ISSN:0003-2654    

DOI:10.1039/AN97196FX033

出版商:RSC    

年代:1971

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97196BX035

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

iV THE ANALYST [September, 1971THE ANALYSTEDITORIAL ADVISORY BOARDChairman: H. J. Cluley (Wembley)*T. Allen (Bradford)*L. S. Bark (Salford)M. T. Kelley (U.S.A.)W. Kemula (Poland)*G. F. Kirkbright (London)G. W. C. Milner (Harwell)G. H. Morrison (U.S.A.)*G. Nickless (Bristol)*J. M. Ottaway (Glasgow)*G. E. Penketh (Billingham)R. Belcher (Birmingham)L. J. Bellamy, C.B.E. (Waltham Abbey)L. S. Birks (U.S.A.)E. Bishop (Exeter)*R. C. Chirnside (Wembley)A. C. Docherty (Billingham)D. Dyrssen (Sweden)*W. T. Elwell (Birmingham)*D. C. Garratt (London)*R. Goulden (Sittingbourne)S. A. Price (Tadworth)D. 1. Rces (London)E. B. Sandell (U.S.A.)A. A. Smales, O.B.E. (Harwell)H. E. Stagg (Manchester)E. Stahl (Germuny)A. Walsh (Australia)T. S.West (London)P. Zurnan (U.S.A.)J. Hoste (Belgium)D. N. Hume (U.S.A.)*J. A. Hunter (Edinburgh)H. M. N. H. Irving (Leeds)A. G. Jones (Welwyn Garden City)*Members of the Board serving on the Executive Committee.NOTICE TO SWBSCRIBERSSubscriptions for The Analyst, Analytical Abstracts and Proceedings should beThe Chemical Society, Publications Sales Ofice,Blackhorse Road, Letchworth, Herts.Rates for 1972(a) The Analyst, Analytical Abstracts, and Proceedings, with indexes . . . . €33.50 $80.40(b) The Analyst, Analytical Abstracts printed on one side of the paper (withoutindex), and Proceedings . . .. .. . . . . . . . . €34.50 $82.80(c) The Analyst, Analytical Abstracts printed on one side of the paper (withindex), and Proceedings .. .. .. .. .. .. . . f40.50 $97.20(Other than members of the Society)sent through a subscription agent or direct to:The Analyst and Analytical Abstracts without Proceedings-(d) The Analyst and Analytical Abstracts, with indexes . . .. .. . . €31.00 $74.40(e) The Analyst and Analytical Abstracts printed on one side of the paper (withoutindex) . . . . . . . . .. .. .. .. .. . . €32.00 $76.80(f) The Ana!yst and Analytical Abstracts printed on one side of the paper (withindex) . . . . . . . . .. .. .. .. .. . , €38.00 $91.20(Subscriptions are NOT accepted for The Analyst and/or for Proceedings alone)Members should send their subscriptions to the Hon. Treasurevi SUMMARIES OF PAPERS I N THIS ISSUE [September, 197 1Summaries of Papers in this IssueThe Application of Non-flame Atom Cells in Atomic-absorptionand Atomic-fluorescence SpectroscopyA ReviewSUMMARY OF CONTENTSTntroductimGeneral considerationsFurnacesFilamentsCathode sputtering cellsOther non-flame cellsConclusionREPRINTS of this Review paper will soon be available from the Societyfor Analytical Chemistry, Book Department, 9/10 Savile Row, London,IVlX IAF, a t 25p per copy, post free.A remittance for the correct amount, made out to The Society forAnalytical Chemistry, MUST accompany every order ; these reprints are notavailable through Tradc Agents.G.F. KIRKBRIGHTChemistry Department, Imperial College, London, S.W.7.Analyst, 1971, 96, 609-623.SLEPT: A Simple Computer Language for Examining DataRecorded on Punched Paper TapeA versatile language has been developed for the computer proccssingof paper tape output from multi-channel analysers.It is intended for usein laboratories in which the work load is too varied to justify writing aprogram, or suite of programs, dedicated to a specific set of operationr.The language consists of a set of commands and associated information, whichis read from carcis and executed by a program written in FORTRAN IV. Theprogram has a modular structure so that new commands can easily be incor-porated into the language as required.C. R. BOSWELLAnalytical Research and Development Unit, Atomic Energy Research Establishment,Harwell, Didcot, Berks.Analyst, 1971, 96, 024-630.The Atomic-emission Spectroscopy of Rhenium in the NitrousOxide - Acetylene FlameKhenium can be determined by atomic-emission spectroscopy by useof a pre-mixed nitrous oxide - acetylene flame supported on a 6-cm slotburner.Thc limits of detection were 0.7 pg ml-l (346.1 nm) and 1-5 pg ml-l(488.9 nm) ; analytical working curves were linear for rhenium concentrationsbelow 200 p g ml-l a t both wavelengths. Spectral interference from