereford/Jersey Red Danish Dairy Home killed Home killed Home killed Home killed Home killed German Home killed Home killed Home killed German “Spann Rippe” Australian Boneless Hungarian Home killed Australian Boneless Hereford /Jersey Hereford /jersey Red Danish Dairy Argentine chilled No details, sausage meat cuts No details, sausage meat cuts No details. sausage meat cuts No details, sausage meat cuts (Frozen) Leg of Mutton Cut, Home Killed Imported frozen Australian (No details) Home killed (No details) Frozen (No details) Lean Meat (No details) Imported Frozen Friesian --Crop (clod, blade & rib) Hereford ;Jersey-Breast (flank and brisket) Hereford /jersey-Breast (flank & brisket) Red Danish Dairy-Breast (flank & brisket) - Hindquarter Fore & Hindquarters Arithmetic mean.Whole carcase Leg. German (No details). German Guernsey Jersey Shorthorr! Guernsey Friesian /Shorthorn Australian Argentine Frozen Argentine Laboratory D F Laboratory E Laboratory D ; M Laboratory K I ; Laboratory K I; M Laboratory K Laboratory D Laboratory E Laboratory D Laboratory E Laboratory E Laboratory L Laboratory D Laboratory E Laboratory D Laboratory L Laboratory F I ;-8 Laboratory J Laboratory D Laboratory F ; M Laboratory K I .I Laboratory K I; M Laboratory K Laboratory F Jackson &Jones Markland & Sutton A.M.I.F. I A.M.I.F.Ldboratory D M Laboratory E Laboratory M Laboratory M Laboratory E Laboratory E 10 F Laboratory F M Laboratory K I 1 Laboratory K 1; M Laboratory K 1 ;-8 Laboratory J F Laboratory E Laboratory N F Laboratory 0 Laboratory L Laboratory L 4; F Laboratory H 6 ; F Laboratory H 4 F Laboratory H 6 F Laboratory H 5 F Laboratory H Laboratory G Laboratory P M Laboratory I of - 4 s 2 s I W I W I W I W I W I W I W 5 w 5 w 5 w I W 8 W I W I W I W I W 3 s I w I W I W I W I W I W I W I W 5 w 5 w 7 w I W I W I W 5 w 5 w I W I W I W I W 5w I 2 w 2 s 7 w I W I W 5 w 5 w 5 w I W I W I W I W I S 29 W 12 w 3 w I W 5 w 5 w I W 4 w 2 w 5 W I W I W I W I W I W I W I W I W I W I W 4 s I W 7 w 2 s 8 s 4 s 2 s 2 s 2 s I S 4 s 2 s 4 s 4 s 5 s 4 s 4 s 127 S I S 4 s I S 16 S 3 s 5 w I S 3 s I S 4 s 2 5 I W 12 s 6 5 45 s 6 s I S 2 s I S 2 s 2 s 2 s I W I W 4 w 2 w 5 w 2 s 28 S I W 113 S 24 S I W I W I W I W I W 5 w 3 w 3 w I I I I I I I I m I I I I I I I I I I i I I I I I I I I I I I I I I I I .I I : T i T 1 I I I 1 I I I I I . I I I I I 1 I I I I I I I . I I 1 I I. I I I I I ’ I I I i l I I I I i I i I I I - 1 I I I I I I I I I I I I I I I I I . I I I i i 1 I I I1 I. I I I I I a I m m . . I . I . m I I . I - r I .I S = Analyses on samples comminuted W = Analyses on whole cuts comminuted Fig. 1. Nitrogen on fat-free contents of various cuts oi Horizontal lines represent the range of nitrogen contents, beef. shoft vertical lines indicate the average valuesJune, 19631 ANALYTICAL METHODS COMMITTEE: NITROGEN FACTORS FOR BEEF 423 relevant to the investigation. To reduce sampling errors as much as possible the Sub- Committee asked for analyses to be carried out by one of two recommended methods.In the first, certain cuts of beef (as marked in Fig. 2, which was sent out to all collaborating laboratories) were to be comminuted, mixed and sampled for analysis. The selected cuts were those known to the trade as factory cuts, ie., flank, shin, plate, brisket and clod with sticking. In the second, all usable meat from the whole carcase was to be comminuted, mixed and sampled for analysis. As the latter method was costly and laborious, only 16 carcases were analysed in this way. In reviewing all the available results, the Sub-committee noticed an interesting distri- bution of figures according to the country of origin (see Table I).Some loss of moisture in carcases transported over long distances was indicated by the slightly higher average nitrogen content of meat from the Argentine and Australia when analysed in this country. TABLE I SUMMARY OF NITROGEN (FAT-FREE) VALUES At country of origin, After import to U.K., % - % Home killed . . .. 3.6 Argentine . . . . .. 3.4 3.5 Australian .. .. 3.7 3.8 .. 3.45 - Danish . . .. 3.5 - Dutch . . .. .. German . . .. .. 3.6 Hungarian .. .. 3-6 - - As found in the previous work on the nitrogen factor for pork, a wide range of values was also obtained for beef, namely, 2.96 to 4.53. The arithmetical mean for all the figures (716 samples) was 3.57, whereas that for chilled and frozen samples only was 3.68 (32 samples).For 16 whole carcases the average value was 3-59, for the factory cuts 3.56 (223 samples) and for the miscellaneous samples, which also represented types of meat used for manu- facturing purposes, 3-53 (260 samples). RECOMMENDATION After due consideration of the types of beef used in the manufacture of comminuted products, the Sub-committee recommends an average nitrogen factor of 3-55 as the best compromise for general use. ACKNOWLEDGMENT The Sub-committee thanks those listed below for their help and communications- Brand & Co. Ltd. Co-operative Wholesale Society. Crosse & Blackwell Ltd. C.S.1 . R. 0. (Australia). Danish Meat Research Institute, Roskilde. Federal Meat Research Institute, Kulmbach, Germany. Hungarian Meat Research Institute, Budapest. Liebig’s Extract of Meat Co. Ltd. (Argentine). Markland, J., and Sutton, R. W. National College of Food Technology, Weybridge. N. V. H. Hartog’s Fabrieken Oss, Holland. 0x0 Ltd. J. Sainsbury Ltd. T. Wall & Sons Ltd. (Meat & Handy Foods) Ltd. REFERENCES 1. 2. 3. 4. 6. Analytical Methods Committee, Analyst, 1961, 86, 557. Stubbs, G., and More, A., Ibid., 1919, 44, 125. Analytical Methods Committee, Ibid., 1952, 77, 643. Jackson, F. W., and Jones, O., Ibid., 1932, 57, 562. Maroney, J. E., and Landmann, W. A., Circular No. 49, American Meat Institute Foundation, Chicago, 1959.