palla-dium, nickel, rhodium, cobalt and large amounts of lanthanum or phosphateoccurred at 346.1 nm and from large amounts of aluminium a t 488.9 nm.A number of elements gave rise to chemical interference, but this was elimi-nated by the addition of sulphuric or phosphoric acid. A monochromatorgiving a spectral band pass of the order of 0.1 nm must be used for thedetermination, and background corrections must be made either by wave-length scanning or by measurement a t the peak and at an adjacent wavelength.R.SMITH and A. E. LAWSONImperial Chemical Industries Ltd., Petrochemicals Division, Billingham, Teesside.Analyst, 1971, 96, 631-639viii SG'llillAlIlES OF PAPERS I N THIS ISSUEA Method for the Chemical Analysis of Magnesites and Dolomites[September , 107 1This paper includes a detailed description of a method for the analysisof magnesites and dolomites; the method has been accepted (in principle)by the British Standards Institution as a standard method. Determinationsinclude SiO, (gravimetric) , TiO,, Fe,O,, Cr,O,, &In0 (colorimetric) and A1,0,,CaO and MgO (complexometric) .Notes on the development of the method andtables of co-operative results obtained by the Refractories Working Groupof the Analysis Committee are included.H. BENNETT and R. A. REEDThe British Ceramic Research Association, Queens Road, Penkhull, Stoke-on-Trent,Agzalyst, 1971, 96, 640-656.ST4 7LQ.4- [Bis(carboxymethyl)aminomethyl] - 3- hydroxy-2-naphthoic Acidas a Fluorescent Indicator for the ComplexometricTitration of Calcium plus Magnesium4-[Bis(carboxymethyl)aminomethyl]-3-hydroxy-2-naphthoic acid, knownalso as l-dicarboxymethylaminomethyl-2-hydroxy-3-naphthoic acid, a spectro-fluorimetric reagent for beryllium, has been found to be an effective fluori-metric indicator for the complexometric titration of calcium plus magnesiumwhen used in conjunction with a suitable fluorimetric titrimeter.A procedureis described in which the indicator is used in the titrimetric determinationof magnesium in silicate rocks.R. L. CLEMENTS, J. I. READ and G. A. SERGEANTDepartment of Trade and Industry, Laboratory of the Government Chemist,Cornwall House, Stamford Street, London, S.E. 1.Analyst, 1971, 96, 656-658.Use of the Halphen Reaction for the Determination of theCyclopropenoid Content of LipidsAn application to cottonseed oils of a quantitative version of the Halphentest for the determination of cyclopropenoid material has been published byother workers, but for other oils containing higher levels of cyclopropenoids,although the absorption at the 495nm peak is linearly related to the con-centration of each oil examined, the relationship differs among the oils.How-ever, transmethylation of oil before applying the Halphen reaction has beenfound to give results that are in better agreement with titration with hydrogenbromide for oils with widely differing cyclopropenoid content. The use ofpressurised capsules for carrying out the reaction with reduced loss of solventhas proved advantageous, as flatter peaks are obtained when optical absorp-tion is plotted against time. The application of the modified technique to oilscontaining a wide range of concentrations of total cyclopropenoid material inthe component fatty acids is described and discussed.T. W. HAMMONDS, J. A. CORNELIUS and L. TANForeign and Commonwealth Office (Overseas Development Administration), TropicalProducts Institute, 56/62 Gray's Inn Road, London, W.C.1.Analyst, 1971, 96, 659-664.Determination of Clamidoxic Acid in Serum byGas - Liquid ChromatographyA method has been developed for the determination of clamidoxic acid[2- (3,4-dichlorobenzamido)phenoxyacetic acid] in serum by using gas chro-matography with electron-capture detection. Clamidoxic acid is extractedfrom acidified serum into toluene and returned to an aqueous sodium hydroxidephase. The amide is hydrolysed to 3,4-dichlorobenzoic acid which, afteracidification, is extracted into toluene containing an internal standard. Theacid is converted into its methyl ester by the addition of a solution of diazo-methane in diethyl ether, excess of diazomethane is destroyed with aceticacid, and an aliquot analysed.L. SHERMAN and (the late) G. A. TAYLORSmith & Nephew Research Limited, Gilston Park, Harlow, Essex.Analyst, 1971, 96, 665-670