ISSN:0003-2654    

DOI:10.1039/AN9638800422

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

424 BISHOP AND DHANESHWAR: SILVER- AND HALIDE-IOX [Analyst, VOI. 88 Silver- and Halide-ion Responsive Electrodes Part I. Preparation, Treatment, Durability and Zero-current Response in Aqueous Media BY E. BISHOP AND R. G. DHAXESHWAR* (Washington Singer Laboratories, The University, Exeter) In the preparation and treatment of silver metal, silver halide, gold and silver amalgam and cation-responsive glass electrodes, the current density, electrolyte composition and duration of deposition of halide films, the con- ditioning of electrodes before use, the durability of the electrodes and methods of cleaning and storage have been studied, and recommendations are made. The response of the electrodes to silver, chloride, bromide and iodide ions in aqueous solutions in the concentration range 10-I to M has been determined under dynamic conditions in the absence of supporting electrolytes, and values of formal potentials, slope factors, response ranges and failure points have been derived.All the electrodes respond first order to silver ion, with efficiencies of 65 to 75 per cent. Silver, clean or fouled, amalgam and halide-coated electrodes, provided the coating is of a halide of atomic number equal to or greater than that of the halide ion in solution, behave as halide electrodes with greater than theoretical slopes. Silver amalgam behaves as an active silver metal electrode, gold amalgam as a mercury electrode. IN studies of the differential electrolytic potentiometry (UEP)l of argentimetric reactions, phenomena have been encountered2 s3 s 4 that are not explicable by classical theories of electrode behaviour.From the behaviour patterns of DEP, electrode behaviour can be deduced with some confidence, but confirmation is lacking. Despite a tacit but vague agreement among electrochemists that silver electrodes respond to halide ions, halide elec- trodes respond to silver ion, slopes may deviate from theoretical and electrodes become non-Nernstian at high dilution, recourse to the literature reveals little supporting evidence. The preparation, reproducibility, surface characteristics and standard potential are abundantly covered for silver chloride electrodes, and adequately for silver bromide5 at 0.1 M and above. Apart from a study of the hydrosulphide and silver ion responses of silver metal677 attention has been confined8s9 to silver concentrations ;> Dual response of electrodes has had scant notice.6~10~11 The influence of bromide ion on chloride electrodes has been noted12 and general directions for the analysis of halide mixtures have been given,13 but no information is available on halide interaction. Electrolytic pre-treatment of electrodes has been men- tioned,14 but neither the effect of continuous passage of current nor behaviour in mixed solvent - water media at lower concentrations has been treated.Electron-transfer systems affect potentiometric response in argentimetry,15 but no quantitative study has been reported. The non-theoretical behaviour of many electrodes is attested by the limited usage of absolute potentiometry. Relative potentiometry, such as null-point potentiometryl6 or conventional pH measurement, also suffers limitations. In summary, most publications deal with the performance of silver and halide electrodes at high constant concentration in respect of their function as reference electrodes, and not with their indicator response characteristics. To set the subject on a quantitative basis and to confirm and extend the interpretation of argentimetric DEP, a complete investigation by the method of response curves has been undertaken.This paper deals with the preparation, treatment and durability of a wide range of electrodes and their responses in aqueous media in the absence of supporting elec- trolytes. Part I1 deals with behaviour in solvent - water media and with the effect of polarising currents ; the final part treats ionic interactions and supporting electrolytes.ELECTRODES- & Co. Ltd.) sealed into soft glass sheaths as previously described" and cut to size. Bombay 28, India. M. EXPER MENTAL Wire electrodes were made from 22 s.w.g. mint silver or fine gold (Johnson, Matthey Micro * Present address : Analytical Division, Atomic Energy Establishment (Trombay), 414A Cadell RoadJune, 19631 RESPONSIVE ELECTRODES. PART I 425 electrodes2 were made from the same wire, cut at the seal and ground square to expose the cross-section of the wire (0.003973 sq. cm). Square sheet silver electrodes of side 1, 0.316 and 0.2 cm cut to a precision of 50.1 per cent. were mounted by welding to 22 s.w.g. wire and sealing into the shrunk end of a glass sheath by means of Araldite so as to avoid heat distortion, and the backs and edges coated with Araldite so as to expose areas of 1, 0.1 and 0.04 sq.cm. Wire-form silver metal electrodes were trimmed to 1.5 cm, coated wires were trimmed to 2 cm before coating. Silver halide electrodes were prepared by making cleaned silver wire electrodes anodic against a silver cathode in the appropriate electrolyte (100 ml of 0.75 M hydrochloric acid, 0.5 M hydrobromic acid or 0.05 M potassium iodide PLUS one drop of 0.01 M sulphuric acid) at the optimum current density for the required time (20 mA per sq. cm for 45 minutes for chloride, 0.2 mA per sq. cm for 4 hours for bromide or 2 mA per sq. cm for 15 minutes for iodide), with subsequent thorough washing and immersion in 0.01 M halide until the potential reached equilibrium.Silver and gold amalgam electrodes were prepared by immersing the mounted and carefully cleaned wire in Specpure mercury for a few seconds. Cation-responsive glass (BH68) electrodes, kindly supplied by Electronic Instruments Ltd., Richmond, Surrey, as GNA23 electrodes, were transported in 0.1 M sodium chloride, rinsed with water and aged in 0.1 M silver nitrate for 8 days before use. The reference electrode was a high capacity mercurous sulphate - saturated potassium sulphate cell15 connected to the test solution via a saturated potassium sulphate salt bridge terminated by a low leakage rate ceramic plug. REAGENTS- distilled water1* of halide content < 10-l1 M was used throughout, acid was used for cleaning the electrodes.prepared by direct weighing, and used as stock. before use by successive ten-fold dilution of the stock solutions in calibrated glassware. PROCEDURE- After suitable cleaning or pre-treatment, sets of the various electrodes were immersed in the appropriate vigorously stirred solution at 20" C and their potentials with respect to the reference electrode observed serially at intervals on a Doran M4989 potentiometric pH meter until the rate of drift became imperceptible (below 0.1 mV per minute). After the electrodes had been thoroughly washed, the next more dilute solution was substituted, and the process repeated until the potential showed no further change with dilution. The process was then repeated in reverse order. They should not be allowed to dry out.All reagents were of AnalaR grade and were analysed as previously de~cribed.~J' Quartz- Transistor-grade nitric Standard 0.1 M silver and halide solutions were More dilute solutions were freshly prepared PREPARATION, TREATMENT AND DURABILITY OF THE ELECTRODES INFLUENCE OF CURRENT DENSITY IN THE PREPARATION OF HALIDE ELECTRODES- The method for chloride, based on that of BatesJ2O proved quick and reliable, and gives reproducible electrodes, stable for considerable periods. The initial current density of 20 mA per sq. cm falls to 15 mA per sq. cm. Above 18 mA per sq. cm a white deposit of poor adhesion is fonnedls; at lower current density the mechanism of deposition changes, resulting in a porous adherent deposit. Jaenicke, Tischer and Gerischer,21 studying the cathodic reduction of chloride films, found that the ohmic resistance of the film rapidly drops as silver metal forms along the inner crystal grain boundaries.The chloride film on a micro cathode spalls under conditions of excessive differentiating current density, when a loose or easily a ttackable boundary film of silver is produced, or in too strongly acid conditions when the fresh metal is rapidly attacked.2 Deposition of bromide by this method failed because the current rapidly decayed virtually to zero. Apparently,22 at 20 mA per sq. cm a non-porous high-resistance film forms and stops the flow of current. This high current density proved usable if the hydro- bromic acid concentration was increased to 10 M, but a non-uniform non-adherent deposit resulted.Rather higher currents and lighter deposits gave the most satisfactory iodide films. Thiocyanate is best Current density affects the nature and electrical properties of the deposit.19 Prolonged deposition at low current gave excellent electrodes.426 BISHOP AND DHANESHWAR SILVER- AND HALIDE-ION [Analyst, Vol. 88 deposited a t 18 mA per sq. cm for 2 hours, with continuous passage of nitrogen, from 100 ml of 0.1 M potassium thiocyanate containing 1 rnl of 72 per cent. perchloric acid and previously de-oxygenated with nitrogen. Reduction of the current density to 12 mA per sq. cm gave a bulky dark grey easily removable deposit instead of the uniform adherent dull white film. AGEING OF ELECTRODES BEFORE USE- Little has been published on electrode ageing other than on silver chloride.23~24 Ageing in aerated solutions has been attributed to- causing a slight decrease in chloride concentration within the pores of the deposit giving a potential slightly positive to that of an air-free electrode.This concentration polarisation may take up to 2 days to vanish. Although the work described here confirms that fresh electrodes are positive to aged electrodes, the role of oxygen is uncertain. A 48-hour pre- ageing treatment of chloride electrodes in 0.01 M potassium chloride has been made standard practice. Thiocyanate electrodes require ageing for 24 hours in 0.1 M thiocyanate solution. In contrast, bromide electrodes reached equilibrium within 2 hours, and iodide electrodes stabilised within a few minutes. Stabilisation of coated electrodes to silver ion is similar.Except for thiocyanate, the ageing period bears some relation to the durability of the electrodes. Silver metal electrodes give an immediate proper response to both silver and halide ion, and the potential in halide media agrees closely with that of an aged halide-coated electrode. Both silver and gold amalgams rapidly reach equilibrium in chloride and bromide media > However, a considerable period of ageing is required before their silver ion response extends to M. Silver-responsive glass requires conditioning in 0.1 M silver nitrate for at least 8 days. DURABILITY OF ELECTRODES- Silver metal is readily reactivated (see below), but other electrodes become “poisoned,” presumably by clogging of the pores of the deposit. Among factors contributing to this de-activation are light, oxygen, impurities and the nature and composition of the solvent medium.Thermal cycling and degree of use also modify the structure of the deposit. Response was checked daily before use and the period of time noted when response became sluggish or the potential or slope deviated from normal. With proper care and storage, chloride electrodes remained active for 21 days, bromide electrodes for 5 to 6 days, but iodide electrodes became inactive after 36 hours and thiocyanate after 48 hours. Amalgam electrodes maintain activity for 6 months or more. CLEANING AND STORAGE OF ELECTRODES- In use, electrode performance, particularly in dilute solution, gradually becomes im- paired by sluggish response and shortening of response range.The effect of different cleaning agents on the potential of a silver electrode in M silver solution is shown in Table I. The electrodes were fouled by immersion for 2 hours in a halide solution before, and washed thoroughly after, they had been cleaned. 2Ag + 2Cl- + 2H+ + $02 = 2AgCl+ H20 M and in silver solutions > 10d2 M. TABLE I EFFECT OF CLEANING AGENTS ON THE POTENTIAL OF A SILVER ELECTRODE IN 10-3 M SILVER SOLUTION Agent Deviation from expected potential, mV Potassium cyanide, 1.0 M . . .. . . - 33 Sodium hydroxide, 10 M . . .. . . - 123 (immediately) Nitric acid, concentrated . . . . .. -1- 30 (immediately) -63 (after 30 minutes) 0 (after 2 hours) Nitric acid (1 + 1) . . .. . . .. 0 Abrasion with No. 0 emery paper .. -5 Cyanide removes grease, but chemical or mechanical exposure of a fresh metal surface is best. Immersion in aqua regia serves for treating gold before ahalgamation; chromic- sulphuric acid is also effective, but may lead to oxidation or chemical contamination resistant to washing. Abrasion with emery paper is effective, but liable to deform the metal surface,June, 19631 RESPONSIVE ELECTRODES. PART I 427 induce strains and mechanically contaminate the electrode ; it is not recommended. Im- mersion in (1 + 1) nitric acid and then washing for 5 minutes with distilled water has been adopted as the standard treatment. Increased nitric acid concentration leads to high potentials that require 2 hours to return to normal; excess of silver ion accumulates in the exchange layer and diffuses only slowly into the solution.25 Unless stripped and re-coated, halide-coated and amalgam electrodes cannot be chemi- cally cleaned, and can be washed only with water, so it is essential that they be properly stored when not employed and not be allowed to dry out.Electrodes are best stored in the medium in which they are to be used, but water serves well. Amalgam electrodes dried in air for an hour or two suffer a change in slope and response range, and display potentials more negative throughout. After exposure to organic solvents, halide-coated electrodes especially require very prolonged washing with water before the potentials return to normal. ELECTRODE RESPONSE B+ +A- +BA For a reaction- an electrode responding to the ion B+ gives a potential- and an electrode responding to the ion A- gives a potential- E = 7 ~ 0 + S 10g[B+] = T O - SpB 1 E = 7 ~ 0 + S log-= 7 ~ 0 + SPA [A 1 where r0 is the formal potential at unit ionic concentration and S is the slope factor.A plot of potential versus ion exponent is a straight line of slope S, making an intercept no at,; pB 55c 77: 80C 775 775 ; 775 Z v1 775 775 770 > E + - L (u * 2 L I I I I I I I -I 3 -I 2 - 9 10 -6 7 -5 -4 -3 -2 -1 Z 2 ? Q) v r3 I I I I I A \r 2 4 6 P% KEY TO CURVES- Silver cleaned in nitric acid (I + I ) I. 2 Silver fouled with chloride 9. Freshly amalgamated gold 3. Chloride-coated silver 10. Aged gold amalgam 4. Bromide-coated silver 5. Iodide-coated silver 12. Micro silver metal,2 0.004 sq. cm 6. Freshly amalgamated silver 13.BH68 glass 7. Aged silver amalgam Response to silver ion 8. Aged silver amalgam allowed to dry in air I I. Aged gold amalgam allowed to dry in air Fig. 1.428 BISHOP AND DHANESHWAR : SILVER- AND HALIDE-ION [A.PzdYSf, VOl. 88 or pA = 0. As pB or pA increases, a point is reached where the slope changes, usually the graph begins to curve; this point defines the lower concentration limit of linear response. Thereafter, the response continues on further dilution with a much reduced or continually changing slope for a further one or two pB or pA units until the failure point, beyond which the potential becomes constant or reverses its direction of change. Occasionally, an inter- mediate slope is interposed between the initial and final slopes. The theoretical equivalents of T, and S are the normal potential E, and the Nernst factor 2.303 RT/nF.Deviation of experimental from theoretical values will be discussed later ; the experimental values presented here are not necessarily true t hermodyanamic equilibrium values, but rather dynamic values designed to relate to the requirements of titrimetric practice. Particularly at extremely low concentrations, potentials continue to drift away from theoretical for 300 to 400 hours, approaching ultimate equilibrium asymptotically, the last few millivolts occupying 99 per cent. of the time. Titrimetrically, a drift rate less than 1 mV per minute is usually acceptable; 0.1 mV per minute is entirely adequate. The latter criterion has been adopted here; frequent checks showed the drift to be less than 1 mV per hour.The term “dynamic” is chosen in distinction from “thermodynamic” and in contrast to “kinetic” or instantaneous measure- ments (e.g., the dotted extension of curves (4) and (13) in Fig. l), which have special virtues to be described elsewhere. INDIVIDUAL RESPONSE CURVES- Little stress is laid in the literature on the response of silver electrodes to halide ions and even less on the response of halide-coated electrodes to silver ion.26 Amalgam electrodes appear to have escaped notice.6~10~11 Electrodes tend to respond to silver or halide ion depending on which is in excess, and change in nature during an argentimetric titration. This accords with the shape of recorded titration curves,17 which indicate a change of slope during titration.Response of various electrodes to the four ions, silver, chloride, bromide and iodide, in aqueous solution free from supporting electrolyte are depicted in Figs. 1 to 4. Many unequivocal sharp discontinuities in slope occur, but curvature is not excluded by discon- tinuous presentation. Critical values of no, S , limits of linear response and failure points are 9 E I aJ Y I Id E U.j 400 I 6 Z p 450 fI ui 450 400 450 350 350 350 350 Fig. 2. liesponse to chloride ion (for key to curves, see Fig. 1)June, 19631 u! I. z s : E $ 2 E Y 5 2 200 200 200 200 225 300 RESPONSIVE ELECTRODES. PART I I I I I I h 0 Z L 2. .I0 -7 .5 .4 I '2 . 3 , e, I I I 1 - 2 4 6 PBr Fig. 3. Response to bromide ion (for key to curves, see Fig. 1) 429 I 2 3 4(Curves 3.4 and 5) 7 8 PI Fig.4. Response to iodide ion (for key to curves, sce Fig. 1)TABLE I1 RESPONSE OF ELECTRODES TO SILVER AND HALIDE IONS IN AQUEOUS SOLUTION AT 20°C (3) (4) (6) AgCl AgBr AgI (6) (7) (8) Silver amalgam 0.832 0.814 - - 40.8 - 2 6.9 - - 12 - 2 8 - (9) (10) (11) Gold amalgam 0.828 0.840 - 36.0 - 2 5-1 - 17 3 6 - - - - (13) BH68 glass Electrodes Response to Ag+- ro (extrapolated), V . . Initial slope, mV/pAg . . Final slope, mV/pAg . . Failure point, pAg . . ro (extrapolated), V . . Initial slope, mV/pC1 . . Intermediate slope, mV/pCl Final slope, mV/pC1 . . Failure point, pC1 . . v0 (extrapolated), V . . Initial slope, mV/pBr . . Intermediate slope, mV/pBr Final slope, mV/pBr . . Failure point, pBr . . 7r0 (extrapolated), V . . Initial slope, mV/pI . . Intermediate slope, mV/pI Final slope, mV/pI .. Failure point, PI . . Limit, pAg . . .. Response to Cl-- Limit, pC1 . . .. Limit, pC1 . . .. Response to Br-- Limit, pBr . . .. Limit, pBr . . .. Response to I-- Limit, PI . . . . Limit, PI . . .. 0.818 0.827 0.821 40.0 48.0 41.6 4.9 4.9 6 18 17 - 6 6 6 0.594 * 41.8 6.1 22 0.820 40.5 6 6 - 0.826 45.7 4 20 6 . . .. .. .. 0.260 68 3 0.262 68 3 - 0.2756 0.2677 0.266 64.5 54.3 56.5 3 4.85 3 45 5 42.5 20.5 26 5 7 7 - - - - 0.350 0.350 0.323 65 65 48.5 5 5 3 - - - 0.345 0.345 0.301 70 65 66 3 3.8 3 78.5 - - 5 10 23.5 46 6 6 5 - - .. .. .. .. .. .. .. 44 5 44 5 - - 23.5 5 5 5 0.136 59 2 66.7 6 6 7 0.136 69 2 66.7 6 5 7 (0.155) 0.126 0.129 - 69 63.4 - 2 6 55 71 3-59 5.6 - 14 22 18 7 8 8 - - 0.149 - 60.2 - 4 - - - - 0.149 - 64.2 - - - 6 - ..+ Z U .. .. X + U M t: .. .. .. I . 1 i; % 0.082 67 4.7 - - 60 7 - 0.088 67.3 5 - - 0.084 erratic 66.3 4 erratic 59.5 6.8 erratic 20 8 6 8 - -0.083 - 77.8 - 3 70 - 4.9 - - erratic - - - - - - - 8 - .. .. . . . . .. .. .. n b x erratic 7June, 19631 RESPONSIVE ELECTRODES. PART I 431 presented in Table 11. Potentials have been corrected to the hydrogen scale. The kinetic, or instantaneous, value of the potentials in any given medium approaches more nearly the theoretical or straight line value at high dilution (compare the dotted extensions to curves (4) and (13) in Fig. l), and the drift is away from theoretical, is unidirectional and does not recover and reverse after prolonged equilibration, The deviation is a reduction in slope, ultimately to zero, and cannot be due to loss of ions from solution by adsorption,2 because this would cause a drift in the opposite direction, i.e., towards a lower than apparent con- centration.CONCLUSIONS 1. All except the freshly amalgamated electrodes respond first order to silver ion similarly to silver metal. The different no value for BH68 glass is admissible, because this membrane electrode responds to univalent cations and is not specific to silver ion. Aged amalgamated silver electrodes give the best performance. 2. Silver ion response slopes are all considerably less than theoretical, showing efficiencies of 65 to 75 per cent., and do not reveal any intermediate changes. 3. Silver metal electrode response to silver ion is independent of electrode area down to 0.04 sq.cm; reduction of the area to 0.004 sq. cm impairs the response range but improves the slope. 4. All electrodes except glass respond similarly to halide ions, provided the halide coating has an atomic number equal to or greater than that of the halide ion in solution. Amalgam electrodes are exceptional only in respect of the no values in chloride media. Bromide-coated electrodes give the best response to chloride ion, and iodide-coated electrodes to bromide ion. 5 . Halide ion initial response slopes are generally greater than theoretical, and inter- mediate changes are frequent. 6. Fouled silver electrodes show a slightly sluggish and reduced silver ion response, but their halide ion response is practically identical to that of clean silver and silver coated with the relevant halide. Combined with other e ~ i d e n c e l , ~ , ~ this indicates that even a zero-current metal electrode acquires an adherent halide coating in halide medium. More- over, a silver electrode used to determine, eg., bromide response, when then used to determine silver ion response gives a curve exactly superimposable on curve (4) of Fig.1. 7. A halide-coated silver electrode responds to both silver ion and its own halide ion; it responds even better to a halide ion of lower atomic number, and its response to a halide of higher atomic number indicates that a slow exchange of halide occurs between the solid and solution phases resulting in an electrode of mixed and erratic behaviour. This change in surface nature has been seen taking place.8. To silver ion, fresh amalgam electrodes show a vestigial response that improves on ageing, aged silver amalgam giving the best response of all. This may be ascribed to migration of silver atoms from the base metal to give an activated surface. Freshly amal- gamated gold and silver show a response due to kinetic exchange between mercury atoms and silver ions, which lose their charge and migrate into the amalgam faster than they reach the surface from the solution when this is dilute. The aged gold amalgam differs a little from the silver amalgam in 7~~ and slope, but why it should behave differently from a freshly amalgamated gold electrode is obscure. Were the electrode aged in a silver ion solution, an exchange between mercury and silver ion would reach equilibrium and a silver ion response would be expected.It appears that growth of the gold atom population in the amalgam surface on ageing slows down migration of silver atoms away from the surface sufficiently to give silver ion response characteristics. The no values indicate that all electrodes respond first order to silver ion, independently of electrode nature. 9. It is to be expected that amalgam electrodes should respond to halide ion and that the response should be better than that of a silver electrode, because of the more favourable formation constants of the mercurous halide. Although the overpotential no - E, of about 80 mV is rather high, chloride response of freshly amalgamated electrodes is ascribed to the reaction 2Hg + 2Cl- + Hg,C12 + 2e, because the silver ion response of silver amalgam electrodes shows that silver electrode kinetics is not established until the electrode is aged.Aged silver amalgam shows a failure point of pC15 and behaves as a mixed electrode, whereas432 BISHOP AND DHAXESHWAR :Analyst, Vol. 88 aged gold amalgam with a failure point of pC1. 6 and a modified slope is a mercury electrode. Aged silver amalgam shows the characteristics of an active silver electrode in bromide and iodide media, and aged gold amalgam acts as a mercury electrode giving good behaviour in bromide media and the expected erratic response in iodide media. This interpretation is supported by the no values and the behaviour of the electrodes in titrations. We thank Electronic Instruments Ltd. for general support of this work, and one of us (R.G. U.) acknowledges the support of a Colombo Plan Fellowship and the grant of stud?; leave from the Indian Atomic Energy Establishment, Trombaj-. REFEXENCES 1. 2 . 3. 4. 5. ti. 7. 8 . 9. 10. 11. 12. 13. 14. 15. 18. 17. 18. 19. 20. 21. ‘> d _L. 23. 24. 25. 26. Bishop, E., Analyst, 1958, 83, 212 and 1960, 85, 422; Mikrochim. Acta, 1956, 619 and 1960, SO3; Bishop, E., and Dhaneshwar, It. G., Analyst, 1962, 87, 845. Bishop, E., Dhaneshwar, I<. G., and Short, G. B., Proceedings of the Feigl Anniversary Symposium on Analytical Chemistry, Birmingham, April, 1962, Elscvier Publishing Co., Amsterdam. Dhaneshwar, R. G., “Study of thc Differential Electrolytic l’otcntiometry of Argentinietric Ikactions,” University of Exeter, 1962. Janz, G. J., and Taniguchi, H., Chem. Reri., 1953, 53, 397. Tamele, M. W., Irvine, V. C., and Ryland, L. 13., Atznl. Chem., 1960, 32, 1002. Schmid, A., Winkelmann, W., and Vogcle. P., Helv. Chim. -4cfa, 1933, 16, 398. Chatterji, A. C., and Singh, K., J . Indian Chem. Soc., 1954, 31, 669. Brehmer, T. E., Tek. Foren. i Finland Forh, 1!)55, 75, 7. Fianda, F., and Nagel, K., Z . Elektrochem., 1951, 55, 606. Northrop, J . H., J . Gen. Physiol., 1948, 31, 213. Pinching, G. D., and Batcs, H. G., J . Hes. Vat. Bur. Stalzd., 1946, 37, 311. Howyers, R. G., Hsu, L., and Goldman, J . A . , .4nal. Chem., 1961, 33, 190; also liolthoff, 1. ill., and Furman, N. H., “I’otcntiometric Titrations,” John M’iley and Sons Inc., New York, 1931, p. 154. Bishop, E., and Short, G. I)., Analyst, 1962, 87, 467, 724 and 860. Purdy, W., Burns, E. X., and Rogers, L. B., .4jzaZ. Cheiiz., 1955, 27, 19x8. Bishop, E., Analyst, 1952, 77, 672. hlalmstadt, H. V., Anal. Chim. Acta, 1961, 24, 91. Bishop, E., and Dhaneshwar, R. G., Analyst, 1962, 87, 207. Bishop, E., and Sutton, J . R. B., Anal. Chzm. Acta, 1960, 22, 590. Ivcs, D. J . G., and Janz, G. J., “Reference Electrodes,” ;lcadcmic Press Inc., New York, 1961, Hates, R. G., “Elcctrometric p H Dctermination,” John Wiley and Sons Inc., Xew York, 1954, Jaenicke, W., Tischer, I<. P., and Gerischcr, I{., Z . Elektvochem., 1955, 59, 448. Nagel, K., “Procecdings of the 7th hlceetirig of C.I.T.C.E.,” Hutterworths Publications, London, MacInnes, D. A., and Parker, I<., J . Amer. Chem. Soc., 1945, 37, 1445. Smith, E. R., and Taylor, J . I<., J . Res. Xat. RUT. Stand., 1938, 20, 837. Gerischcr, H., and I’ielstich, IV., Z. Elektrochenz., 19.52, 56, 380. Samson, S., .4nal. Chint. .4cfa, 1955, 13, 473. p. 187. p. 206. 1957, p. 154. 1iect:ivcd October 31st, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800424