ISSN:0003-2654    

DOI:10.1039/AN97196FP133

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

September, 19711 THE ANALYST xiCLASSIFIED ADVERTISEMENTSThe rate for classified advertisements is 35Pa line (or sficeequivalent of a line) with an extra charge of l o p f o r theuse of a Box Number. Semi-displayed classijedadvertisements are f;4 for single-column inch.Copy required not later than the 8th of the month pre-ceding date of publication which is on the 16th of eachmonth. Advertisements should be addressed toJ . Arthur Cook, 9 Lloyd Square, London, W.C.I.Tel.: 01-837 6315FOR SALEThe Analyst-9 volumes 1962-1970. Analytical Abstracts-11 vol-umcs 1962-1970. Indexes--13 volumes 1960-1950. Offers please t oBox No. 223, c/o J. Arthur Cook, 9 Lloyd Square, London, W.C.l.APPOINTMENTS VACANTANALYTICAL CHEMISTWith initiative and drive required in Food andDrugs Laboratory.Most of the work of thelaboratory is of an official or legal nature andpreference will be given to a man or woman hold-ing the M.Chem.A. qualification who can be ap-pointed as a deputy Public Analyst. The personappointed should be able to control staff andaccept responsibility. Commencing salary up toL3,OOO per annum. Write giving full particularsto Thomas McLachlan & Partners, 4 HanwayPlace, London WlP 9DF.BOOKSMONOG RAPHSREPRINTSorders for all publications ofthe Society (except journals)should be sent direct or througha bookseller to-THE SOCIETY FORANALYTICAL CHEMISTRYBook Department9/10 Savile Row,London, WIX IAFWANTEDSpectrop hotometersFlame photometersAbsorptiometerspH meters, balancesVacuum equipment, recordersall general laboratory instrumentsany age or condition.V.N. BARRETT, & Co. Ltd.,1 Mayo Road,CROYDON CRO 2QPSurrey.01-684 9917BUREAU OF ANALYSEDSAMPLES LTD.are pleased to announce their appointmentas sole agents in the U.K. and Europe for:-STEEL HYDROGENANALYSIS STANDARDSprepared by The Welding Institute. Sixdifferent hydrogen levels, 0.05 to 1.10 mlare available and each sample is suppliedwith a Certificate of Preparation statingthe precise hydrogen content.For further details please write to us at :-NEWHAM HALL, NEWBY,M I DDLES B ROU G H , TE ESSl D E, EN G LAN D.TS8 9EAor Te1ephone Middlesbroogh 37216 (STD 0642xiv SUMMARIES OF PAPERS IN THIS ISSUEResidues of Prophylactics in Animal ProductsPart I.The Determination of Sulphaquinoxaline in Eggs andPoultry by Gas - Liquid ChromatographySeptember, 19711A procedure for the detection and determination of residues of sulpha-quinoxaline in eggs and poultry is described. Sulphaquinoxaline is extractedfrom the sample with acetonitrjle and, after a partition clean-up process,is hydrolysed to 2-aminoquinoxaline. The trifluoroacetyl derivative of thisamine is a suitable compound for determination by gas - liquid chromato-graphy with electron-capture detection. The method is applicable to residuesat concentrations in the range 0.1 to 5 mg kg-1.S . CRISPDepartment of Trade and Industry, Laboratory of the Government Chemist,Cornwall House, Stamford Street, London, S.E.1.Analyst, 1971, 96, 671-674.Colorimetric Method for the Determination of Iron inPyrethrum ExtractsIt has been shown that pyrethrum extract is contaminated with ironduring processing and that it is necessary to destroy the organic matter beforedetermining the metal. Wet oxidation has been carried out with a mixtureof sulphuric and nitric acids, but this treatment may lead to the precipitationof an iron complex. It was found that addition of potassium sulphate priorto oxidation prevented the precipitation of this complex. The iron can bedetermined spectrophotometrically at 480 nm by use of its thiocyanate com-plex. Nitrous acid residues from the oxidation must be removed by boilingwith water and nitric acid added to oxidise iron(I1) ions.The resultantmethod for determining microgram amounts of iron is both rapid and precisewith small weights of extract, but when weights greater than 0.1 g aredigested loss of sulphuric acid by evaporation causes small errors.R. A. G. MARSHALLDepartment of Chemistry, Thames Polytechnic, London, S.E. 18.Analyst, 1971, 96, 675-673xvi THE ANALYST September, 19711Reprints of Review PapersREPRINTS of the following Review Papers published in The Analyst since January, 1963, areavailable from The Society for Analytical Chemistry, Book Department, 9/10 Savile Row, London,WlX 1AF (not through Trade Agents). Orders MUST be accompanied by a remittance for thecorrect amount made out to “Society for Analytical Chemistry.”