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

June, 19631 BISHOP AND DHANESHWAR 433 Silver- and Halide-ion Responsive Electrodes Part 11. The Behaviour in Mixed Solvent - Water Media, and the Effect of Restoring Currents on Electrode Response BY E. BISHOP AND R. G. DHANESHWAR* (Washington Sivtger Laboratories, The Upziversity, Exeter) A selection of suitable solvents is made on a basis of titrimetric per- formance and potential shift of chloride ion response. Potential shift increases with solvent concentration, rapidly above 80 per cent., and with decreasing ionic concentration. Fouling, or even poisoning, of electrodes by solvent adsorption is pronounced and difficult to remove. Dynamic zero- current response curves and formal potentials, slope factors and failure points are recorded for clean silver, silver fouled with chloride, silver chloride, silver amalgam and gold amalgam electrodes responding to silver ion in 80 per cent.methanol and acetone, and to chloride ion in 80 per cent. methanol. -4cetone gives rise to remarkable responses and severe electrode poisoning. Restoring currents extend the linear silver ion response of macro electrodes to pAg = 11.7, of micro electrodes to pAg = 10.7, and the chloride ion response of macro electrodes to pC1 = 8.7. The silver ion restoring current increases with pAg to a sharp maximum and then declines, in the same manner as does the double-layer differential capacitance. The function of the current is therefore to charge the double-layer capacitor. The chloride current demand is many times greater than for silver, and attains a stationary maximum.The effects of counter-, pre-, time-decay and induced polarisation are demon- strated. The results are used to account for phenomena encountered in potentiometric and differential electrolytic potentiometric titrimetry IN Part I1 of this series the preparation, conditioning, storage, cleaning and durability of electrodes responding to silver ion and halide ion were described. Zero-current response curves in aqueous media free from supporting electrolyte were presented for clean and dirty silver metal, silver chloride, silver bromide, silver iodide, silver amalgam, gold amalgam and cation-responsive membrane (RH6S glass) electrodes, from which values of dynamic formal potentials and slope factors were derived. In argentimetric titrimetry at titrant concentrations < lo4 S!, organic solvents or solvent - water mixtures are beneficial, primarily on account of the resultant depression of the solubility of the silver halide.Studies of such systems,2 and of the differential electrolytic potentiometry (DEP) of argentimetric reactions at high and extreme dilution on both micro3 and macro4 scales revealed phenomena in- explicable from theoretical Nernstian electrode behaviour, but understandable from the behaviour pattern of DEP.5t6 Such deductions required confirmation, and further information on electrode behaviour at low ionic concentration in mixed solvents was desirable. Search of the literature revealed some work on concentrated solutions and on standard potentials,'~s but nothing on low concentrations and little on amalgam electrode^.^ Consequently, an investigation was undertaken of the behaviour of metal, typical coated metal and amalgam electrodes in mixed solvent - water media, free from supporting electrolytes, both at zero current and under the influence of restoring currents.EXPERIMENTAL The preparation and treatment of the electrodes and the determination of response curves have been described.l y 2 Polarisation was effected according to the requirements of DEP, i.e., the currents were heavily stabilised by high value series ballast resistors such that the product of source voltage and ballast resistance fell within the range 1O1O to 1014 volt-ohms, by methods previously described.l0j1l All potentials have been converted to the hydrogen scale. Electrodes are numbered to correspond with Part 1.l * Present address : Analytical Division, Atomic Energy Establishment (Trombay), 414A Cadell Road, Bombay 38, India.434 BISHOP AND DHANESHWAR: SILVER- AKD HALIDE-ION [Analyst, vol.88 NATURE AND CONCENTRATION OF THE SOLVENT- Selection of solvent-Primarily, the solvent must be miscible with water and must exert the greatest effect on the electrode potentials without making them unstable. Exploratory work indicated that a solvent to water ratio of 80 to 20 was advantageous. Chloride ion responses were therefore taken in a mixture of 20 ml of 0.1 M potassium chloride and 80 ml of solvent. Acetone and methanol TABLE I 80 per cent. v/v SOLVENT - WATER MEDIA The more interesting results are recorded in Table I. ELECTRODE POTENTIALS VERSUS K.H.E.I N 2 X 10-2M CHLORIDE Ih’ Electrodes and potentials in millivolts Solvent IVater .. . . Methylated spirit . . Ethylene glycol . . Ethanol . . . . Dioxan . . .. ;kctone . . .. Methanol . . .. Dielectric constant E 80.4 37 24.3 2.2 20.7 33.6 - (1) Ag 375 295 338 264 208 193 155 (2) Ag dirty 375 302 340 302 25 1 228 259 (3) AgCl 385 310 339 302.5 263.5 278 322 (7) Xg/H?2 460 307 335 297 253 233 320 (10) A u / W 456 322 361 309 277 240 314 give the best shifts and dioxan causes instability, in agreement with titrimetric findings.394 All solvents cause negative shifts, an effect in terms of aqueous response equivalent to in- creasing the chloride concentration by up to three orders of magnitude. Electrode fouling-Clean and halide-fouled (i.e., simply washed with water after ex- posure to a halide solution) silver metal electrodes differ little in response in water,l but differ considerably in solvent media.Solvent-fouling is even more pronounced ; a silver metal electrode transferred from the ethanol solution to the methanol solution with an intervening thorough wash with water, showed a potential of 312 mV, and only after it had been cleaned three times in (1 + 1) nitric acid was the stable potential of 155 mV reached. Because halide-coated or amalgam electrodes cannot be chemically cleaned, adsorbed solvent molecules are not easily removed and reproducibility is impaired, in the extreme to the extent of +15 mV. The silver metal electrode is therefore a more attractive indicator electrode in such media.The proportion of solvent-The effect can be illustrated in two ways, for changing and for constant formal ionic potential. In Fig. 1, the blocks of curves (A) for silver ion response and (C) for chloride ion response were taken in solutions of x ml of 10-1 M ion diluted to 100 ml with methanol, so that 100 - x represents the percentage of methanol; curves (la) and (lc) show the response on dilution with water. With decrease of formal concentration, the silver ion potential instead of falling as in curve (la) actually rises, and the chloride ion potential instead of rising as in curve (lc) actually falls, i.e., the potentials move in the direction of the apparently increasing concentration, and this contrary movement is more marked for chloride than for silver ion.All curves show a break at 80 per cent. of solvent, and if replotted in terms of deviation from aqueous response show a marked steepening at this point.’ This indicates the condition of substantial titrimetric benefit combined with economy. The effect, moreover, increases with decreasing ionic concentration as well as with increasing solvent proportion, and it is profitable3 to increase the proportion of solvent to 95 per cent. in micro titrations at extreme dilution. The effect a t a constant formal concentration of l O P 3 ~ is shown in block (B) of Fig. 1 for the silver ion response of silver metal (lb) and silver bromide (4b) electrodes, and increases rapidly in methanol proportions above 80 per cent. For chloride ion response, the normal potential E, of a silver chloride electrode12 is plotted on a solvent concentration (w/w) scale in block (D), along with the dielectric constant, E, of methanol- water mixtures12 on the same scale.The behaviour indicates a drastic change in solvation of both active ions and electrode surface ; adsorption has already been mentioned. Free energy of solvation has been recognised as a factor in electrode behaviour in mixed solvent media,13J4 but no distinct relation with dielectric constant can be extracted.June, 19631 RESPONSIVE ELECTRODES. PART I1 435 4b 720 740: 0 20 40 60 80 I Methanol, % > Methanol, % Fig. 1. Group A: electrode potentials in a solution of x ml of 10-1 M silver nitrate diluted to 100 ml with methanol. Curve l a shows the response in aqueous solution of equivalent concentration Group B : electrode potential a t constant formal silver ion concentration of M ; curve 4b is for the silver - silver bromide electrode Group C: electrode potentials in a solution of x ml of 10-1 M potassium chloride diluted to 100ml with methanol.Curve lc shows the response in aqueous solution of equivalent con- centration Group D : E,, normal potential of the silver - silver chloride electrode plotted on a per cent. w/w scale; E , dielectric constant of methanol - water mixture plotted on a per cent. w/w scale Curves numbered as in Table I1 Effect of ratio of solvent to water on electrode response. 2 4 6 8 10 PAg Fig. 2. Response to silver ion in 80 + 20 (by volume) methanol - water. Dotted extension to curves 3, 7 and 10 indicate the effect of induced polarisation due to the field of currents indicated a t IA in Fig.5 Curves numbered as in Table 11.436 BISHOP AND DHANESHW-AR SILVER- -4SD HAILIDE-IOX Fig. 3. Hesponse t o chloride ion in 80 I 20 (by volume) methanol - water. Curves numbered as in Table 11. l h t t e d extensions to curves 3, 5 and 10 indicate the effect of intlucect polarisation by the currents indicated at I h in Fig. 6 800 800 800 800 800 I - z t 2 2 U E ? 4: w -2 7 -I .I 0 -7 -3 0 Z 2 9 h aJ I I I 2 4 6 8 Kesponse to silver ion in 80 + 20 (by volume) acetone - water PAg Fig. 4.June, 19631 RESPONSIVE ELECTRODES. PART I1 437 ZERO-CURRENT ELECTRODE RESPONSE Earlier work has been confined to standard potentials at fixed concentration of the silver chloride electrode in pure methanol,15g16 ethanoP and acetonitrile.12 y18 Janz12 reports good stability and reproducibility in the alcohols, but instability in acetonitrile, presumably because of complexation.From the foregoing results and their titrimetric performance,394319 80 per cent. v/v methanol and acetone were chosen for extended study as a first entry into the problem, and solutions for measurement were prepared by diluting 20.00 ml of a standard aqueous solution of silver or chloride ion to 100ml with solvent. Silver chloride being titrimetrically de~irable~9~9ll and having a reasonable durability1 was chosen to represent halide-coated electrodes. Amalgam electrodes have not previously been examined, but have performed well in t i t r i m e t r ~ . ~ The silver ion response curves in 80 per cent.v/v methanol are shown in Fig. 2 , and those for chloride ion in Fig. 3. Silver ion response in 80 per cent. v/v acetone is shown in Fig. 4. The critical values, following the notation of the earlier communication,l calculated from the curves are shown in Table 11. Repro- ducibility (a) of different specimens of the same type of electrode in a given solution and (b) of a given electrode transferred with intervening washing from one solution to another of the same composition, is about + 2 mV. Unless cleaning or washing of electrodes is ex- tremely thorough, the reproducibility may fall to * 15 mV. Equilibration, especially at extreme dilution, is a little slower than in water. TABLE I1 Potentials in these media are neither so reproducible nor so stable as in water.RESPONSE OF ELECTRODES TO SILVER OR CHLORIDE ION IN SOLVENT - WATER MIXTURES AT 20°C Electrodes Response to Ag+ in 80 per cent. v/v methanol- ro (extrapolated), V . . . . 0.877 Limit, pAg . . . . . . 6.7 Initial slope, mV/pAg . . . . 42.21 Intermediate slope, mV,/pAg . . - - Limit, pAg . . .. . . Final slope, mV/pAg . . . . 16 Failure point, pAg . . . . 7.8 Response to Cl- in 80 per cent. v/v methunol- ro (extrapolated), V . . . . 0-205 Initial slope, mV/pCl . . . . 685 Limit, pC1 . . . . . . 3.7 Final slope, mV/pC1 . . . . 433 Failure point, pC1 . . . . 5.7 no (extrapolated), V . . . . 0.910 Initial slope, mV/pAg . . . . 59 Limit, pAg . . . . . . 3.7 Failure point, pAg . . . . 5.7 Response to Ag+ in 80 per cent. v/v ucetone- Final slope, mV/pAg . .. . 115.5 ( 2 ) Ag dirtyt 0.870 38.8 3.7 - - 40.79 6.7 0.197 75-5 3.7 5.7 0.912 59 3.7 115-5 5-7 52s (3) AgCl 0.868 3.7 6.7 3 7.7 0.203 72.5 4.4 51 5.7 0.909 57.8 3.7 4.7 39 409 232 (7) Ag/Hg aged 0.917 - - - - 51 6.6 0.2521/ erratic 3.7 5.7 0.903 59.8 3.7 4.7 52 $ 219 (10) Au/Hg aged 0.878 41-81 6.3 - - 10 7.7 0.229 66.5 4.1 45 5.7 0.907 59.5 3.7 4.7 200 * Cleaned in (1 + 1) nitric acid and washed for 5 minutes with distilled water. t Immersed in 0-01 M potassium chloride for 1 hour and washed thoroughly with distilled water. $ Slight deviations from linearity, t-2 mV. 5 Considerable deviations from linearity, i 10 mV. Electrode behaviour in methanol differs little from that in water. 1. All electrodes respond first order to silver ion, with less than theoretical slopes. Silver 2.All electrodes respond second order to chloride ion, with greater than theoretical Generally, there is a marginal improvement in response range over water; the principal From mean slope. amalgam again gives the best slope and response range. slopes ; fouled silver metal behaves similarly to silver chloride. differences are described below.438 BISHOP AND DHAKESHWAR : SILVER- AXD HALIDE-ION [AIzaEySt, VOI. 88 3. A considerable improvement in slope, accompanied by a decrease in response range for the silver ion response of silver amalgam. A slight improvement in slope, together with some extension of range €or the silver ion response of clean silver. 4. A major decrease in slope combined with an extension of response range in the silver ion response of micro (0-004 sq.cm) silver electrodes-the reverse of the size effect in aqueous media. 5, An increase of about 30 mV over the aqueous value in the overpotential, T,, - E, = 70mV, for both chZoride and silver ion. Electrode behaviour in acetone is considerably different from that in either water or methanol. 1. The overpotentids, no - E,, are considerably greater. 2. The initial slopes are greatIy improved and are practically theoretical, hut the response range remains limited and is the same for all electrodes. 3, At the end of the initial range, the slopes, instead of decreasing, suddenly increase by a factor of 2 for silver metal or 4 for coated electrodes, and are maintained at this very high value for a further one (coated electrodes) or two (metal electrodes) pAg units, and then the response fails abruptly, all electrodes becoming stable constant-potential generators.This remarkable behaviour2 accords with titrirnetric e ~ i d e n c e ~ ? ~ that acetone is an excellent solvent for chloride determinations a t titrant concentrations down to 10-3 M, but is useIess at lower concentrations and in the determination of other halides because the electrodes are rapidly poisoned. Even higher slopes have been depicted, but not commented on, for the silver sulphide system.20 The response curves indicate a profound change in the energetics of the electrode surface layers, and an explanation may lie in the shape of the organic molecule and the adsorption of solvent molecules within the pores of the electrode surface ; this would also account for the difficulty in re-acti17ating electrodes poisoned by such materials. The observations recorded so far confirm the difficulty of generalising on the effect of organic solvents, even within a homologous series, or of enunciating guiding rules.The effect of each solvent is specific and must be individually investigated. EFFECT OF RESTORING CUKRF;NTS OS ELECTRODE RESPOSSE Zero-current electrode response fails, even in soIvent media, at active ion concentrations below 10-5 to IO-' M. The titrimetrkally beneficial increase in silver halide formation constant in solvent media is not accompanied by any significant extension of electrode response range. Consequently, halide determinations at tit rant concentrations below about 103 Sigx will run through a significant region around equivalence, where the zero-current indicator electrode is unpoised except by chance of induced polarisation (vide i~zfm) whether in aqueous or solvent media; at higher titrant concentrations this region is traversed by SO small a titrant increment as to be unimportant.At titrant concentrations of 102S\,, the zero-current potentiometric method fails entirel~r.~ Nevertheless, micro voIurnes of 10-9 M bromide and iodide and M chloride and macro volumes at slightly higher concentration have been titrated volumetricdly with silver solutions of commensurate concentration by l)EP,3,4 and even 1O-I' mole of chloride at a concentration of 2-4 x lo-* 31 has been titrated coulometrically by kinetic response DEP.396 At extreme dilution the polarising current does more than merely differentiate.2q3v4 The potential of an electrode polarised by a fixed and not too large current depends on the bulk concentration of the active ion, is highly reproduci- ble, is diffusion controlIed5 and responds rapidly to change in poIarising current .21 Stable rcproducible potentials can therefore be expected from electrodes carrying appropriate heavily stabilised currents. Because d l the work at extreme diluti0n~9~ has been done in solvent media, and in view of the poisoning effect of acetone, attention has been confined t o 80 + 20 v/v methanol - water solution and information sought on the behaviour of both macro4 and micro3 electrodes in respect to silver and chloride ion response.The electrode response at zero current was taken by the progressive dilution mcthodl until the failure point was reached.From this point on a polarising current was applied,lOP anodically for chloride ion response and cathodically for silver ion response, a second silver electrode immersed in the same solution being used as the counter electrode. The polarisingJune, 19631 RESPONSIVE ELECTRODES. PART I1 439 current was gradually incrementally increased from zero, until the stable electrode potential fell on a straight line or the extrapolation of the zero-current response curve. The equili- bration speed was high compared with zero-current potentials; even at pAg = 11, the potentials reached equilibrium within 30 minutes, and thereafter varied by not more than k0-1 mV over a period of many hours.The results for silver ion response are shown in Fig. 5, group (A) being for micro electrodes and group (B) for macro electrodes. The cathodic curves (e) indicate that linear response has been restored to pAg = 10.7 for micro and pAg = 11-7 for macro electrodes. Beyond these points the potentials begin to oscillate slightly. The zero-current response curves (c) are included for comparison, and the improvement in range is evident. The restoring current IA rises to a maximum a t pAg = 9.7 and thereafter declines; for macro electrodes the IA returns to zero. Failure of zero-current electrode response is probably due to the energy barrier presented by the double layer on the electrode surface, the ionic population at low concentration being unable to maintain an adequate exchange current.Polarographically, capacitance charging current of the DME becomes equal to the faradaic current at about M, and Hickling and Taylor22 have identified the major function of a polarising current on the silver ion response of a silver electrode with the charging of the double-layer capacitor. The restoring current may be expected to fulfil this function. Freyberger and de B r ~ y n ~ ~ have shown that the differential capacitance of the silver iodide electrode is a function of the silver ion con- centration and reaches a maximum at pAg = 7 to 9 depending on the ionic strength, p, of the solution. The present value of pAg = 9.7 in the absence of supporting electrolyte and p equal to the ion product of the solvent are in good agreement.The reason for the increased current density demand of micro over macro electrodes is at present obscure, but is probably connected with the low slope, and may relate to mechanical stress in the metal. In Fig. 6, the anodic curve (2) (a) indicates that the chloride ion response has been restored compared with the zero-current curves (2) or (3) in Fig. 3. The termination at pC1 = 8.7 in a pure halide medium is set by the behaviour of the cathode, curve (1) (e), and would not necessarily apply in a solution containing silver ion. The restoring current demand shown on a logarithmic scale as (IA) is much higher than for silver ion, which confirms the deduction3 that, at moderate current densities, the DEP anode is a constant-potential 800: 800 750 U ul 5 700 > E " T + z c n 550 ; V 500 4 650 Z v) v) 3 450 450 Fig.5. Effect of restoring currents on the response to silver ion in 80 f 20 (by volume) methanol - water. (A) micro electrodes (0.004 sq. cm) ; (B) macro electrodes. Curves a, anode potcntial; b, rest potential of anode after switching off current; c , zero-current responsc; d, rest potential of cathode after switching off current; c , cathode potcntial ; Ih, linear plot of restoring current density for curves a and e440 BISHOP AND DHANESHWAR: SILVER- AND HALIDE-ION [A.lZalySt, VOl. 88 generator leading to Z-shaped differential curves, but chloride ion response is restored at very high current densities leading to the normal peak form of differential curve. Because the restoring function demand on the polarising current varies with concentration at low concentrations, it follows that the residue of the current that serves for differentiation in DEP will vary if the total current is kept constant, and so will give rise to the observed aberrations in anodic, cathodic and differential titration c ~ r v e s . ~ ~ ~ I t also follows that, although what has previously been said about the fundamental unsoundness of potentio- metric measurements with singly polarised electrodes a t customary concentrations remains true,5 such electrodes with appropriate control of the current can be useful in the hitherto unexplored regions beyond the failure limit of zero-current electrodes.COUNTER POLARISATION- A working polarised electrode implies a counter electrode either ixl the same solution, as here and in DEP, or isolated in a separate compartment.In the cathodic silver ion response measurements in Fig. 5, the counter anode was a fouled silver electrode whose response is given in the curves (a). The counter anode assumes a potential somewhat positive to the zero-current electrode (c) becoming a constant-potential electrode at high pAg values. The reaction is stripping of silver from the surface leading to a slightly enhanced silver ion con- centration on the electrode surface. In the anodic chloride response measurements of Fig. 3, the counter cathode was a silver electrode (e) whose potential ran strongly negative in the absence of any depositable metal ion, leading to a parasitic discharge of solvent ion, i.e., hydrogen ion.PRE-POLARISATION- Anodic or cathodic pre-treatment of electrodes improves their titrimetric perf~rmance~~**~ and is effective in cleaning.26 Pre-cathodisation of a silver electrode, followed by zero-current response measurement, showed an extension of linear response range by one pAg unit, but at pAg > 7 continuous passage of current is necessary. OTHER EFFECTS OF CURRENTS AND THEIR FIELDS In DEP titrimetry silver ion would be present to prevent this. 700 60 0 > E + T 5 500 < > 5 400 30C I I I I I 2 4 6 " 8 PCI Fig 6. Effect of restoring currents on chloride ion response in Curves labelled as in Figs. 3 80 + 20 (by volume) methanol - water. and 5 ; IA is a logarithmic plotJune, 19631 RESPONSIVE ELECTRODES. PART I1 441 TIME-DECAY POLARISATION- At concentrations below the zero-current failure point, if the polarising current is stopped the potentials drift for a short time, coming to rest at a value intermediate between the polarised and zero-current values, as illustrated by the anodic time-decay curves (b) and catho- dic time-decay curves (d) in Figs.5 and 6. In the extreme, the rest potential is the same as the zero-current potential, as with micro electrodes. The drift is in a direction of an apparent increase in the indicated ion concentration, and freshly prepared or activated electrodes show this effect more markedly. Similar results have been recorded for the hydrogen ion response of platinised platinum electrode^.^' IN D u CED POLARI SATION- Zero-current electrodes immersed in a solution containing other electrodes through which current is passing display active response beyond the normal failure point, as illustrated by the broken line extensions to curves (3), (7) and (10) in Figs.2 and 3. Initially, this response is in the proper direction, but, as the ionic concentration decreases, aberrations occur and the response becomes spurious, When the polarising current is stopped, normal zero-current response is restored. Such polarisation induced by fields from current-carrying electrodes explains certain phenomena observed in titrations at low concentration.* We thank Electronic Instruments Ltd. for general support of this work, and one of us (R.G.D.) acknowledges the support of a Colombo Plan Fellowship and the grant of study leave from the Indian Atomic Energy Establishment, Trombay.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. REFERENCES Bishop, E., and Dhaneshwar, R. G., Analyst, 1963, 88, 424. -- , “Proceedings of the First Australian Electrochemistry Conference, ” Pergamon Press, Oxford, London, New York and Paris, Paper 61, in the press. -- , , APzaZyst, 1962, 87, 845. Dhaneshwar, R. G., “Study of the Differential Electrolytic Potentiometry of Argentimetric Bishop, E., Dhaneshwar, R. G., and Short, G. D., “Proceedings of the Feigl Anniversary Sym- Reactions,” University of Exeter, 1962. posium, Birmingham, April 1962, ” Elsevier Publishing Co., Amsterdam. Press, Oxford, London, New York and Paris, Paper 5. in the press. --- , , “Proceedings of the First Australian Electrochemistry Conference,” Pergamon Harned, H. S., and Mavison, J . O., J . Amer. Chem. SOC., 1936, 58, 1908. Harned, H. S., and Nestler, F. H. M., Ibid., 1946, 68, 665. Coghill, E. C., and Kirkland, J. J., Anal. Chem., 1955, 27, 1611. Bishop, E., and Short, G. D., Analyst, 1962, 87, 467. Bishop, E., and Dhaneshwar, R. G., Ibid., 1962, 87, 207. Janz, G. J., and Taniguchi, a., Chem. Rev., 1953, 53, 397. Feakins, D., and French, C . M., J . Chern. Soc., 1956, 3168; 1957, 2284 and 2581. Harned, H. S., and Owen, B. B., “Physical Chemistry of Electrolyte Solutions,” Third Edition, Reinhold Publishing Co., New York, 1958. Nonhebel, G., and Hartley, G. S., Phil. Mag., 1925, 50, 729. Austin, J. M., Hunt, A. H., Johnson, F. A., and Patron, H. N., in Conway, B. E., Editor, “Elec- trochemical Data,” Elsevier Publishing Co., Houston, Texas, 1952. Woolcock, J . W., and Hartley, H., Phil. Mag., 1928, 5, 1133. Ulick, H., and Spiegel, G., 2. phys. Chem., 1936, A177, 103 and 187. Bishop, E., Analyst, 1952, 77, 672. Tamele, M. W., Irvine, V. C., and Ryland, L. B., Anal. Chem., 1960, 32, 1002. Short, G. D., and Bishop, E., Analyst, 1962, 87, 724. Hickling, A., and Taylor, D., Disc. Faraday SOL, 1947, 1, 277. Freyberger, W. L., and de Bruyn, P. L., J . Phys. Chem., 1957, 61, 586. Purdy, W., Burns, E. A., and Rogers, L. B., Anal. Clzem., 1955, 27, 1988. Allen, P. L., and Hickling, H. A., Anal. Chim. A d a , 1954, 11, 467. Lingane, J . J ., “Electroanalytical Chemistry,” Second Edition, Interscience Publishers Inc., New Bryant, P. M., and Coates, G. E., Disc. Faraday SOL, 1947, 1, 115. York, 1958, p. 284. Received October 31st, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800433