“Classification of Methods for Determining Particle Size,” by the Particle Size Analysis“Methods of Separation of Long-chain Unsaturated Fatty Acids,” by A.T. James (August,“Beer’s Lam7 and its Use in Analysis,” by G. F. Lothian (September, 1963).“A Review of the Methods Available for the Detection and Determination of Small Amounts“Circular Dichroism,” by R. D. Gillard (November, 1963).“Information Retrieval in the Analytical Laboratory,” by D. R. Curry (Sovember, 1963).“Thermogravimetric Analysis,” by A. W. Coats and J . P. Redfern (December, 1963). Price“Some Analytical Problems Involved in Determining the Structure of Proteins and Peptides,”“The Faraday Effect, Magnetic Rotatory Dispersion and Magnetic Circular Dichroisni,” by“Electrophoresis in Stabilizing Media,” by D.Gross (July, 1965).“Recent Developments in the Measurement of Nucleic Acids in Biological Materials,” by“Radioisotope X-ray Spectrometry,” by J. R. Rhodes (November, 1966).“The Determination of Iron(11) Oxide in Silicate and Refractory Materials,” by H. N. S.“Activation Analysis,” by R. F. Coleman and T. B. Pierce (January, 1967).“Techniques in Gas Chromatography. Choice of Solid Supports,” by F. J . Palframan“Heterocyclic Azo Dyestuffs in Analytical Chemistry,” by R. G. Anderson and G. Nickless“Determination of Residues of Organophosphorus Pesticides in Food,” by D. C. Abbott and“Radioactive Tracer Methods in Inorganic Trace Analysis: Recent Advances,” by J. W.“Gamma-activation Analysis,” by C. A. Baker (October, 1967).“Precipitation from Homogeneous Solution,” by P.F. S. Cartwright, E. J . Newman andD. W. Wilson (November, 1967).“Industrial Gas Analysis,” by (the late) H. N. Wilson and G. M. S. Duff (December, 1967).Price 35p.“The Application of Atomic-absorption Spectrophotometry to the Analysis of Iron andSteel,” by P. H. Scholes (April, 1968).“Inorganic Ion Exchange in Organic and Aqueous - Organic Solvents,” by G. J. Moody andJ. D. R. Thomas (September, 1968).“Radiometric Methods for the Determination of Fluorine,” by J. K. Foreman (June, 1969).Price 25p.“Techniques in Gas Chromatography. Developments in the van Deemter RateTheory of Column Performance,” by E. A. Walker and J . F. Palfranian (August, 1969).Price 25p.Choicz of Detectors,” by T. A. Gough andE.A. Walker (January, 1970).Sub-committee of the Analytical Methods Committee (March, 1963).1963). Price 25p.Price 25p.Price 25p.of Cyanide,” by L. S. Bark and H. G. Higson (October, 1963). Price 25p.Price 15p.Price 15p.25p.by Derek G. Smyth and D. F. Elliott (February, 1964).J . G. Dawber (December, 1964).Price 25p.Price 25p.Price 25p.H. N. Munro and A. Fleck (February, 1966). Price 25p.Price 25p.Schafer (December, 1966). Price 25p.Price 25p.Part I.and E. A. Walker (February, 1967).(April, 1967). Price 25p.H. Egan (August, 1967).McMillan (September, 1967). Price 25p.Price 25p.Price 25p.Price 25p.Price 25p.Price 25p.Price 35p.Part 11.“Techniques in Gas Chromatography.“Laser Ranian Spectroscopy,” by P. J. Hendra and C. J. Vear (April, 1970).“Ion-selective Membrane Electrodes,” by Ern0 Pungor and KlAra T6th (July, 1970). Price“X-ray Fluorescence Analysis,” by K. G. Carr-Brion and K. W. Payne (December, 1970).“Mass Spectrometry for the Analysis of Organic Compounds,” by A. E. Williams and H. E.Part 111.Price 25p.Price 35p.35p.Price 25p.Stagg (January, 1971). Price 35p