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

442 BISHOP AND DHANESHWAR: SILVER- AND HA4LIDE-ION [A%a&St, 1'01. 88 Silver- and Halide-ion Responsive Electrodes Part 111. Ionic and Electron Transfer Interaction and the Effect of Supporting Electrolytes BY E. BISHOP AND R G. DHANESHWAR* ( R'ashington Singer Labovatoi,ies, The University, Exeter) The interactions of cations and electron transfer systems on the silver ion response of RH68 membrane electrodes, of diverse halide ions on the halide response and of electron transfer systems on the silver- and halide-ion responses of metal, halidisect and amalgam electrodes have been examined and tolerance limits established. Supporting electrolvtes improve response slopes, but shorten response ranges of the electrodes, and saturation of the solutions with silver halide has no effect on the halide-ion response charac- teristics.EAHI,IEH p a p e r ~ l . ~ . ~ reported the preparation and handling of electrodes, their zero-current response in aqueous and mixed solvent - water media free from supporting or other elec- trolj-tes, and the restoring effect of polarising currents on linear response. This has placed on a quantitative basis many of the speculative explanations of phenomena encountered in potentiometric and differential electrolj.tic potentiometric (DEP) argentimetric titrimetr\~,*~j but several questions remain unanswered, and are posed below. Experimental details have been given.' The general method of determining the tolerance limit for a potentially interfering ion of the electrode response to an indicated ion was to prepare solutions of a constant indicated ion concentration containing increasing concen- trations of the interfering ion and to measure the potential of the proper zero-current indicator electrode in the solutions.The solution for which the electrode potential deviated by t 1 mi' from the expected value,' i . e . , a &4 per cent. alteration in the indicated concentration, contained the maximum tolerable concentration of the interfering ion. INTERACTION OF CATIONS OK THE SILVER ION RESPONSE OF BH68 GLASS MEMBRANE ELECTRODES Electron-transfer systems interfere seriously with potentiometric argentimetry,C and, since glass is indifferent to redox systems, it was hoped that the new cation responsive glasses would provide a silver responsive electrode free from this defect. Although particular responses can be emphasised,' these membrane electrodes respond to most univalent cations, and the tolerance of the silver response for such ions had to be determined.The speed, range and specificity of response of BH68 membrane electrodes (kindly supplied by Electronic Instruments Ltd., Richmond, Surrey, as GNA23 electrodes) improves considerably on ageing in a solution of the selected The response of such electrodes to a single ion of activity a is reported8 to accord with- E = E o $ 2-303 RT 1 F 'g1oa and after conditioning for 8 days in 0.1 M silver nitrate the linear response to silver ion was found1 to extend to pAg = 5, with kinetic' or instantaneous response down to pAg = 7, though tlie equilibrium potentials deviated from linearity at pAg > 5.In the presence of other univalent cations, e.g., sodium, the mixed response is reported8 as- Sa+ 2.303 RT F E = E o + where K, and K, are constants for the particiilar glass. the tolerances listed in Table I. takes control at pH = 2. Bombay 28, India. The conditioned electrode showed The hydrogcn ion deviation is not serious at pH = 3, but Increase in potassium ion concentration gives first an increased * Presont address : Analytical Division, Atomic Energy Establishment (Trombav), 41 4.\ Catlell Road,June, 19631 RESPONSIVE ELECTRODES. PART 111 TABLE I TOLERANCE FOR V'SIVALENT CATIONS OF THE SILVER ION RESPONSE OF BH68 GLASS MEMBRANE ELECTRODES CONDITIONED FOK 8 DAYS IN 0.1 M SILVER NITRATE Indicated ion Tolerable concentration concentration [-4g+] Interfering ion of interfering ion, n1 0.1 I< + 0.07 0.1 Na+ 0.01 0.1 H+ 10-4 443 and then a constant potential, whereas the potential increases continuously with increase in sodium ion concentration.Halide ions are without effect, even on a conditioned electrode, which merely takes up the potential displayed in water or in silver ion at < M silver solution iron11 ion concentrations of 5 x lo-*, 5 x 10-3 and 5 x M caused potential deviations of +SO, +36 and -5 mV, respectively, thus confounding the hope expressed above ; a conclusion confirmed by titrimetric experience.5 The iron11 ion concentration must not exceed one-hundredth of the silver ion concentration. The reason for this is obscure. Xo chemical reaction between iron11 and silver ion could be discerned, nor was any change in the nature of or deposition of metallic film upon the electrode surface detectable.M. Iron111 ion is tolerated up to one-fifth of the silver ion concentration, but in a 5 x INTERACTION OF HALIDE IONS Pinching and Bates9 showed that 0.01 mole per cent. of bromide alters the potential of a standard silver chloride electrode by 0.1 to 0-2 mV. BateslO makes the further surprising statement that the potentials of the chloride electrode are not greatly affected by traces of iodide or cyanide. Tremblayll found that the bromide response of silver metal electrodes failed in the presence of chloride at an equal concentration. We have found' that silver bromide and silver iodide electrodes respond to chloride ion better than do silver chloride electrodes, silver iodide electrodes respond to bromide ion better than do silver bromide electrodes, silver chloride electrodes respond feebly or not at all to bromide or iodide ions and silver bromide electrodes behave similarly to iodide ion.Use has been made of these facts in increasing the selectivity and sensitivity of halide titrations at extreme dilution by DEP.4 Gaps remain in the quantitative knowledge of the interaction of halide ions on halide response. Interactions on chloride and bromide response are shown in Table 11. Chloride and bromide ion do not interfere with iodide ion response, even at a ratio of [Cl-]/[I-] of 10'. The chloride response of all electrodes fails in the presence of 0.1 mole per cent. of bromide, but amalgam electrodes show a better tolerance to iodide than do silver or silver halide TABLE I1 INTERFERENCE OF BROMIDE AND IODIDE ION WITH CHLOKIDE ION RESPONSE The maximum tolerable concentration of the interfering ion is given AND OF CHLORIDE AND IODIDE ION U'ITH BROMIDE RESPONSE Indicated ion M C1- lo-' M C1- M Hr- lW3 M Br- Interfering ion Br-, M I-, ni C1-, nf I-, n1 Elcctrode Silver .. .. .. - . 10-5 < 10-8 10- 3 10-7 Silver bromide . . .. .. 10-5 < 10 - 8 1 0 - 2 10-7 Silver amalgam .. .. 10-5 10 -6 10-2 10-6 Gold amalgam . . .. .. 10-5 10 -6 10-3 1 0 - 6 Silver chloride . . . . .. 10-5 < 10-8 erratic erratic electrodes, for which, contrary to the previous statement,1° less than 1 part of iodide in lo7 parts of chloride seriously interferes. An equal or greater chloride concentration is tolerated in bromide ion response, but 0.01 mole per cent.of iodide interferes; amalgam electrodes again show a favourable performance. The tolerance in the response of one ion to an interfering ion is qualitatively in the order of the solubility of the silver halides. Deter- mination of a halide in a mixture is best performed with either an amalgam electrode or a silver electrode coated with the required halide.4$5444 BISHOP AND DHANESHWAR: SILVER- AND HALIDE-ION [~4ndySt, VOl. 88 INTERFEREKCE FROM ELECTROK-TRANSFER PROCESSES Redox systems such as CrVI/III, FeIII/II and quinhydrone completely upset potentio- metric halide determinations, but the process is obscure as the effect is due to the oxidant with chromium, and the reductant with iron.g One system was chosen to examine its effect on electrode response.TronIII, pace oxidation of iodide, had little or no influence on the electrode potentials, but iron11 caused serious disturbances, increasing in intensity for the responses in the order iodide, bromide, chloride and silver ion, as shown in Table 111. TABLE I11 THE MAXIMUM TOLERABLE CONCENTRATION OF IRONII ION IS ELECTRODE RESPONSE TO 1 0 - 3 ~ SILVER AND HALIDE ION Indicated ion ( 1 0 - 3 ~ ) .. '2g+ c1- Br- I- Silver chloride . . . . . . 10-5 10-5 10-4 10-3 Silver bromide . . . . . . 10-5 10-5 10-4 10-3 Silver amalgam . . . . 10-5 10-5 10-4 10-3 Gold amalgam . . . . .. 10-5 10-4 10-4 10-3 No chemical reaction could - be detected, the difference between amalgam and other electrodes is slight, and the effect is hardly changed by alteration of the electron transfer potential by addition of ironIII.Electron response of the electrodes is possible, but im- probable, and membrane electrodes show a similar behaviour ; complexation is equally unlikely. THE EFFECT OF SUPPORTING ELECTROLYTES Electrode Silver . . .. . . .. 10-5 10-4 10-4 10-3 Most earlier work, particularly that directed towards constant or standard potentials, has been done at constant, usually fairly high, ionic strength. Titration media in the equivalence-point region have a much higher ionic strength of counter ions than of active indicated ions. Supporting electrolytes have proved effective in reducing double-layer charging currents and charge transfer over potential^^ at extreme dilution.I t has been suggested12 that the low slope factor of membrane electrodes1 is due to the lack of background electrolyte. Many responses were therefore re-determined at constant supporting electrolyte concentration. Micro- titration by DEP has shown the presence of, e.g., 7 p.p.m. of bromide in potassium chloride, 2 p.p.m. of chloride in uranyl nitrate, 1 p.p.m. of chloride in perchloric acid and 15 p.p.m. of chloride, 22 p.p.m. of bromide and 63 p.p.m. of iodide in barium nitrate, all reagents of the highest grade. Transistor-grade nitric acid (kindly supplied by British Drug Houses Ltd.) proved to have a halide content of less than 10-l1 M, and since hydrogen ion at a concentration less than 0.15 M has proved to be without effect on the electrodes, this reagent was used to give a constant concentration of 0.1 M.Kitric acid is obviously unsuitable for pH sensitive glass membrane electrodes, but these are inert to halide ion, so a trace of halide impurity insufficient to affect the silver ion concentration within the response range is tolerable, and a background of 0.1 and 0.01 M potassium nitrate7 was used for these electrodes. Silver ion response slopes increased from about 40 to 56 mV per unit of pAg for both membrane and metal electrodes. Chloride ion response slopes were reduced to 60 mV per unit of pC1 in the range pC1 = 1 to 2, and to 52 mV per unit of pC1 in the range pC1 = 2 to 3. However, at concentrations below to M, the slopes rapidly tailed off. The failure points generally fell by 1 p unit, so that the response range was considerably shortened.For silver ion response, the potential span was much the same with and without supporting electrolyte, but was covered in a shorter response range in the presence of supporting electrolyte. The use of supporting electrolyte is therefore of benefit at moderate concentrations, > 1 0 - 4 ~ , but is deleterious a t lower concentrations. The halide response curve^^,^^^ were taken in solutions free from silver ion. This may lead to a diffusion-controlled electrode imbalance, which may affect the potentials. This possibility was examined by repeating a selection of the determinations on solutions pre- saturated with silver halide. There was no detectable difference in either slope or response range under these conditions.Electrolytes of a suitably low halide content were difficult to find. Generally, the effect of supporting electrolyte on response slopes was beneficial.June, 19631 RESPONSIVE ELECTRODES. PART I11 445 OTHER INFLUENCES ON ELECTRODE POTENTIAL The rate of stirring has a profound effect on potentials in argentimetric titrimetry when a suspension of precipitate is present6 and on DEP titrimetry, which is a diffusion controlled process.13 In response measurements in the absence of suspended matter, the effect is relatively small, is greater for halide ion than for silver ion response, and increases with decreasing ion concentration. Potentials are always higher in unstirred solutions, The difference in potential between quiescent solutions and solutions stirred at cavitation speed for pAg = 1 is 1 mV, for pC1= 1 is 7 mV, for pAg = 7 is 4 to 5 mV and for pC1 = 7 is 8 to 9 mV.Above a certain critical, fairly low, speed the potentials become independent of stirring rate. Exposure to light changes the colour of the film on halide coated electrodes, but has little effect on the potential? at concentrations above to 1 0 - 4 ~ . At low concentrations the effect becomes considerable. We thank Electronic Instruments Ltd. for general support of this work, and one of us (R. G. D.) acknowledges the support of a Colombo Plan Fellowship and the grant of study leave from the Indian Atomic Energy Establishment, Trombay. 1. 2. 3. 4. 5. 6. 7 . 8. 9. 10. 11. 12. 13. REFERENCES Bishop, E., and Dhaneshwar, R. G., Analyst, 1963, 88, 424. -- , “Proceedings of the First Australian Electrochemistry Conference,” Pergamon Press, __- , Analyst, 1963, 88, 433. ~- , Ibid., 1962, 87, 845. Dhaneshwar, R. G., “Study of the Differential Electrolytic Potentiometry of Argentimetric Bishop, E., Analyst, 1952, 77, 672. Mattock, G., Ibid., 1962, 87, 930. Electronic Instruments Ltd., Technical Data Sheet ELECT-9, Issue 3, 29th September, 1961. Pinching, G. D., and Bates, R. G., J . Res. Nut. Bur. Stand., 1946, 37, 311. Bates, R. G., “Electrometric pH Determination,” John Wiley and Sons lnc., New York, 1954, Tremblay, J. L., Naturaliste Canadian, 1936, 63, 269 and 317. Mattock, G., personal communication, 1962. Short, G. D., and Bishop, E., Analyst, 1962, 87, 724 and 860. Oxford, London, New York and Paris, Paper 61, in the press. Reactions,” University of Exeter, 1962. p. 206. Received October 31st, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800442

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

446 MORRISON AND WILSON : THE ABSORPTIOMETRIC [Analyst, Vol. 88 The Absorptiornetric Determination of Silicon in Water Part HI*. Method for determining the Total Silicon Content BY I. R. MORRISON AND A. L. WILSON (Central Electricity Research Laborafories, Cleeve Road, Leatherhead, Szwrey) A method is described for determining the total silicon content of water a t low concentrations. I t consists in evaporation and subsequent fusion of samples with sodium carbonate ; the resulting sodium silicate is determined absorptiometrically as reduced or-molybdosilicic acid. The absorptiometric procedure is precise, and the major errors, a t low concentrations, arise in the fusion procedure. The standard deviation, for samples between 2 and 20 ml, was about 0.3 pg of silica over the range 0 to 60 pg of silica.Four samples can be analysed in duplicate per day. A METHOD was required for determining total silicon in the make-up water, feed-water, condensate and condensed steam (0.002 to 0.2 p.p.m. of silica), and the raw water and partially purified raw water (up to 20 or 30 p.p.m. of silica) of power stations. The results were also needed for calculating the “non-reactive” silicon content of samples by subtracting their “reactive”7 from their total silicon contents. As this difference may sometimes be small compared with the analytical results, the greatest possible precision was required. A fusion technique was considered necessary to convert into sodium silicate any form of silicon that might occur in samples; fusion with sodium carbonate was chosen as being convenient and suitable.Absorptiometric determination of the silicate was chosen because of the sensitivity of this technique. Suitable conditions for determining sodium silicate absorptiometrically were reported in Part I,2 and precise results have been obtained with a method based on reduced p-molybdosilicic acid.l However, the sensitivity of such methods is decreased by large amounts of salts, and the effect of neutralised sodium carbonate on the method was therefore determined; 0.5 g of sodium carbonate was used, as this was con- sidered the minimum desirable for use with 30-ml platinum crucibles. The results of five determinations were erratically lower (by 0.3 to 4.3 per cent.) than expected in the absence of sodium salts. The proposed method is therefore based on reduced a-molybdosilicic acid, which is not affected by large amounts of salts3 From the experimental results in Part I, a suitable procedure was chosen; it is described below, together with the tests made of it.EXPERIMENTAL APPARATUS, REAGENTS AND TECHNIQUE- The procedure given under “Method,” p. 449, was used for the work described in this section, except when indicated otherwise. Optical-density measurements were made in 4-cm cuvettes at 742 mp with a Hilger Uvispek spectrophotometer, distilled water being used in the reference cuvette. The temperature of the laboratory varied between 18” and 25” C during the work. Dilute suspensions of clay in water were prepared by mixing about 60 g of finely ground bentonite clay with one litre of water, and then repeatedly spinning in a centrifuge at about 2200 r.p.m., and decanting the suspensions.Suspension No. 1, prepared in this manner, had no visible turbidity, whereas suspension KO. 2 was slightly turbid. Other suspensions were prepared by dilution of these with water. For all the statistical significance tests applied in this work, the 95 per cent. confidence level was used. Before determinations of total silicon were carried out, all surfaces in the working area and its immediate vicinity were wiped with a wet cloth to remove dust. The laboratory was not air-conditioned. * For details of prcvious parts of this series, see reference list, p. 455. t “Reactive” silicon-mainly monomeric and dimeric silicic acid-is defined in this paper as those forms of silicon that react with ammonium molybdate in 10 minutes to form molybdosilicic acid under the conditions of the method given in Part I1 of this series.’June, 19631 DETERMINATION OF SILICON I N WATER.PART I11 447 VARIATION IN CONCENTRATION OF REAGENTS- The effect of variations in the volumes of ammonium molybdate, 48.5 per cent. sulphuric acid and reducing agent solutions dispensed was determined by using the same experimental design and analysis as in the analogous experiment in Part 1I.l The concentration of silicon in the final solution was 0.5 p.p.m. of silica. The results showed that a +20 per cent. change in the concentration of ammonium molybdate or sulphuric acid gave a k0.2 per cent. change in optical density, and that a t 2 0 per cent.change in the concentration of reducing agent had no significant effect compared with a coefficient of variation of k0.24 per cent. EFFECT OF OTHER SUBSTANCES- Except when otherwise stated, the eva.poration and fusion stages were omitted and the solutions of the substances were treated as described in the procedure for calibrating samples or reagent blank solutions. Duplicate analyses were made for each condition tested, and in each batch blank solutions and silicate solutions containing no added substance were also treated. The effects of phosphate and chloride, which both interfere, have been described previously.2 However, during a fusion platinum crucibles lost only 0.1 to 0.2 mg of platinum, equivalent to 1 to 2 p.p.m. of platinum in the final solution; the effects are therefore negligible.The effects of cupric copper and nickel are probably caused by the colour of their ions in acid solutions; both substances Table I shows the effect of other substances. Table I shows that both ionic forms of platinum interfere. TABLE I EFFECT OF OTHER SUBSTANCES All concentrations are in the final solution Platinous Platinic . . Ferrous . . Ferric . . Cuprous Cupric . . Y ickelous Chromic.. Aluminium Zinc . . Magnesium .. . . .. . . .. . . . . . . . . .. .. . . Substance . . . . . . . . .. . . . . . . . . . . .. . . .. .. , . . . .. . . . . . . . . . . .. .. . . .. .. .. .. . . . . . . .. Concentration of substance, p.p.m. .. 10 20 .. 10 50 100 .. 10 50 . . 2 10 50 . . 2.5 . . 100 . . 50 100 . . 10 . . 10 .. 10 . . 100 . . 100 CaGium .... . . MolybdenumVI,. tungstenV1, titaniumIV, vanadiumv, manganous, stannous and cobaltous . . . . Potassium . . . . .. . . .. . . Sulphate . . . . . . . . . . . . Nitrate . . . . . . . . . . . . . . Fluoride . . . . . . . . .. . . Morpholine, cyclohexylamine, octadecylamine and hydrazine sulphate* . . . . . . . . ~4lkylarylsulphonate (detergent powder containing 42 per cent. of active material and no added phosphate or silica)* . . . . .. . . 0.5 g of sodium carbonate, neutralised by 10 ml of N sulphuric acid . . . . .. . . . . Significant effects are greater than . . .. .. 0.1 of each 4500 5500 155 5 50 of each 10 - Difference, silicon found less silicon added (as p.p-m. of silica), a t silicon concentrations of- 0 p.p.m. of 0.5 p.p.m. of A I > silica silica 0.001 0.002 0.003 0.004 0.002 0-003 0.006 - 0.001 0.006 - 0.023 - 0.000 0.001 - 0.008 - o*ooo 0.001 -0.005 - - 0.057 0.002 0.002 0.049 0-052 0.005 0.004 0.01 1 0.009 0.000 0~000 - 0.001 - 0.00 1 o*ooo 0~000 - 0.001 0.000 0-004 0.003 0.000 0.009 0.009 0.004 0.004 0.01 1 0.019 0.01 1 0.002 0-000 0.006 0.006 0.003 0.004 0.009 0.010 0.01 1 0.003 * The entire procedure was carried out.448 MORRISON AND WILSON : THE ABSOHPTIOMETRIC [Analyst, Vol.88 in 2.5 N sulphuric acid gave effects of similar magnitude to those reported in Table I. Some of the interference effects shown in Table I are possibly caused by silicon in the substance added. FUSIOX PROCEDURE- Contamination-From time to time during the development work on this method erratically high results were obtained and were traced to contamination from various sources. These are discussed below.One of four platinum crucibles was shown to give consistently higher results than the others. This effect was little changed by bisulphite or carbonate fusions, but was reduced considerably by repeated evaporation of hydrofluoric acid from the crucible, and may have been caused by a piece of siliceous material embedded in the wall of the crucible. Initial difficulty was also experienced from airborne contamination during the evaporation stage. However, by using the hood described under “Method,” the incidence of unduly high results was reduced from 13 out of 82 to 3 out of 89. Further tests showed that the general level of airborne contamination during the evaporation stage was equivalent to less than 0.2 pg of silica.Some contamination occurred when the sodium carbonate was being fused ; the magnitude of the effect depended on where the fusion was carried out. This effect was measured by comparing the optical densities of fusion blank solutions with those of similar solutions prepared by fusing the carbonate with the crucible lids left on all the time. In a fume cupboard, this contamination was variable and averaged 1 pg of silica. On one occasion on a bench in the laboratory it averaged only 0.1 pg of silica (one result, of 1.1 pg, out of 12 was excluded). On other occasions, although it was not directly checked, it appeared to be no more than about 0.1 pg of silica. Although the effect was small when fusions were carried out on the bench, it could not be guaranteed to be absent, and therefore fusion blank values were always checked before and during a run of sample determinations; the results were then examined for evidence of contamination.Duplicate portions of samples were always analysed, and when the results differed by unusually large amounts, further determinations were done and the discordant values rejected. Silicon in the sodium carbonate-Various methods of reducing and homogenising the silicon content of AnalaR sodium carbonate were tested, the most effective being simply to filter a solution of the carbonate and then evaporate to a powder; details are given under “Method.” Millipore filters, with pore sizes of 450 and 10 mp, were no more effective than Whatman No. 542 filter-papers in reducing the silicon content of the carbonate.Table I1 shows the fusion blank values (see “Method,” p. 449) obtained with sodium carbonate pro- duced in this maner and also with various commercially available grades. TABLE I1 FUSION BLANK VALUES FOR VARIOUS GRADES OF SODIUM CARBONATE Number of 17 Sodium carbonate determinations AnalaR, batch 1, mixed mechanically Micro-analytical reagent : Specpure, batch 1, not mixed . . 3 batch 1, not mixed . . 9 batch 2, mixed mechanically . . 18 B.Z).H, batch 2, not mixed . . 9 9 .. . . . . 9 .. . . . . 18 { H. & W., batch 1, not mixed Specially purified (see Method) : batch 1 . . batch 2 . . . . Number of unduly high rcsults, excluded from calculations 4 0 1 0 1 3 0 0 Kemaining results 7 7 Standard Mean, deviation, pg of silica pg of silica 4.8 8.1 8.4 1.5 2.4 0.21 0.8 0.25 0.8 0.10 3.6 0.25 1.7 0.21 2.8 0.25 Efect of duration and temperature of fusion-The procedure described under “Method” (i.e., fusion at bright red heat for 6 minutes) was compared with what should be the worstJune, 19631 DETERMINATION OF SILICON IN WATER.PART I11 449 conditions for (a) incomplete fusion and (b) losses by vaporisation. Fusion for 3 minutes- the shortest time in which it was possible to rotate the crucible so that the sodium carbonate came into contact with the entire inner surface-at the lowest temperature for just melting sodium carbonate, gave results for clay suspensions 0-5 per cent. lower than those by the normal procedure; this difference was not significant. Fusion for 18 minutes with three Meker-type burners using town-mains gas to heat each crucible gave results for sodium silicate solutions 0.6 per cent.higher than those for the normal procedure; this difference was not significant. Therefore, within the ranges likely with the apparatus and procedure specified, any effects caused by the temperature and duration of fusion were less than 1 per cent. Efect of order of dissolving and neutralising the melt-Two alternative procedures were compared with the procedure described under “Method” (Le., addition of a solution of the melt to excess of sulphuric acid). Addition of sulphuric acid to a solution of the melt from a clay suspension gave results 0.6 per cent. higher than those by the normal procedure; this increase was not significant.However, dissolution of the melt in sulphuric acid and dilution with water gave a 2-6 per cent. increase with a clay suspension and a 0.6 per cent. increase with a sodium silicate solution. The limits for detecting a significant difference were 1.9 per cent. and 1-2 per cent., respectively; the result for the clay suspension is, therefore, statistically significant. The procedure described under “Method” was adopted in spite of these figures, because the effect was small and it is more difficult to avoid losses and con- tamination when sulphuric acid was added to the melt. METHOD APPARATUS- Evaporating hood-An inverted large glass funnel, placed about 2 mm above the surface of the hot-plate, forms a suitable hood. Pass a gentle stream of air (about 500 ml per minute) filtered through two Whatman No.542 filter-papers (or suitable alternative filter) down the stem of the funnel so that condensation does not take place on the inside of the funnel. A record of fusion blank values from each crucible will show whether any particular crucible gives results consistently different from the others. PoZythene bottles-Eight-ounce narrow-mouthed polythene bottles are used in the colori- metric stage. As some bottles have initially given systematically high results, treat all bottles as described below before first using them. Allow 20 ml of 2.8 per cent. sulphuric acid to stand in each bottle for 30 minutes, then add 10 ml of ammonium molybdate solution and 100ml of distilled water, and heat for 1 hour in a boiling-water bath.Discard the solutions, rinse the bottles well with distilled water, and then test them as described below. Allow 10 ml of 2.8 per cent. sulphuric acid to stand in each bottle for 30 minutes, and then add 5 ml of ammonium molybdate solution, and treat the solutions further as described in the procedure for the colorimetric stage. Use these optical densities to detect the bottles giving high results. Platinum crucibles-Thirty-millilitre platinum crucibles are suitable. Further cleaning of these particular bottles may make them satisfactory. REAGENTS- Only silicon in the reagents added before the 48.5 per cent. sulphuric acid solution contributes to the reagent blank value ; therefore polythene-ware or well-rinsed glassware is preferably used in their preparation.Reagents should be stored in polythene bottles. If techniques are chosen so that the maximum error involved in each of (a) weighing out chemicals, (b) adjusting reagent solutions to volume and (c) adding the required volume of reagent solution to samples, is less than 5 per cent. (except when other limits are specifically stated), errors from these sources should be small. Water-Distilled water from a Manesty still, stored in polythene, was found suitable and usually contained less than 0.005 p.p.m. of silica. If the water used for dissolving the fusion melt contains silicon, then in general the variations in the amounts of water used will cause errors, the magnitude of which may be estimated from the “reactive” silicon content of the water. The high-sensitivity modification of the method for “reactive” silicon described previously1 is suitable for determining the All reagents should be of analytical-reagent quality unless otherwise stated.450 MORRISON AND WILSON : THE ABSORPTIOMETRIC [AnaZyst, vol.88 silicon content. The total silicon content of the water may be higher than the“reactive” silicon content and for some purposes, e.g., careful checks for bias, should be determined by the method, and the appropriate corrections made. Sodium carbonate-Determine the homogeneity of the material by fusing about ten 0.5-g portions for a few minutes with the lids left on the crucibles, treating them further by the normal procedure and calculating the standard deviation and mean silica content of the final solution.Also obtain an estimate of the within-batch standard deviation of the reagent blank value, and use all these figures to decide whether or not the material gives adequate precision. Some batches of M.A.K. sodium carbonate (obtained from British Drug Houses Ltd.) were suitable for our purposes after being mixed mechanically for several hours. If a suitable commercial product is not available, prepare one as described below. Dissolve 200 g of analytical-reagent grade sodium carbonate, Na,CO,, in 1000 ml of water in a polythene beaker; use a polythene stirring rod. Filter this solution through a Whatman No. 542 filter-paper held in a polythene funnel, and collect the filtrate in a polythene beaker; reject the first 50 ml, and keep the beaker well covered throughout.Evaporate this solution to dryness in batches; use platinum dishes (100 ml or larger) that have been cleaned by sodium carbonate fusion, and maintain a cover about 5cm above the dishes during the evaporation. As the carbonate crystallises, break up large lumps with a spatula before they become dry and solid, as they are difficult to grind. Grind the material to a fine powder in a metal container with a metal rod; polythene was once found to cause contamin a t’ ion. Mix the powder well, by mechanical means if possible, and store in a polythene bottle. Test as described above. Ammonium moZybdate solution-A 5 per cent. w/v solution of ammonium molybdate tetrahydrate in water. A white deposit may appear on the container walls, but the reagent was usable for at least 2 months.Sul$huric acid, 48.5 per cent. v/v, aqueous. SuZphuric acid, 2.8 per cent. v/v-Dilute with caution 28 0.5 ml of concentrated sul- phuric acid to 1 litre with water. Prepare sufficient to last for an entire batch of sodium carbonate, otherwise the fusion blank value (see later) may alter. Sulphuric acid, 0.14 per cent. v/v, aqueous-Prepare sufficient to last for an entire batch of sodium carbonate. Stannous chloride solutioiz-Prepare not more than 3 hours before use. Allow a clean stoppered 50-ml cylinder or flask to drain thoroughly (about 60 seconds), place in it a 0-1-g piece of tin foil (obtainable as 0-1-g pieces from the British Drug Houses Ltd.), add 1.5 $ 0.3 ml of concentrated hydrochloric acid, and allow to dissolve. Small black particles left in the solution may be ignored if no bubbles of gas are evolved from them.Dilute to 50 ml with water. Standard soZzitions of silicorz-Fuse 1.000 g of pure dry silica with 5 g of analytical- reagent grade anhydrous sodium carbonate at red heat in a platinum crucible; dissolve the cooled residue in water, and dilute to exactly one litre. This solution contains 1000 p.p.m. of silica,. The solutions containing 1000 p.p.m. were stable within $0.5 per