ISSN:0003-2654    

DOI:10.1039/AN97196BP141

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

SEPTEMBER. 1971 Vol. 96, No. 1146 THE ANALYST The Application of Non-flame Atom Cells in Atomic-absorption and Atomic-fluorescence Spectroscopy A Review* BY G. F. KIRKBRIGHT (Chemistry Department, Imperial College, London, S . W.7) SUMMARY OF CONTENTS Introduction General considerations Furnaces Filaments Cathode sputtering cells Other non-flame cells Conclusion IN the past decade, atomic-absorption spectroscopy (AAS) has been widely demonstrated to provide a sensitive and selective technique for inorganic trace analysis. Atomic-absorption spectrophotometers are now available from more than twenty instrument manufacturers throughout the world. Atomic-fluorescence spectroscopy (AFS) has been shown by several groups of workers to be a complementary technique that may permit higher analytical sensitivity than AAS for the determination of some elements. In both techniques the attainable sensitivity and precision are limited by the characteristics of the primary radiation source, and by the technique with which atoms of the analyte element are formed from the sample to be examined. Although considerable effort has been devoted to the develop- ment of improved hollow-cathode lamps and electrodeless discharge-tube sources, and the selectively modulated and pulsed operation of sources, the most long-standing and widely used method of atomisation has been that with the flame.All commercial instrumentation available is equipped with flame atomiser facilities, and this situation will undoubtedly continue for some time. The fact that flames were inherited as AAS atom cells from the older technique of flame-emission spectroscopy may account in part for their popularity, although they also have the following advantages for use in the analytical techniques of AAS and AFS.(i) They are convenient to use,reliable and relatively free from a tendency to memory effects. Most flames in common use can be made noiseless and safe to operate. (ii) Most burner systems are small, durable and inexpensive. Sample solutions are easily and rapidly handled by the use of relatively simple nebuliser assemblies. (iii) A wide variety of flames is available to allow the selection of optimum conditions for many different analytical purposes. (iv) The signal-to-background and signal-to-noise ratios obtainable are sufficiently high to allow adequate sensitivity and precision to be obtained in a wide range of analyses at different wavelengths between 200 and 800nm.Flame-atomisation systems exhibit some disadvantages, however, which limit the attain- able sensitivity and convenience in their use for analysis. These disadvantages have led many workers to devise techniques for the atomisation of samples for analysis that are not based on chemical flames. This paper reviews some of the non-flame cell devices that have been described for use in AAS and AFS. * Reprints of this paper will be available shortly. For details see Summaries in advertisement pages. 0 SAC and the author. 609610 KIRKBRIGHT: APPLICATION OF NON-FLAME ATOM CELLS IN [Analyst, VOL 96 GENERAL CONSIDERATIONS Some of the possible disadvantages of flames for analytical work are given below.(a) The volume of the sample solutions available may frequently be less than that required for use with an indirect nebuliser system. For low analyte concentrations it may not be possible to dilute the solution to overcome this limitation. With indirect nebulisers used with pre-mixed flames only a small fraction of the sample solution enters the flame. (b) Flame cells are only rarely able to atomise solid samplesldirectly. (c) Flame background absorption and emission at the wavelength of the resonance line of the analyte element, or thermal emission from the analyte or concomitant elements at this wavelength, may give rise to unacceptable signal noise with consequent loss of precision. (d) In some locations it may be inconvenient to use high-pressure cylinders of support and fuel gas, and in closed automated systems where no operator is in attendance it may not be desirable to