cent. for at least 2 years in polythene bottles, and the solutions containing 10 p.p.m. for at least 3 months. Prepare, by dilution, a solution containing 10 p.p.m. of silica. PROCED u KE- Size of sample-The selection of sample size is important in this method in order to obtain the highest precision. A balance should be struck between: (a) a large error contri- bution from the fusion blank with small volumes and (b) increased evaporation times with large volumes (at low temperatures) or increased spectrophotometric errors with high optical densities (at high concentrations).In our work, 20-ml samples were the largest used. Samples larger than 10 ml should be evaporated in 10-ml portions. Evaporation and fusion stage-Clean the platinum crucibles by fusing in them about 1 g of analytical-reagent grade sodium carbonate, rotating each crucible so that the melt touches the entire inner surface. Allow to cool, and wash out the melt. (Omit the fusions if the previous determinations in the crucibles were by this method.) Rinse the crucibles well inside and outside, and dry the outside with clean filter-paper; do not dry the inside.June, 19631 DETERMINATION OF SILICON IN WATER.PART I11 45 1 By pipette, place a suitable portion of sample (not more than 10 ml) in a clean crucible, and evaporate to dryness on a hot-plate under a hood. The hot-plate should be as hot as possible short of causing the sample to boil or spit. If the required volume of sample was more than lOml, evaporate further 10-ml portions in the same crucible. With the crucible being kept covered whenever possible, add 0.5 k 0-002 g of purified sodium carbonate, and fuse by heating to bright red heat over a burner. Rotate the crucible in the flame so that the melt touches the entire inner surface, except where the tongs are holding. Allow to cool, nearly fill the crucible with water from a polythene wash bottle, and heat gently to dissolve the melt rapidly without bumping or boiling.Add 10 A 0.2 ml of 2.8 per cent. sulphuric acid to a clean narrow-mouthed 8-02 polythene bottle; in cleaning, the walls should be well rinsed, e.g., by twice quarter filling with water, and inverting. Quantitatively transfer the contents of the crucible through a polythene funnel into the bottle; pour the solution out in one continuous stream, and avoid as far as possible having it run down the outside of the crucible, where it may be contaminated. Do not tip the crucible back before the rinsing has been completed. Rinsing should be controlled so that the final volume in the bottle is 60 t 5ml. These solutions may be left overnight if necessary, until all the bottles for a batch have been filled.Colorimetric stage-Put 5 ml of ammonium molybdate solution into the polythene bottle, and cover the mouth with a loose-fitting polythene cover. Heat the bottle in a boiling-water bath for 1 hour k5 minutes. Place the bottle in cold water, allow to cool to room tempera- ture or preferably below, add 15ml of 48.5 per cent. sulphuric acid, noting the time, and mix. One minute t-5 seconds after the sulphuric acid has been added, add 2 ml of stannous chloride reagent, and mix. Transfer the liquid quantitatively to a 100-ml calibrated flask, place the flask in cold water, and cool to 20" 2" C (or some other chosen standard tempera- ture; other limits on this temperature may be chosen to give the required precision2). Adjust the liquid to the mark with water. Within one hour after adding the reducing agent, measure the optical density of the solution in 4-cm cuvettes with distilled water in the reference cuvette; keep the solutions out of direct sunlight.The best wavelength to use is 742 mp (the peak is sharp and should be checked on each instrument), but Ilford 608 or equivalent filters can be used if necessary. Subtract the optical density of the reagent blank solution from that of the sample, and calculate the concentration of silica in the final solution from calibration results. Calibyation samples and reagent blank tests-For calibration samples, place by pipette the required volumes of standard silicon solution (7-5 ml of the solution containing 10 p.p.m. gives an optical density of about 0.95 at 742 mp) in clean 8-02 polythene bottles, and add sufficient water to make the volumes 50 & 5 ml; for reagent blank solutions merely place 50 Add 10 ml of 0.14 per cent.sulphuric acid, and carry out the procedure of the colorimetric stage. One uniform batch of water should be used for the samples, reagent blank solutions and calibration samples in each batch. Between-batch errors may be allowed for, if large, by including calibration samples (say 0.5 p.p.m. of silica) with each batch of samples, and using their optical densities in the calculations. When adding liquids from a pipette into polythene or platinum vessels, draining the pipette with its tip held against the container walls causes relatively large errors (coefficient of variation about k0.15 per cent.).It is preferable to place the tip of the pipette just under the surface of the liquid and to leave it there for the required draining period (coefficient of variation about k 0-015 per cent.). Fusion blank test-Weigh out 0.5 2 0.002 g of sodium carbonate into a clean platinum crucible, and place on a hot-plate under the hood for the same period as for samples. Carry out the rest of the procedure in the same manner as for samples; ensure that the crucibles are rotated in the flame when fusing the carbonate. In these laboratories, the variation of the fusion blank value was more than the errors of the colorimetric determination, and therefore the fusion blank figure used in the calculation of results should be the mean of many determinations (at least about ten) on the batch of sodium carbonate used.I t is advisable to carry out a further determination with every few samples as a check on the procedure and on the reagent. Replace the lid, and heat for 1 to 2 minutes. 5 ml of water in the bottles. Alternatively, the usual calibration curves may be used.452 MORRISON AND WILSON : THE ABSORPTIOMETRIC [AmZyst, 1’01. 88 Calculation of results-Subtract the mean fusion blank value (as p.p.m. of silica in the final solution) from the measured concentration of silica in the final solution, and multiply by the factor 100/sample volume, to give the concentration of silicon in the sample, expressed as p.p.m. of silica. RESULTS LINEAKITY OF CALIBRATION C ~ V E - cent. to at least 1 p.p.m. of silica in the final solution.742 mp, was 0.790 p.p.m. of silica in the final solution per unit of optical density. PRECISION- Precision of the colorimetric stage-The experimental design used to estimate precision was similar to that previously describedl The samples were (i) sodium silicate solutions, analysed in batches during 2; years and (ii) synthetic soft water (containing sodium silicate plus 15 p.p.m. of Ca2+, 7 p.p.m. of Sa+, 1-5 p.p.m. of Mg2+, 0.5 p.p.m. of K+, 0.5 p.p.m. of A13+, 0-25 p.p.m. of Fe3+, 0.25 p.p.m. of Cu2+, 0-1 p.p.m. of Mn2+, 46 p.p.m. of SO,2-, 11 p.p.m. of C1- and 0.8 p.p.m. of NO,-) analysed in batches over a period of 16 days. As 50-ml samples of synthetic soft water were used, the concentrations of the elements mentioned above in the final solution were half those stated.Table 111 shows the standard deviations of the differences-optical density of sample minus optical density of reagent blank solution expressed as p.p.m. of silica. TABLE I11 PRECISION OF COLORIMETRIC STAGE All concentrations are in the final solution A calibration curve prepared as described under “Method” was linear within k0.5 per The slope, with 4-cm cuvettes at Sodium silicate solutions Synthetic soft water b f Rcagcnt 0.5 p.p.m. Op.p:m. 0.1 p.p.m. 0.5 p.p.m. blank of silica of silica of silica of silica A > Within-batch standard deviation, p.p.m. of silica 0*0006* 0.0012t 0.00095 0.0009 0-001 1 Betwccn-batch standard deviation, p.p.m. of silica -- 0.00135 0*0005$ 0.00035; 0.0012 Degrces of freedom . . . . . . . . . . 50 62 9 10 10 Degrees of freedom .. . . . . .. . . 50 9 9 9 - * This figure is for reagent blank solution only and not for the difference: sample minus blank. t One extremely low result excluded from thc figures. $ Not statistically significant that a between-batch error exists. Table 111 shows that both the within-batch and between-batch standard deviations were not significantly affected by the additional impurities present in the synthetic soft water samples. Detailed analysis of the results for the sodium silicate solutions showed that the standard deviations remained fairly constant over 2 j years. The between-batch coefficient of variation was about 0.25 per cent. and was possibly largely caused by variations in the temperature of the final solution; a t-2” C change in this temperature causes a 20.3 per cent.change in optical density.2 Precision of the entire procedure-The entire procedure was repeatedly carried out on (a) sodium silicate solutions and (b) dilute clay suspensions (prepared from suspensions Kos. 1 and 2). For each solution or suspension, three portions were evaporated and fused in a batch (fusion batch) and the neutralised solutions from three fusion batches were treated together in a colorimetric batch. The optical densities of samples were corrected for reagent blank values and converted to p.p.m. of silica in the final solution; no corrections were made for the mean fusion blank value. The analyses were carried out during a period of 8 weeks, and statistical analysis showed no evidence for between-fusion batch and between-crucible effects except for the sodium silicate solution containing 0.211 p.p.m.of silica. The over-all standard deviations (those of one determination carried out in one colorimetric batch) were therefore calculated, and are shown in Table IV. The standard deviations for clay suspensions 2A and 2R are considerably greater than those for the other solutions. This was possibly caused by non-uniform samples ; the turbidity of suspension No. 2, from which the dilute solutions were prepared, supports this view, asJune, 19631 DETERMINATION OF SILICON I N WATER. PART I11 TABLE IV PRECISION OF ENTIRE PROCEDURE All concentrations are in the final solution Type of sample Fusion blanks . . .. Triple-distilled water . . Clay suspension No. 2Dt Sodium silicate solution Clay suspension No.lA* Clay suspension No. 2Ci Sodium silicate solution Clay suspension No. 2At Sodium silicate solution Clay suspension No. 2Bt Sodium silicate solution Silicon Volume found, Degrees ml silica freedom of sample, p.p.m. of of .. - 0.019 38 . , 20 0.023 24 . . 20 0.062 8 . . 20 0.211 8 . . 20 0.235 18 . . 20 0.226 7 . . 2 0.403 8 . . 2 0.426 7 .. 10 0.494 8 . . 10 0.533 8 . . 2 0.571 7 * Prepared by dilution of clay suspension No. 1. Prepared by dilution of clay suspension No. 2. Standard deviation, p.p.m. of silica 0.0026 0.0028 0.001 8 0.0015 0.0032 0.0034 0.0021 0.0062 0.0035 0.0056 0.0033 453 Coefficient of variation, Yo 14 12 2.9 0.7 1.4 1.5 0-5 1.5 0.7 1.1 0.6 it indicates the presence of relatively large clay particles.These results were therefore rejected as not being representative of the precision attainable by the method, although they give information about the homogeneity of the samples. The remaining results show that the standard deviation was relatively independent of silicon concentration in the final solution and of sample size. BIAS ARISING IN EVAPORATION AND FUSION STAGES- The fusion procedure must convert all forms of silicon that might be present in power- station waters to sodium silicate or monomeric silicic acid. A direct test of this is difficult to devise, as these forms are not known; some possibilities are quartz, clays, magnesium silicates and polymeric silicic acid. A further difficulty is that of preparing solutions of known concentration of the substances tested.However, three kinds of tests for bias were made and are described below. Fusion of standard szlbstances in powder form-The weighed dried powders were fused with 0.5 or 1.Og of sodium carbonate, the melt was partly dissolved in water and added to excess of acid as described under “Method,” and then diluted to 1 litre in a calibrated flask, The standard silicon solution used for calibration purposes was analysed gravimetrically (two dehydrations with perchloric acid), and the determined concentration was used in the calculztions; the results are shown in Table V. Although clay and felspar are the only two of the substances likely to be present in water, analysis of the other materials gives some idea of how the fusion copes with various siliceous substances.Part of the bias shown in Table V may be due to errors in the gravimetric analysis of the standard silicon solution, in which the final weight of silica was only 0.1 g. If the Spectrosil rod is regarded as a standard-its impurity content is less than 1 p.p.m.-the maximum bias would be reduced to -2.5 per cent. The remaining bias is relatively small, and the cause has not been traced further; the effect of dissolving the melt in acid may be relevant to this problem (see “Experimental,” p. 446). Fusion of sodium silicate solzttions-Fifteen 5-ml portions of standard silicon solutions containing 9.66 p.p.m. of silica were evaporated and fused as described under “Method.” They were then treated by the colorimetric procedure together with further 5-ml portions of the same standard silicon solution.After correction for reagent blank values and the mean fusion blank value, the fused samples gave results 0.2 & 1.0 per cent. lower than for those treated by the colorimetric procedure only. Fusion of clay solutions-If a series of solutions of decreasing concentration is prepared by progressive dilution of one master solution, the relative concentrations of these solutions Suitable portions were then analysed by the colorimetric procedure.454 MORRISON AND WILSON : THE ABSORPTIOMETRIC [AnaZyvst, Vol. 88 TABLE V TOTAL SILICON CONTENT 3 F STANDARD SAMPLES Weight of Silicon, as silica- Difference, Weight of sodium (-A- as percentage Sample sample, carbonate, certified, found, of silica g g % Yb con tent* Spectrosil roc1 (synthetic fused silica) 0.01000 0.5 100.0 98.8 - 1.2 American Bureau of Standards, Standard Samples- No.98, Plastic clay . . .. . . 0~01000 0.5 59.11 57.41 --.3.0 0*1000 0.5 57.6 0*1000 1.0 57.2 57.3 0~1000 1.0 57.5 0~1000 1.0 57.0 0.1000 1 .o So. 102, Silica brick . . . . . . 0*01000 0.5 0.1000 1.0 S o . 91, Opal glass . . . . . . 0~1000 1.0 67.53 65.5 - 3.0 S o . 99, Soda felspar . . . . . . 0*01000 0.5 68.66 g;} 67.1 5 - "2 93.94 ;;:;} 92.4 - 1.6 * For each determination, the 95 per cent. probability limits are estimated to be about kO.6 per cent. of the silicon content; the differences from the certified values are, therefore, statistically significant. will be known. Comparison of these known relative concentrations with those experimentally determined should reveal any bias that is not directly proportional to concentration.Such a series of results was obtained during the reproducibility tests carried out on dilute clay suspensions (see Table V). TABLE VI MEASURED CONCENTRATION OF SILICON IN CLAY SUSPENSIONS Table VI shows the relevant figures. Total silicon in sample h Volume of r \ Ratio of silicon found Suspension sample, Expected, Found,* to silicon expected,* 2 - Used as standard 1790 9 100 & 0.5 2A 2 20.0 20.4 0.2 102 1 2R 10 4.99 6.14 0.04 102.5 1- 0.8 2c 20 1 .oo 1.01 & 0.01 101 & 1 2 I> 20 0.20 0.20 111 0.0076 100 5 4 No. ml p.p.m. of silica p.p.m. of silica % * Error limits are for t t s a t the 96 per cent. confidence level. The results for suspensions 2A, 2B, 2C and 2D are consistent with one another, and, therefore, show no significant evidence of bias not proportional to concentration.However, they are not all consistent with the result for suspension No. 2. This could have been caused by bias, but, as it is not evident with the remaining suspensions, it is thought more likely to arise from the non-uniform nature of clay suspension No. 2, which could have caused an error when the concentrated solution was diluted. The results, therefore, indicate that, if bias is assumed to be constant, it is probably less than about 0.01 p.p.m. of silica in the sample. DISCUSSION OF THE WORK With 20-ml samples, the criterion of detection and the limit of detection, which may be related to the standard deviation of the fusion blank value: were about 0.02 and 0.04 p.p.m.of silica, respectively (95 per cent. confidence level; duplicate determinations). This is higher than ideally desirable, and it is useful to consider ways of improving the limit of detection by a factor of, say, 10. Examination of the precision figures given above shows that the largest contribution to the random errors at low concentrations was made by the fusion procedure; if its contribution had been zero, the criterion of detection would have been about 0.004 p.p.m. of silica (for 20-ml samples). Silicon in the sodium carbonate and con- tamination are thought to be the major sources of these errors, but as neither seems easy to reduce much further with a procedure similar to the proposed one, no more work was done on this aspect of the methodJune, 19631 DETERMISATION OF SILICOS I N WATER. PART I11 455 A more valuable approach would be to increase the sample size. The errors arising in the fusion stage would be unaffected, and thus errors in terms of the concentration in the sample would be reduced. We have not used samples larger than 20 ml, as the time required for their evaporation would then become rather long. However, greater volumes may always be used when necessary; increased contamination during the longer evaporation time may then place a limit on the expected improvement in precision. There is clearly a need here for a means of rapidly evaporating a large volume of water on a small area of platinum, which is then subsequently fused with sodium carbonate. The method is not suitable for use in the presence of more than about 0.05 to 0.1 p.p.m. of phosphate in the final solution.2 To determine total silicon in the presence of phosphate, the fusion procedure could be combined with a reduced p-molybdosilicic acid absorptiometric finish.2 The precision would then probably be worse, and special calibration samples would have to be taken through the entire procedure. The method is fairly convenient to use, but requires about three times as many manipu- lations as “reactive” silicon methods. Four 20-ml samples can be analysed in duplicate per day when 3 platinum crucibles are used. This paper is published by permission of the Central Electricity Generating Board. RE~ERENCES 1 . -. 3 - - , Ibid., 1963, 88, 88. 3. 4. ?;oTE-References 1 and 2 are to Parts I1 and I of this series, respectively. Morrison, I . R., and Wilson, A. L., Analyst, 1963, 88, 100. Anderson, L. H., .4cta Chem. Scand., 1958, 12, 495. Roos, J. B., Analjist, 1962, 87, 832. lieceived Febvuary 7th, 1963

ISSN:0003-2654    

DOI:10.1039/AN9638800446

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

456 CRUMP: ANALYSIS OF EDIBLE OILS [A~zalyst, Vol. 88 The Analysis of Edible Oils Contaminated with Synthetic Ester Lubricants BY G. B. CRUMP* (“Shell” Hesearch L t d . , Thornton Research Centre, P.O. Box 1, Chester) The analysis of edible oils contaminated or adulterated with synthetic ester oils used as lubricants (e.g., for aviation gas-turbine engines) presents the analyst with novel problems. Unlike mineral lubricating oils, the presence of which can readily be detected, synthetic oils have many of the properties of edible oils, such as high saponification values. Saponification and iodine values can be used as an early indication of contamination, and simple chemical techniques can be used to separate the contaminated oil into recognisable fractions. Properties are described of the saponification products of edible oils, synthetic ester lubricants and additives, such as tritolyl phosphate, as well as those of various unsaponifiable additives, such as phenothiazine and chlorinated biphenyl.Thin-layer silica-gel chromatography is used to distinguish between synthetic ester oils, glycerides and mineral oils. Paper Chromatography is used for detecting phenols (in the saponification products) and antioxidants containing nitrogen. Some applications of ultraviolet and infrared absorption spectrometry, mass spectrometry and gas - liquid chromatography are considered. THERE have been occasions, fortunately rare, when edible oils have become contaminated, accidentally or deliberately, with lubricating oils. Analytical investigation can be trouble- some because of the wide range of possible contaminants and the lack of any systematic method of isolating and identifying them.Although most lubricating oils consist mainly of petroleum mineral oil, which can be readily detected as a contaminant of fatty oils,lY2 some, such as aviation gas-turbine lubri- cant~,~,4Jj consist mainly of synthetic esters whose properties and reactions are similar to those of edible fatty oils. Moreover, most modern lubricating oils, whether of petroleum or synthetic origin, contain one or more of a wide variety of additives in sufficient quantity to complicate the analysis. This paper presents a scheme of analysis utilising the essential chemical differences between fatty and synthetic ester oils for separating a mixture of such oils into identifiable fractions, and at the same time isolating and identifying additives likely to be present.The scheme not only gives an early qualitative indication of contamination, but permits its quantitative determination. CHEMICAL CHARACTERISTICS OF VARIOUS EDIBLE OILS The two most immediately useful parameters for assessing the purity of edible oils are saponification values and iodine value.’ Some typical values are listecf in Table I. TABLE I SAPOXIFICATION AND IODINE VALUES OF SOME EDIBLE OILS* Saponification value, Iodine value, 0 il mg of potassium hydroxide g of iodine per Per 6 100 g of oil F’alm (kernel) . . .. 242 to 262 14to 19 Coconut . . .. .. 255t 7.0 to 9.5 Maize . . .. .. 188 to 193 103 to 125 Olive . . .. . . .. l88t 79to 88 Groundnut .. .. . . l88t 82 to 99 Sesame . . .. .. 188 to 193 105 to 118 Soya bean . . .. .. 190t 129 to 143 Cottonseed . . .. .. 192t 103 to 113 *Values taken from B.S. 628-32 and 650-56 : 1950. t Not less than. * Present address : “Shell” Research Ltd., Central Laboratories, Whitehall Lane, Egharn, Surrey.June, 19631 CONTAMINATED WITH SYNTHETIC ESTER LUBRICANTS 457 In general, whereas edible glycerides are esters of monocarboxylic acids (generally C,, to C,,) and glycerol, synthetic ester oils33415 are formed from- (i) monohydric alcohols and dicarboxylic acids ; (ii) polyhydric alcohols (NOT glycerol) and monocarboxylic acids (generally C, to C,). Synthetic ester oils have saponification values greater than 250 and iodine values of If a synthetic ester oil contaminates an edible oil, the saponification value of the The change of saponification and zero. latter will increase and its iodine value will decrease.* iodine values with degree of contamination is considered later (p.459). CHEMICAL CHARACTERISTICS OF SYNTHETIC ESTER OILS The major components of most synthetic ester oils are compounds of the types recorded Some typical synthetic aviation turbine oil compositions are shown in Table 111. in Table 11. DISCUSSION OF TABLE II- of various synthetic ester oils. for glycerides. 884 for glycerin trioleate). Table I1 records molecular weights, saponification values and saponification products The saponification values are generally higher than those The molecular weights are much lower than those of the glycerides (about Formula Monohydric akohols- Dioctyl sebacate C8H17OOC-(CH,),-COOC,H,, Dinonyl sebacate CgH1900C-(CH,)8-COOCgH19 Dioctyl azelate CaH1700C-(CH,) 7-COOCsH17 Dinonyl azelate Dioctyl adipate CgH1g00C-(CH,)7-COOCgH19 C8H17OOC-(CHs)*-COOCsH,, Dinonyl adipate C9H1gOOC-( CHJ,-COOC9H19 Trihydric alcohols- 1, 1, 1-Trimethylolpropane trivalerate CH,00CC4Hg C,H5-L-CHzoOCC4H9 I CH,OOCC,H, TABLE I1 ESTERS Molecu- Mols of Saponi- lar KOH fication weight per mol value Hydrolysis products 426 2 263 2-Ethylhexanol Sebacic acid 454 2 247 3,5,5-Trimethylhexanol Sebacic acid 412 2 2i2 2-Ethylhexanol or 2,2,4- Trimethylpentanol Azelaic acid 440 2 370 2 398 2 386 3 255 3,5,5-Trimethylhexanol Azelaic acid 303 2-Ethylhexanol Adipic acid Adipic acid 2 82 3,5,5-Trimethylhexanol 435 1,1,l-Tr:methylolpropane Valeric acid 1,1,l-Trirnethylolpropane tri-n-heptanoate CH200C(CH,),CH3 470 3 357 1, 1,l -Trimethylolpropane n-Heptanoic acid I I C,H,-C-CH,OOC( CH2),CH3 CH,OOC(CH,),CH, * However, some synthetic ester oils, e.g., dioctyl sebacate, have iodine values of zero and saponifi- cation values close t o those of edible oils, such as coconut oil, that contain high proportions of the glycerides of lauric and myristic acids.458 CRUMP: ANALYSIS OF EDIBLE OILS TABLE I1 : EsrExs-continued [Analyst, Vol.88 Forniu la 1, 1 , 1-’rrimethylolpropane tri-n-nonanoate CH200C(CH2),CH3 I I C2F15-C-CH200C (CH,),CH, CH,OOC(CH,) ,CH3 Xeopentyl di-n-nonanoate CH, I I CH3 CH3(CH2).,COO-CH,-C-CH2-OOC (CH,),CH, Neopentyl di-n-hexanoate CH3 1 I CH3 CH,( CH2),COOH2C-C-CH,OOC(CH2),CH3 Polyhydric alcohols- Pentaerythritol tetra-n-heptanoate CH,OOC( CH2),CH3 I i CH3 (CH2)5COOH2C-C-CH200C (CH2)5CH3 CH200C(CH,),CH3 Pentaerythritol tetra-n-hexanoate C H200C6Hl, I I C,H,,COOCH2-C-CH20OCC5H11 CH200CC5H,, Pentaerythritol tetrabutyrate CH200CC,H,CH3 I I C3H,COOCH2-C-CH20OCC3H, CH200CC3H, Molecu- 3101s of Saponi- lar KOH fication weight per mol value Hydrolysis products 534 3 304 1, 1, I-Trimethylolpropane n-Xonanoic acid 3x4 2 328 2 584 4 528 4 416 4 292 Xeopcntyl glycol n-Nonanoic acid 342 Xeopentyl glycol n-Hexanoic acid 384 Pentaerythritol n-Heptanoic acid 124 Pentaerythrito; n-Hexanoic acid 538 Pentaerythritol n-Butyric acid TABLE I11 TYPICAL SYNTHETIC ESTER OIL FORMULATIOSS -4ND COMPOSENTS 1.Uioctyl sebacate . . .. . .. . . . .. .. . . major constituent * * } minor constituent . * Phenothiazine . . . . .. . . . . .. .. Tritolyl phosphate . . . . .. . . . . . . .. .. 2. Dinonyl sebacate . . . . . . . . .. .. .. . . major constituent Thick ester based on polyoxyethylene glycol 200 31, sebacic acid and minor constituent Phenothiazine . . .. .. .. .. .. . . .. . . minor constitucnt 3. Dioctyl sebacate . . . . .. .. .. .. .. . . major constituent Thick cster based on 2-ethyl-2-methylpropane-1,3-diol, sebacic acid minor constituent Phenothiazine . . . . .. .. .. . . .. . . . . minor constituent .. .. .. I= . . . . .. . . I= 2-ethylhexanol . . . . .. .. .. and 2-ethylhexanol . . . . . . 4. (a) Thick esters C8H17OOC(CH2),CO [ (P.E.G. *)OOC(CH2)8CC)!nOC8H17 ( b ) C~H~700C(CH~)~CO[OC~H~~OOC(CHJ~CO~nOC~H~~ In both ( a ) and (h) the values of n can range from 1 t o 6.* P.E.G. I= polyoxyethylene glycol.TABLE IV SCHEME OF ANALYSIS OIL (Mixture of edible and synthetic oils) Dilute with 10 volumes of <40° C light petroleum I I I LATE (I ) Shoke with saturated brine (2) Extract with ether RES ETHEREAL PHASE Evaporate of the ether FRACTION I r Octanols 1 BRINE SOLUTION (Including ethanol) REjECT DUE, (K+ Salts; GI ycols) (I ) Acidify with HaO, (2) Ether extract (continuous) L and nonanols J ETHEREAL NOTE I PHASE FRACTION 2 [Dicarboxylic acids] glycerides, antioxidants, etc. 1 SOL~BLES Mainly simple esters, [ (I ) Distil af light petroleum (2) Saponify with 0.5 N ethanolic KOH (3) Steam distil. DISTILLATE K+ Salts of mono and dicarboxylic acids K+ Salts of phenols K+ Salts of phosphoric acid Glycerol (I ) Treat with brine (2) Ether extract Glicols Hydrocarbons Chloro hydrocarbons ETHEREAL PHASE BRINE SOLUTION FRACTION 4 DISCARD r Octanols and 1 I Extract with 60' to 80' C light petroleum L nonanols J NOTE 2+ LIGHT PETROLEU'M PHASE (NOTE 8) ALKALINE (Aq) PHASE (K+ Salts; I ACID (Aq) PHASE FRACTION 3 [ G I ycol s] NOTES- I and 2 If 2-isopropyl-2-methylpropane-1,3-diol or 2,2-dimethylpropane-I .3-diol are the esterifying glycols, part of them will be steam volatile.On evaporating the ether phase to low volume the diols will be precipitated. + If tritolyl phosphate had been present, phenols would appear in this fraction. 3 (i) On saponification certain chlorine compounds are converted into highly soluble ( i i ) Dipentaerythritol has lower water solubility and may be partly precipitated. pol yo Is.4 5 and 6 7 8 9 If phenothiazine is present it will also appear in later fractions. Ion-exchange resins may be used here t o remove undesirable ions. Solvents of higher polarity than benzene may be necessary to elute some antioxidants. On gassing this solution with dry HCI the hydrochlorides of most secondary aromatic amines present will be precipitated. Part of this material can also appear in Fractions 5 and 6. with benzene with acetone [Amine type] [Polyalkylene] FRACTION 5 FRACTION .7 antioxidants oxides NOTE 4 NOTE 7 NOTE 9 I RESIDUE Extract with ether 1 ETHEREAL PHASE (I ) Evaporate to dryness (2) Extract with light polyhydroxy cmpds.) (I ) Acidify with (2) Steam distil sulphuric acid I DISTILLATE ACID (Aq) PHASE FRACTION 9 Pol yg I ycol s, [ glycols, ] glycerol NOTE 5 SOLUBLES INSOLUBLES FRACTION I I FRACTION 10 [Monocarboxylic (fatty) acids] [Dicarboxylic acids] (I ) Dissolve in excess of NaOH (2) Saturate with CO, ( 3 ) Ether extract I 1-1 (I) Acidify with sulphuric acid ETHEREAL PHASE ACIDIC PHASE FRACTION 8(b) DlSCAR D Monocarboxylic acids Low molecular weight]June, 19631 CONTAMIS-ATED WITH SYNTHETIC ESTER LUBRICANTS 459 PKELIMIKAKY MEASUREMENT OF CONTAMINATIOS Indications of contamination can often be obtained during the early stages of an analysis.Many synthetic ester oils after exposure to air develop characteristic colours ; for instance, an oil containing phenothiazine can become deep red. The presence of substituted naphthyl- amines confers a bloom to an oil.Synthetic esters occasionally contain traces of free alcohols (e.g., 2-ethylhexanol) and these impart characteristic sweet odours. Further, if after the determination of saponification value the saponification liquor has an ethereal odour (other than that of ethanol), the presence of octanols and nonanols is indicated. The determination of such quantities as the saponification value, iodine value and phosphorus content of a suspect edible oil frequently allows the degree of contamination to be calculated, particularly when this is high. I t is desirable that the saponification value and iodine value of an unadulterated specimen of the oil should be known. The type of calculation is shown below. ( a ) E f e c t of dioctyl sebacate on the saponification value of olive oil-Contamination of olive oil by dioctyl sebacate (10 per cent.by weight) raises the saponification value from 192 (for example) to 199. If the saponification values for samples of olive oil received from the same area were normally 192, then this difference of 7 units would suggest contamination or admixture with another glyceride, e.g., palm kernel oil. (b) Phosphorus content of olive oil coiztaminated witJz a syntlzetic ester oil containing tritolyl phosphate-If a contaminating ester oil contains dioctyl sebacate and tritolyl phosphate (5 per cent. by weight), determination of the phosphorus content would give a figure of 0.043 per cent. by weight for 10 per cent. by weight contamination. XoTE-It is unlikely that the analyst will encounter phospholipids in refined edible oils, hence the presence of phosphorus is usually an indication of contamination.How- ever, some unrefined oils, e.g., unrefined soya bean oil, contain phosphatides that on saponlfication yield glycerol, fatty acids, phosphoric acid and choline. Most other synthetic ester oils will alter drastically the saponification values of most edible oils. However, the saponification values of an edible oil and an ester oil can be similar (e.g., palm oil 242 to 252 and a dioctyl sebacate - tritolyl phosphate oil 260), and, in this instance, the saponification value is not a criterion of purity. Further, palm oil has a low iodine value and contamination with, for example, dioctyl sebacate would not alter it greatly.As before, a high phosphorus content would render such an oil suspect, but otherwise it is necessary to carry out more extensive analysis to detect contamination. The presence in edible oils of contaminant additives such as tritolyl phosphate, diphenyl- amine, phenothiazine, silicon compounds and titanium compounds introduces alien elements. Given sufficiently sensitive methods, these elements can be quantitatively determined at low concentrations, and hence provide proof of contamination. SCHEME OF ANALYSIS A scheme for analysing mixtures of glycerides with synthetic esters is presented (see Table IV); it takes advantage of the ease of saponification of carboxylic esters. I t permits both saponified and unsaponified materials to be separated into groups of compounds having common functional groups.Such compounds can frequently be isolated in a high degree of purity, making it possible to identify them by quantitative elemental microanalysis, by derivatives and by characterisation reactions. The nature of the original mixed esters can be ascertained from the composition of the saponification products. The scheme of analysis is designed to meet most eventualities, and can often be simplfied on the evidence provided by the preliminary analyses. If the aim of the analyst is to give a rapid qualitative assessment of contamination, the determinations listed below will generally serve his purpose- ( a ) Saponification value. ( b ) Iodine value. (c) Qualitative analysis of component acids. (d) Qualitative analysis of hydroxy compounds.The isolation of such compounds as amines, phenols and chlorohydrocarbons serves as Although these compounds may be difficult to charac- further confirmation of adulteration. terise fully, this in no way impedes the ester characterisation.460 CRUMP: -4E;ALYSIS OF EDIBLE 011,s [AnaZyst, Vol. 88 \!'hen members of homologous series of organic compounds, such as methyl esters of normal monocarboxylic acids, occur toget her, particularly when their boiling-points are close, separations are often extremely tedious, involving fractional crystallisation, vacuum dis- tillation, etc. In these instances the more recent techniques, particularly gas - liquid chromatography, have come into prominence.8 These methods are considered later (p. 465).ANALYSIS OF FRACTIOXS ISOLATED ACCOKDIKG TO TABLE IV To date few other aliphatic alcohols appear to have been used in synthetic ester lubricants; sometimes a complex mixture of alcohols will be isolated, usually C, to C14. The boiling-points of octanols and nonanols are close together, but the compounds can be separated by distillation. Their purity can be assessed from-- These will usually be mixtures of octanols and nonanols. (i) Boiling-poin t determinations. (ii) Active hydrogen or hydroxyl content (by a~etylation).~ Some properties of these alcohols are shown in Table V. TABLE 1' ?llONOHYDRIC ALCOHOL COMPONEXTS OF SYNTHETIC ESTER LUBRICANTS Molecular Boiling-point, Hydroxyl content, Alcohol weight "C yo by weight 2-Ethylhexanol . . .. . . 130 184 13.1 2,2,4-Trimethylpentanol .. .. 130 165 13.1 3,5,5-Trimethylhexanol . . . . 144 194 11.8 FRACTIONS 2 AND 10- Determinations of equivalent-weights and melting-points can be used to identify pure acids, and will indicate the presence of mixtures. Methylationlo $11 of the acids with diazomethane, boron trifluoride - methanol or methanol - sulphuric acid will yield methyl esters, which can be separated by fractional distillation. Three dicarboxylic acids are commonly encountered in this type of analysis ; they are listed in Table VI with some of their properties. These fractions will consist of dicarboxylic acids. TABLE VI PROPERTIES OF SOME DICARBOXYLIC ACIDS Molccu lar Roiling-point of Melting-point of Acid weight Melting-point, diethyl ester, P-nitrobenzyl ester, "C "C "C ..idipic .. . . .. 146 152 241 106 Azelaic . . . . .. 188 106 29 1 44 Sebacic .. . . .. 202 134 308 73 FRACTIOS 5- This fraction will consist almost entirely of antioxidant materials containing nitrogen. These can be characterised by various colour reactions, melting-points, derivatives and nitrogen contents. Table VII contains details of the materials likely to be encountered in this fraction. They can also be analysed by paper chromatography.12 Small portions of the nitrogen compounds (or the suspect oil) dissolved in acetone are spotted on a Whatman No. 1 paper impregnated with dipropylene glycol. The chromatogram is developed with cyclohexane saturated with dipropylene glycol. The individual compounds are located by spraying the paper with a solution of 9-nitrophenyldiazonium fluoroborate in acetone.The resulting azo dyes have characteristic colours, which, together with R, values, permit the amines to be identified. FRACTION 6- and hydrocarbon oils rarely occur together.) This fraction consists mainly of aromatic chloro-compounds. (Chlorinated polyphenyls The presence of chloro-compounds is readilyJune, 19631 COKTA4MINATED WITH SYNTHETIC ESTER LUBRICANTS 461 indicated by means of Beilstein's copper wire test. Elemental analysis will confirm an aromatic nucleus. Cl c1 A typical material likely to appear in this fraction would be- The elemental analysis is C, 49.3 per cent.; H, 2.1 per cent. ; C1, 48.6 per cent. The low Any mineral oil present in a contaminated hydrogen content indicates an aromatic nucleus. edible oil could appear in this fraction.TABLE VII PROPERTIES OF SOME ANTIOXIDANTS CONTAIKING NITROGEK Compound Diphenylamine -Melting- Nitro- Sul- Melt- point of hlole- gen, phur, ing- deriva- cular %.by %.by point, tive, weight weight weight "C "C Other details* 169 8.28 -- 54 S-scetyl RF 0.55 198 219 6.39 - 62 c16H13x N-phenyl- I-naph- thylamine N-acetyl Rp 0.57 115 H Phenothiazine H I /\%/\ 3,7-Dioctylpheno- I I! !I I thiazine /\/-'\/\ C8H,, C8H17 H ~- Di-(p-octylpheny1)- C,H,,-/=\-h -/=\C8Hli 393 3.56 - amine \-/ \-/ * RF values obtained by using Delves system.12 219 6.39 - 108 S-acetyl RF 0.43 93 Ratio sulphur tonitrogen = 199 7.03 16.1 185 - 2.29 RF 0.19 Intense carmine 423 3.31 7.56 168 - colour with conc. H,SO, RF 0.82 104 HF 0.86 FRACTION 7- This fraction will contain compounds such as polypropylene oxide, polyethylene oxide or copolymers of ethylene oxide and propylene oxide- where n and m may be any integral number, but values between 4 and 20 are to be expected.Polyethylene oxide [HO(C,H,O),H] has a carbon t o oxygen ratio of between 1 and 2, whereas polypropylene oxide has a carbon to oxygen ratio of between 1.5 and 3, the exact values being dependent on molecular weight. HO(C3HGO),(C2H40)rnH462 CRUMP: ANALYSIS OF EDIBLE OILS [AnaZyst, VOl. 88 A copolymer of both oxides will have a carbon to oxygen ratio falling within the range 1 to 3, but its molecular weight will be too low for the corresponding value for polypropylene oxide and too high for the corresponding value for polyethylene oxide. Thus a knowledge of the molecular weight and carbon to oxygen ratio of a polymer can assist in itscharacter- isation.On oxidation with a potassium dichromate - sulphuric acid mixture, polypropylene oxide and its copolymer with ethylene oxide give a copious evolution of acetaldehyde. A pure ethylene oxide polymer undergoes rapid oxidation, but does not form acetaldehyde. Polj-- ethylene oxides are usually readily soluble in water, whereas polypropylene oxides are not. FRACTION 8(a)- This fraction contains phenolic material, which will only be present if aryl phosphates or alkylaryl phosphates occurred in the untreated oil. It will have a characteristic carbolic odour, and give a green or purple colour with aqueous ferric chloride solution. In all prob- ability it will be a complex mixture of phenol, cresols, xylenols and higher phenols.Resolu- tion of such a mixture by conventional means is almost impossible on the small amounts of material isolated. However, it can be analysed readily by paper chromatography.13 The phenols are coupled in alkaline media with diazotised p-nitroaniline ; the solution is then acidified with dilute mineral acid and the dyes are extracted with ether. A small portion of the azo-dye mixture is spotted on a Whatman No. 1 paper impregnated with 20 per cent. v/v of formamide in acetone and developed with cyclohexane - benzene solution (70 + 30 bj- volume) containing 3 per cent. v/v of dipropylene glycol. The azo-dyes are characterised by R, values and the colour reactions of the spots with ammonia or aqueous sodium carbonate.Gas - liquid chromatography can also be used to resolve small amounts of complex mixtures of phenols.'* FHACTIOX 8(b)- most likely to be encountered are mentioned below. (;) water-soluble acids n-butyric acid isovaleric acid n-valeric acid n-hexanoic acid (ii) water-insoluble acids n-heptanoic acid n-octanoic acid n-nonanoic acid This fraction contains any low molecular weight carboxjdic acids present. They can be divided into two groups- The acids The two groups can be further separated into their pure components by distillation of the methyl esters. The acids isolated will usually be a mixture, e.g., hexanoic, heptanoic and nonanoic acids, so that equivalent-weight determinations will give average values. Gas - liquid chromatographic separations of the methyl esters15 is the most rapid method of quantitative analysis for each component.FRACTIONS 3 AND 9- These fractions contain polyhydroxy material. If the oil being tested contains any fat a t all, then glycerol will be present and can be oxidised quantitatively by potassium periodate. This fraction also contains potassium salts and excess of mineral acid, which must be neut ralised before proceeding further. The aqueous solution of salts and hydroxy compounds is evaporated to dryness, under vacuum, on a steam-bath. The resulting salt cake is dried by azeotropic distillation with anhydrous isopropyl alcohol, and the hydroxy compounds are then dissolved with the same solvent. (Alternatively, the ionic material can be removed by percolating the saline - acid solution through ion-exchange resins, e.g., Amberlite 400 (OH-) and Amberlite 120 (H+).) The isopropyl alcohol is distilled off, and any hydroxy compounds remaining can be characterised. Any glycerol present can be destroyed by periodate oxidation (prolonged treatment can also degrade pentaerythritol) .Table VIII shows the properties of t h e most common hydroxy compounds that will be encountered in this fraction.June, 19631 CONTAMINATED WITH SYNTHETIC ESTER LUBRICANTS 463 FRACTION 11- This fraction contains high molecular weight monocarboxylic acids derived from glycerides. Some properties of these acids are shown in Table IX. TABLE v m PROPERTIES OF SOME POLYHYDROXY COMPOUSDS hlolccular Melting-point, Boiling-point, OH, .yo Compound Formula weight “C “C by weight C3H, i propane- 1,3-diol I CH3 132 58 - 25.8 2-Isopropyl-Z-methyl- IIOH2C-C-CH,OI3 CH3 I 1,3-diol I CH3 CZH, I I I I I 104 123 2,2-Dimethylpropane- HOH,C-C-CH,OH 146 Liquid 2-Ethylhcxane-1,3-diol HOH,C-C-CH-C,H, H OH CH,OH 134 59 1,1,1-Triniethylolpropanc C,H,-C-CH,OH CH,OH CH,OH Pen taery thritol HOH,C-C-CH,OH 136 261 to 262 I I CH,OH Dipen taerythri to1 P.E.G.200 C H,O H I I CH, I 0 I H OCH,-C-CH,O FI CH, I 2 54 222 134 (5 mm) HOCH,-&H,OH I CI I,OH hIainly tctra- ethylene-glycol 194 Tiquid 328 32.7 83.3 38.1 60.0 40.1 17-5 TABLE IX PROPERTIES OF HIGH MOLECULAR WEIGHT FATTY ACIDS Molecular Roiling-points of rlcid \v eigh t methyl esters, “C Other properties Lauric . . . . . . 200 87 ( 1 mm) p-bromophenacyl ester m.p. 76” C Palxnitic .. . . . . 856 111 ( 1 mm) p-bromophenacyl ester m.p. 84” C Stearic . . . . . . 270 130 ( I mm) p-bromophenacyl cstcr m.p. 90” C Iinolcic . . . . . . 266 168 to 170 ( 1 mm) Tetrabromo-derivative n1.p. 180’ C Oleic . . . . . . 268 160 to 162 (2 to 3 mm) - The chemical separation and identification of these acids in mixtures is extremely tedious. The application of gas - liquid chromatography8 *16 and sometimes thin-layer chromato- g r a p h ~ ~ ~ quickly gives both qualitative and quantitative identification of each component acid.464 GRUMP: ANALYSIS OF EDIBLE OILS [Analyst, Vol. 88 Let us assume that an olive oil (saponification value 192) contains 50 per cent. by weight EXAMPLE OF AK ANALYSIS of a synthetic ester oil having the composition- 94 per cent.by weight of dioctyl sebacate 5 per cent. by weight of tritolyl phosphate ] Saponification value about 260” 1 per cent. by weight of phenothiazine J The saponification value of the mixed oils will be 226, and the saponification liquor will have a strong odour of octanols. The iodine value (say 82) of the olive oil will be decreased to 41 by the synthetic ester. The phosphorus content of the oil will be 0.210 per cent. by weight. In all probability the untreated oil mixture will also have an odour of octanols and will be red. The fractions listed below will be obtained when the oil is analysed- Fraction 4-Octanol (2-ethylhexanol) . Fraction 5-Phenothiazine. Fraction 8(a)-Phenols. Fraction 10-Sebacic acid. Fraction 1 l-C16, C,, saturated and unsatqated fatty acids.Fraction 9-Glycerol. (The silica gel column used to prepare this fraction will be pale green after elution with acetone as phenothiazine is present.) Fractions 4, 10 and 11 can be weighed and fraction 9 determined by periodate oxidation. The phenothiazine can be characterised by paper chromatography (on the original oil mixture) and determined via the sulphur content of the oil (0-5 per cent. by weight of phenothiazine ~ 0 - 0 8 per cent. by weight of sulphur). Tritolyl phosphate can be determined from the phosphorus content of the oil (1 per cent. by weight of tritolyl phosphate ~0.084 per cent. of phosphorus), and the constituent phenols can be characterised (after saponification) by paper chr~matographl-.~~ In balancing the weights of the fractions, allowance must be made for the water gained on saponification, e.g.- COOH, COOH // / R + 2H,O + K -+ 211,OH \ COOH, ‘COOH If this allowance is not made the mass balance should exceed 100 per cent.recover>- (a low recovery is probably due to the volatility of fraction 4). RECOVERY OF TRITOLYL PHOSPHATE- If it is required to concentrate the tritolyl phosphate this can be achieved b!- percolating the suspect oil through silica gel. Tritolyl phosphate is strongly adsorbed, whereas most of the esters are not and can be eluted with benzene. Some glycerides however are eluted with tritolyl phosphate, e.g., glycerides of erucic acid found in mustard oils. The tritolyl phosphate is dcsorbed with acetone, and about a five-fold concentration is achieved, which facilitates mass spectrometric analysis.A quantitative assessment of the molecular-weight distribution of aryl phosphates present can thereby be obtained (see Table X). TABLE X MASS SPECTROMETRIC ANALYSIS OF A COMMERCIAL TRITOLYL PHOSPHATE Molecular Distribution, Phosphorus compound Phosphorus, U/ weight O’O (typical of molecular weights in column 1) / O 340 0.1 Diphenyl tolyl phosphate 9-12 382 54-0 Ditolyl xylyl phosphate 8-12 410 0.4 Trixylyl phosphate 1 - 0 1 421 0.3 Higher phosphate 7.33 354 0.5 Phenyl ditolyl phosphate 8.T5 368 41.5 Tritolyl phosphate 8.43 5.84 396 2.9 Tolyl dixylyl phosphate r “7 * Tritolyl phosphate will also be saponified.June, 19631 CONTAMINATED WITH SYNTHETIC ESTER LUBRICANTS 465 MORE RECENT ASPECTS OF ANALYSIS Although the scheme of analysis outlined permits the edible oil contaminants to be characterised within narrow limits and allows a quantitative determination of contamination to be made, it still poses analytical problems that are not readily solved.For example, the complex mixture of acids isolated as Fractions 10 and 11 can only be separated chemically by tedious means, and fractional distillation is slow. Further, such methods are usually only semi-quantitative. The quantitative analysis of such mixtures of acids has been radically simplified by the use of gas - liquid chromatography,8y15y16 which will separate the methyl esters of mixtures of closely related acids, such as azelaic and sebacic acid, or stearic, oleic, linoleic and linolenic acids. Gas - liquid chromatography can also be used to separate octanols and nonanols (e.g., Fractions 1 and 4), and resolve a complex mixture such as cresylic acid14 (Fraction 8(a)).If the contaminant esters are dioctyl sebacate, dinonyl sebacate or dinonyl adipate they can be identified by gas - liquid chromatography without the need for hydrolysis. Pheno- thiazine, diphenylamine and possibly substituted diphenylamines can also be determined by this technique. A valuable method of determining secondary aromatic amines is ultraviolet spectroscopic analysis. Minute concentrations of amine compound, e.g., (0.01 per cent. of phenothiazine, can be determined in the untreated oil. Although some unsaturated carboxylic acids also absorb in the ultraviolet region, they do not interfere. Infrared spectroscopic analysis affords clear preliminary indications of the major func- tional groups present in the contaminated oil and in the various fractions, e.g., the aromatic character of chlorinated biphenyls is readily shown.Mass spectrometric analysis can be extremely useful in the examination of such com- ponents as nitrogen compounds and aryl phosphates. A rapid assessment of the possibility of contamination can be obtained by the use of thin-layer silica-gel chromatography.Is Edible oils and synthetic ester oils give characteristic chromatograms that can be readily distinguished, and hence the presence of the synthetic ester is detected. Further, when the chromatogram is sprayed with 9-nitrophenyldiazonium fluoroborate, most antioxidants present can be revealed as brightly coloured zones. 1. 2. 3. 4. 5. 6. s. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. I- REFERENCES Scott, W. VT., and Furman, N. H., “Standard Methods of Chemical Analysis,” Fifth Edition, D. Van Nostrand Co. Inc., New York, and The Technical Press Ltd., London, 1939, Volume 11, p. 1774. Hadorn, H., and Jungkunz, R., M i t t . Lebensmitt. Hyg., Bern., 1949, 40, 96. Pearson, J . W., and Waight, F. H., EYdol u. Kohle, 1961, 14, 527. Elliott, J . S., and Edwards, E. D., J . I n s t . P e t . , 1961, 47, 39. Panov, V. V., and Sobolev, Yu. S., “Lubricating Oils for Aviation Gas Turbines,” W.A.S.A. British Standards 628-32 and 650-66 : 1950. Ibid., Appendix E. Fontell, K., Holman, R. T., and Lambertsen, G., J . Lzpzd Res., 1960, 1, 391. Mitchell, J., jun., Kolthoff, I. &I., Proskauer, E. S., and Weissberger, A., Editovs, 09. cit., 1953, Volume I, p. 30. Metcalfe, L. D., and Schmitz, 4. X., Anal. Chewa., 1961, 33, 363. Vorbeck, M. L., Mattick, L. R., Lee, F. A., and Pederson, C. S., Ibid., 1961, 33, 1512. Delves, R. B., J. Inst. P e t . , 1962, 48, 283. Crump, G. B., J. Chromatog., 1963, 10, 21. Brooks, V. T., Chem. & I n d . , 1959, 1317. James, A. T., and Martin, A. J . P., Bzochem. J . , 1960, 50, 679. Craig, B. XI., Chem. 6 Ind., 1960, 1442. Morris, I,. J., Ibid., 1962, 1238. Crump, G. B., Nature, 1962, 193, 6i.2. Technical Tvanslation F-21, May 1960. “Vegetable Oils.” Appendix F. Received Septembev 13th, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800456