use a flame as the atom cell.In addition to these practical disadvantages, other more fundamental factors act in flames to limit the sensitivity and selectivity that may be achieved. These are as follows. (1) The attainable atom concentration in flames is limited by the dilution effect of the relatively high flow-rate of unburnt gas used to support the flame and to transport small volumes of sample solution to the flame. The atom concentration is also limited by the flame gas expansion that occurs on combustion. (2) Precise control over the chemical environment of the analyte and concomitant atoms in flame cells is not possible, The degree of control of chemical composition that can be obtained by variation of the fuel-to-oxidant concentration ratio is accom- panied by simultaneous changes in the flame temperature and its spectral absorption and emission characteristics.For many elements, particularly those that form thermally stable oxides, the efficiency of free atom production from the sample introduced into the flame is low. (3) In the technique of AFS, the flame that produces the greatest freedom from inter- element effects may also produce low fluorescence efficiency through quenching of radiationally excited analyte atoms by flame gas molecules. For the analysis of small liquid or solid samples, and for the determination of trace amounts of many elements in larger samples, it would be advantageous to achieve a higher concentration of atoms in a small cell volume than is possible with flames with which solution- nebulisation techniques are used ; this can be accomplished in non-flame cells.Winefordner,l for example, in a study of a typical graphite cell technique, has demonstrated that a sub- stantially higher peak concentration of atoms may be expected in such a cell than in a flame. This gain results directly from avoidance of the limiting sample dilution effects that occur in flame cells, To assist the formation and maintenance of a high free-atom fraction for the analyte element for AAS and AFS, it is also an advantage in non-flame cells that the chemical environment of the cell can be controlled by the use of an inert gas atmos- phere; also in AFS the selection of a suitable inert supporting gas may result in greater fluorescence efficiency by the reduced quenching of radiationally excited atoms.With suitably designed non-flame cells a considerable advantage over flames in the recorded noise caused by background absorption or emission or sample emission, which originates from the cell, would result. It is apparent, therefore, that non-flame devices may find widespread application in AAS and AFS, provided their simplicity and reliability are comparable with those of flame cells. FURNACES- The use of high-temperature furnaces offers the possibility of a high concentration of atoms within a well defined volume with very low background emission and noise for AAS and AFS.Several furnace devices that permit a long path length for AAS have been des- cribed. However, for those in which the operating temperature is below about 1500 "C,September, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 61 1 the range of elements that can be atomised is restricted, and chemical interferences may be encountered with complex sample matrices. These disadvantages may be outweighed by the potentially high sensitivity obtained for particular elements in this type of cell. Mislan2 has described the construction and performance of an atomic-absorption spectrophotometer that involves the use of a 36-cm silica tube of 2.5 cm i d . heated to a maximum temperature of 1250 “C by a wire-wound resistance furnace.Sample solutions were transferred to the absorption tube through a conventional indirect nebuliser - spray chamber assembly. The device was used for the determination of cadmium and excellent detection limits were obtained. Hudson3 reported AAS measurements on sodium vapour produced in a stainless-steel absorption cell heated by a resistance Wire. Atomic-absorption studies by using furnaces have also been made by Vidale,* Choong and Loong-Sen@ and by Tomkins and ErcoL6 U Fig. 1. Graphite tube as- sembly of L’vov furnace (from L‘vovle): 1, electrode with sample; 2, crucible; and 3, graphite contacts placed inside coolers Several satisfactory devices involving graphite furnaces of the type used by King7 have been constructed for analytical use in both AAS and AFS, the earliest and best known of which is the furnace device first described by L’vov.