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

466 LUNT: RAPID ALKALINE REDVCTIOX OF NITRO-GROUPS [Analyst, Vol. 88 The Rapid Alkaline Reduction of Nitro-groups and the Kjeldahl Determination of Nitrogen BY T. G. LUT\JT* (Faculty of Technology, Univettsity of Manchestev) Alkaline stannite rapidly reduces certain nitrophenols so that the nitrogen can be quantitatively determined by the Kjeldahl method. The method may permit easier analysis of moisture-sensitive compounds. This appears to be the first recorded pre-treatment in alkaline solution, although many acidic methods have been described. These methods are reviewed briefly, with special reference to the determination of nitrogen in isomeric nitrophenols. ALTHOCGH the Kjeldahl nitrogen determination was discovered in 1883 and now incorporates additional features to broaden its scope, there are still certain types of nitrogen compounds with the element in a more oxidised form, and a few heterocyclic structures, pyridine and acridine,l which do not yield all their nitrogen as ammonium sulphate in any of the Kjeldahl modifications.Various pre-treatments are used for nitro-, nitroso-, azo-, oximino- and similarly bound nitrogen compounds. Most are acid reductions: with sulphur and sulphuric acid2; zinc and fuming sulphuric acid3; iron filings and sodium osalate and sodium formate in sulphuric acid4; stannous chloride and hydrochloric acid in a sealed tube a t 120" C5s6; steel filings, copper oxide and 18 N sulphuric acid7; zinc and hydrochloric acid or a non-oxidising acid8$9$10; alcoholic sodium dithionite (Na2S204)11 9 l 2 ; activated copper and zinc couple13 ; glucose, on a semi-micro and micro scalel49l5; but compare Takeda and Senda16 and Abrahamczikl' ; chromous chloride, obtained as required from chromic chloride and zinc amalgamla; titanous chloridelg; hypovanadous salts20; phosphonium iodide21 ; hydrogen and a nickel catalyst.22 Acid reduction in the final stage has also been employed.23 The work of Margosches and co-workers has shown that o-hydroxynitro-compounds are degraded completely, whereas a $oriented hydroxy-group does not destabilise the ring sufficiently for quantitative analysis ; m-isomers are similar to the latter.24 These workers also showed that addition of phenolsulphuric acid without heating, for an initial period, afforded quantitative analyses of nitrates.25 Resorcinol or phloroglucinol were also shown to sensitise nitrates to degradation.26 This is doubtless due to nitration of the strongly nucleophilic phenols.The use of salicylic acid,23 o-mercaptobenzoic acid27 *28 or a mixture of equal amounts of 1-naphthol and pyr~gallol~~-see also Tirouflet30-has been found to ensure quantitative decomposition of aromatic nitro-compounds. Their successful use is to be connected with reversible de-nitration of the nitro-compound, with subsequent more rapid nitration of the phenols, which are then decomposed by the medium of hot sulphuric acid. The thiol group of the acid C6H4(SH)C0,H exerts a separate reducing action. Some of these pre-treatments employed have also found use with osazones31; azo- compoundsll ,19 932; hydrazine and its uni- and bivalent substituted products.7 s33 s34 Other Kjeldahl pre-treatments employed are addition of formaldehyde to an azide degradation to prevent loss of gaseous nitrogen3j ; conversion of organic thiocyanates to cuprous thiocyanate, which is then more readily ; gradual oxidation with manganese dioxide in the Kjeldahl determination3' ; oxidation with potassium permanganate below 8Fj0 C, effective also with pyridine compounds, to animonium and nitrate ions.The latter are then reduced with a Raney alloy38; alkaline peroxide also aids decomposition of p y r i d i n e ~ ~ ~ ; addition of benzoic acid in the degradation of benzonitrile, benzophenoneoxime and azo- b c n ~ e n e , ~ ~ and addition of cuprous 0xide.j.l During the analysis of some organo-boron compounds obtained from boron trichloride and phenols and phenols with 0- or 9-nitro-groups or both, e g ., (N02.C6H,0),BC13-,, where x was 1 or 3, the determination of boron was hindered on account of volatile nitrophenols distilling over with methyl borate.42 9-Xitrophenol as well as the o-isomer and 2,4-dinitro- phenol are sufficiently volatile in methanol to collect in the condensate and, because of the * Present address: 342 Poole Road, Branksome, Poole, Dorset.June, 19631 AND THE KJELD-AHL DETERMIEATION OF NITROGEN 467 deep yellow colour of the phenolate ion, obscure the phenolphthalein end-point of a titration of boric acid and mannitol with sodium hydroxide. To prevent this, and to obtain the nitrogen in a form suitable for determination by the Kjeldahl method, the nitro-compound (0.2 to 0.4 g) was reduced with alkaline stannite.Approximately 1.5 times the stoicheiometric weight of stannous chloride liexahydrate neces- sary for complete reduction to amine was dissolved in the minimum of hydrochloric acid with warming. After the solution had been cooled, 30 per cent. sodium hydroxide solution was added carefully, with shaking and cooling in cold water (under a running tap). At first, crystalline sodium chloride was precipitated; almost at once, on further addition of alkali, a pale yellow curdy precipitate was formed, and no further alkali was added after the precipitate had dissolved. This solution was added immediately to the sample to be reduced in aqueous methanol containing a drop of alkali.Reduction was complete after the solution has been warmed gently for a few moments; occasionally, stronger heating to the boil with swirling may be required and the process will be complete when the deep yellow colour has been discharged. This was always a definite change. This solution was cooled and diluted with water before addition of concentrated sulphuric acid and catalyst for a Kjeldahl reaction, or with methanol for the determination of boron. If, in the preparation of the alkaline stannite, a darkening and precipitation of fine black particles occurred, indicating disproportionation to metallic tin and stannate, it was found better to prepare a fresh sample. This disproportionating, which is accelerated by warming, may be avoided by slower addition of the alkali.The concentrated acid was used, but more dilute acid might suffice. 1 . 2. 3. 4. 5. 6. 8. !#. 10. 1 1 . 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. r 1 . REFERENCES Marzadro, M., Mikrochenz. Mikrocliinz. Acta, 1953, 40, 359. Eckert, X., Monatsh., 1913, 34, 1957. \Veizmann, M., Yofe, J., and Kirzon, B., 2. physiol. Chenz., 1930, 192, 70. l,Vuntler, M., and Idascar, X., An9z. Chivz. Anal,, 1914, 19, 329. Sachs, A. P., .I. Soc. Chem. Ind., 1917, 36, 015. Kruger, hl., Chem.-Ztg., 1894, 27, 833. Kiirschner, I<., Z . anal. Cliem., 1926, 68, 209. >.la, T. S., I,ang, K. I<., and McKinley, J . D., jun., Mihvochiin. A d a , 1957, 368. nickinson. IfT. E., Anal. Chenz., 1968, 30, 992.Steyermark, X., McGce, 13. I < . , Bass, 13. .\., and Iiaup, 11. I<., Ibid., 1958, 30, 1561. Simek, H . (;., Chenz. Listy, 1931, 25, 322. Schaefer, \V. E., and Recker, \Y. \V., ,4nal. Clieni., 1947, 19, 307. Arnd, Th., Angew. Ckenz., 1932, 45, 22. Hartc, I<. A., I n d . Eng. Chem., Anal. led., 193A, 7, 432. El&, A . , and Sobotka, H., J . Anzev. Client. SOC., 1926, 48, 501. Takcda, A , , and Senda, J., H e v . Olzava Inst. I-andwitsch. Biol., Okuycima CTniu., 1956, 10, 241. Xbrahamczik, E., Monatsh., 1947, 77, 376. Bclchcr, l<., and Bhatty, 31. I<., Analyst, 1958, 81, 124. Somcrs, 1’. D., jun., Pvoc. Indiana .4cad. Sci., 1945, 54, 117. Ellis, C. M., and Vogcl, A . I., Analyst, 1956, 81, 093. Jullig, T., and Barbicre, J . , Mew?. Yozldves, 1937, 27, 127.I-Iccrtjcs, P. &I., Chenz. Weekblnd, 1937, 34, 827. Cope, I\-. C., .I. Ind. Eng. Chein., 1916, 8, 592. Margoschcs, H. Jf., Kristen, lV., and Scheinost, lc., 13ev., 1923, 56B, 1043. Margoschcs, n. >I., Scheinost, E., and IYoyner, V., /bid., 1925, 58B, 1850. Margoschcs, 13. &I., and Scheinost, E., Ibid., 1925, 58B, 1857. McCutchan, I)., and Roth, \V. lc., A n d . Clienz., 1!152, 24, 369. .Raker, P. K. W., Analyst, 1955, 80, 481. Bradstrcct, li. 13., Anal. CJwnz., 1954, 26, 235. liroutlct, J . , Bull. SOC. S c i . Hvetagne, 1948, 23, 129. I)orfmullcr, G., Z. I V i ~ l s . Zuckevind., 1930, 80, 407. Sisley, P., and David, >I., Bull. Soc. Cliini. Fmtzce, 1929, 45, 312. l;ish, V. B., Anal. Chew., 1952, 24, 760. Yerrot, li., and Barghan, .4., Proc. I X t h I . U.P.A.C. Congvess, 1947, 2, 247. Richmond, H. D., Analyst, 1908, 33, 179. Elmorc, J . \V., J . Ass. O f f . Agvic. Chein., 1945, 28, 363. Thuau, U. J., and de Korsak, 1.’. (Collegium, 1910, 364), C h e w Zenfv., 1910, IT, 1247. Gautier, J . A., and lienault, J., Ann. Chim. Anal., 1946, 28, 86. Crane, F. E., jun., and Fuoss, R. nl., A n d . Ckenz., 1954, 26, 1651. Fleury, P., and I,evalticr, H., J . Plzav,n. Chiwz., 1924, 29, 137. Kuznetsov. V. I . . Zavod. Lab.. 1940. 9. 1039. , .. 42. Thomas, L. H., .J. Chenz. Soc:, 194&, 820. Received Novenzber 301h, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800466

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

468 STEWART : QUANTITATIVE DETERMISATIOX [Analyst, 1701. 8s The Quantitative Determination of Non-ionic Detergents of the Type Formed by the Condensation of Nonyl Phenol with Ethylene Oxide BY R. G. STEIYART ( X e w Zealand Wool Industries Reseavch Iizstitzite, 444 George Stveef, Dztizedijz, X e w Zealand) The method described involves the measurement of the optical density, a t 490 mp, of dilute solutions of dichlorofluorescein in acetic acid; small amounts o f non-ionic detergent of the type formed when iionyl phenol is condensed with ethylene oxide appreciably reduce the optical density, and this effect can be used in determining the amount of detergent in an unknown solution. Anionic and other non-ionic detergents in similar amounts have ;L much smaller effect on the optical density.A calibration curve must first be prepared from known amounts of the cletergent. The effect of the detergent on optical density is dependent on pH, and the sensitivity of the method can be altered by working a t different pH values, with acetic acid - sodium acetate buffer solutions. NON-IOXIC detergents of the Lissapol N tj'pe are widely used in Xew Zealand for scouring raw wool and in the later stages of wool processing. Methods are available for the quantitative determination of these materials, usually by gravimetric analysis,l 92 but they are not recommended for amounts of less than 10 mg. Other more sensitive absorptiometric methods are available in which the colour of the complex formed by the detergent with molybdo- phosphoric acid3 or 2,4-dinitrophenylhydrazirie4 is measured to determine the amount of I o L 2 - - 0 I 3 I 4 Detergent present, mg Fig.1 . Optical density of a 12.5 PM solution of dichlorofluorescein containing detergent at : curve A, pH 4-65; curve H, pH 4.0; curve C, pH 2-9; curve L), pM 2.6 detergent in a mixture. Schonfeldt5 also describes a method dependent on the addition product formed when the ethylene oxide chain of the detergent molecule reacts with ferro- cyanic acid, the excess of acid being determined bj- titration with zinc sulphate. The method described in this paper, though sensitive to one tj'pe of non-ionic detergent only, has the advantages of simplicity and speed.June, 19631 OF NON-IONIC DETERGENTS 469 Lovelock and Xash6 reported that traces of some substances, among them some anionic detergents, appreciably reduced the optical density of a dilute solution of dichlorofluorescein in acetic acid.They attributed this to micelle formation and suggested that the reaction might be applied to the quantitative determination of non-ionic detergents. This was investigated in these laboratories and found to be a quick and accurate method of quanti- tatively determining detergents of the type formed when nonyl phenol is condensed with ethylene oxide. EXPERIMENTAL An approximately 0.005 M solution of dichlorofluorescein in glacial acetic acid was prepared, and 2-5 ml were diluted to 500 ml in a calibrated flask. For an investigation of the effect of pH, known amounts of acetic acid or sodium acetate were added to this solution before it was made up to the mark.By pipette, 25 ml of the diluted dichlorofluorescein solution were transferred to a 50-ml calibrated flask containing a known amount of detergent, and the solution was diluted to the mark with distilled water. The optical density'of the solution was measured in 1-cm Corex cells with a Beckman DU spectrophotometer at 490mp against distilled water. By using, initially, the non-ionic detergent Lissapol N450 curves of optical density against con- centrations of detergent were prepared (see Fig. 1). The pH shown is that of the dichloro- fluorescein - detergent solution whose optical density was being measured. From an examination of the results obtained it was concluded that the measurement of the optical density at pH 4 would provide a satisfactory method for the quantitative determination of the detergent.METHOD The method can be used to determine detergent in the range 0.2 to 3.0 mg. Dichlorojuorescein. Acetic acid, glacial. Sodium acetate, crystalline. PROCEDURE- Weigh out 100mg of dichlorofluorescein, and dissolve in 30 ml of glacial acetic acid, warming if necessary. Cool, transfer to a 50-ml calibrated flask, and make up to the mark with glacial acetic acid. Transfer 2.5 ml of this solution by pipette to a 500-ml calibrated flask containing 1.532 g of sodium acetate dissolved in approximately 450 ml of distilled water, and make up to the mark. By pipette, transfer 25 ml of this diluted dichlorofluorescein solution to separate 60-ml calibrated flasks containing known amounts (0.2 to 3 mg) of detergent, dilute each solution to the mark with distilled water, and measure the optical densities at 490 mp against distilled water.Prepare a calibration graph of optical density against amount of detergent present. Take a suitable sample of a solution of unknown detergent concentration, add 25 ml of dichlorofluorescein solution, dilute to 50 ml with distilled water, and measure the optical density. By reference to the calibration graph, determine the amount of detergent in the sample solution. A more sensitive variation of the proposed method is made possible (range 0.1 to 1.0 mg) by omitting the sodium acetate. However, the disadvantages are that the diluted dichloro- fluorescein solution is not particularly stable and must be used within 2 hours, and that the spectrophotometer cell must be well rinsed with dilute alkali and distilled water after each optical-density measurement.DISCUSSION OF RESULTS REAGENTS- This solution is stable for at least 24 hours. An examination of Fig. 1 shows that, at pH 2.6, the curve obtained is of no value for the quantitative determination of detergent. At pH 2-9 a large change in optical density was associated with a small change in detergent concentration. The curve shows a shoulder in the region of 1 mg of detergent, but it was found possible to use this curve for determining amounts of less than this. Recovery of detergent was reasonably good (see Table I), but, at this pH, contamination of the cell surface owing to adsorption of dichlorofluorescein occurred fairly rapidly, and it was necessary470 STEWART [Analyst, VOl.88 to clean the cell with dilute alkali after each determination. Also, the dilute solution of dichlorofluorescein deteriorated after a short length of time and had to be used within 2 hours of preparation. TABLE I RECOVERY OF LISSAPOL N450 AT pH 2.9 Lissapol N450 present, mg . . . . 0.30 0.50 0.70 0.90 1,issapol N450 found, mg . . . . 0.28 0.48 0.71 0.92 Recovery, yo . . . . . . . . 93 96 101 10% The optical density at higher pH values was not so sensitive to the presence of detergent, but the usable range was greater. Also, dichlorofluorescein was not adsorbed on the cell walls to the same extent, and the diluted solutjon could be kept for at least 24 hours without deterioration. Thus measurement of optical density at pH 4 forms the basis of the proposed method. Table I1 shows the recovery obtained with three commercially available detergents of the same type.TABLE I1 RECOVERY OF DETERGENT AT pH 4 Lissapol N450 Renex 688 Triton NlOO 7h----7 - 7- Detergent present, mg . . 1.20 2.40 1.20 2.40 1.20 2.40 Detergent found, mg . . 1.24 2.38 1-22 2-37 1.21 2.38 (I.C.I.) (Atlas) (Rohm and Haas) Recovery, % .. . . 103 98 102 99 101 99 The scope of the method is unfortunately narrow, since it was found that only the non-ionic detergents formed by the condensation of nonyl phenol with ethylene oxide could be determined. Other non-ionic detergents, eg., condensation products of tridecyl alcohol with ethylene oxide and detergents of the alkylolamide type, had only a relatively small effect on the optical density of dichlorofluorescein solutions.A few anionic detergents were also examined, among them primary and secondary alkyl sulphates and salts of fatty acids. These too had only a small effect on the optical density, reducing it by approximately 5 per cent. of the reduction expected when the same amount of non-ionic detergent of the suitable type was present. Because of this small effect exerted by other detergents the method is not applicable to mixtures containing appreciable unknown amounts of other non-ionic or anionic detergents. Cationic detergents brought about a change in colour to rose pink, owing to salt formation, and would also interfere in any determination. The presence of some water-miscible solvents (those investigated were ethanol, isopropanol and acetone) in comparatively large amounts also reduced the optical density; 5 ml of each solvent were equivalent to approximately 0.7, 0.9 and 1.2 mg of the non-ionic detergent, respectively. However, such interference could, presumably, be easily avoided by evaporating the solvent from the sample of detergent before it was diluted.The reduction in optical density when non-ionic detergents of a suitable type were present appeared to depend on the hydrophobic portion of the detergent molecule. This was indicated when it was found that substituting tridecyl alcohol for nonyl phenol greatly reduced the effect. A series of non-ionic detergents was examined of which the hydrophobic portion of the molecule in each instance was nonyl phenol, but the hydrophilic portion consisted of ethylene oxide chains of different lengths (Renex detergents obtainable from Atlas Powder Co.). When equivalent amounts of these detergents were used, the reduction in optical density was similar for each detergent . REFEKESCES 1. Joint Committee of the Association of British Chemical Manufacturers and the Society for Analytical Chemistry, “Recommended Methods for the Analysis of Trade Effluents,” W. Heffer & Sons Ltd., Cambridge, 1958, p. 91; Analyst, 1957, 82, 826. 2. Smith, W. B., Analyst, 1959, 84, 77. 3. Stevenson, D. G., Ibid., 1954, 79, 504. 4. Gatewood, L., jun., and Graham, H. D., Anal. Chenz., 1961, 33, 1393. 5. Schonfeldt, N., J , Amer. Oil Chem. SOC., 1955, 32, 77. 6. Lovelock, J . E., and Kash, l., Natwe, 1958, 181, 1263. First received December l l t h , 1961 Amended, October 29th, 1962

ISSN:0003-2654    

DOI:10.1039/AN9638800468

出版商:RSC    

年代:1963

数据来源: RSC