* During the past 10 years, L’vov and his co-workers have undertaken extensive AAS studies with various models of the original graphite f~rnace.~ to l7 This work has been reviewed by L’vov in his textbook16 and else- where.16 A diagram of the basic graphite tube assembly used in later models of the device is shown in Fig.1. Graphite cylinders 30 to 50 mm in length and of 2.5 to 3 mm id., depending on the analytical requirements, can be used. The furnace can be heated electrically by a.c. current from a 4-kW transformer at 10 V, and the temperature to which the tube is heated is regulated by changing the voltage supply to the primary circuit of the transformer. In early forms of the furnace the inside wall of the tube was lined with tantalum or tungsten foil to eliminate diffusion of vapour through the porous walls.More recently, preference has been given to tubes lined both inside and outside with a layer of pyrographite, which has low gas permeability, high heat conductivity and resistance to oxidation, so that it precludes diffusion of vapour through the walls and ensures uniform heating of the tube and a longer lifetime. The sample (solid or liquid) is introduced into the tube on an auxiliary carbon-rod electrode 6 mm in diameter, the tip of this rod being shaped to fit an orifice in the tube wall. In operation, the sample to be analysed can be placed on this electrode as a solution or powder. The electrode is positioned below the graphite tube, the chamber enclosing the tube assembly is closed and purged with argon, and the tube is heated (for 20 to 30s) to the required temperature. The auxiliary electrode is moved into place to fit the orifice in the tube wall and is heated electrically by an a.c.current from the secondary winding of a 1-kW step-down transformer (220/15V), from the electrode to the tube. The sample is vaporised and the atomic-absorption signal is recorded. In the original form of the L’vov furnace, the pulse vaporisation of the sample was achieved by using a d.c. arc between an electrode in the crucible and the auxiliary electrode mounted under the crucible. The graphite crucible pulse-vaporisation method has been used by L’vov with a two-channel atomic- absorption spectrophotometer that permits simultaneous recording of absorption for two elements, one of which can be an internal standard.A continuum source can be used to correct for any interference caused by non-selective molecular absorption and scattering of source radiation. The high-intensity atomic-line sources used may be either high-frequency electrodeless discharge tubes or hollow-cathode lamps operated from a pulsed power supply,612 KIRKBRIGHT: APPLICATION OF NON-FLAME ATOM CELLS IN [ATWbSt, VOl. 96 The radiation emitted by the two sources is combined by means of a semi-reflecting mirror. This beam is then modulated, merged with the continuum source radiation by using a rotating mirror chopper and passed through the graphite crucible atom cell. Table I shows some results for sensitivity obtained by L'vov;l6 the absolute amounts of different elements actually measured are given, together with the corresponding experi- mental conditions used and the extrapolation of the sensitivity results to the absolute amount of each element that would produce 1 per cent.absorption for a 2.5-mm graphite tube. The attainable precision claimed for the graphite tube technique operated with small liquid samples (2 to 5 1.1) corresponds to a relative r.m.s. error in a single measurement of 5 to 8 per cent. Element linelnm Ag 328.1 A1 309-3 AU 242.8 B 249.8 Ba 553.5 Be 234.9 Bi 306.8 Ca 422.7 Cd 228.8 CO 240.7 Cr 357.9 CS 852.1 CU 324.7 Fe 248.3 Ga 287.4 Hg 253.7 In 303-9 K 404.4 Li 670.8 Mg 285.2 Mn 279.5 MO 313.3 Ni 232.0 Pb 283.3 Pd 247.6 Pt 265-9 Rb 780.0 Rh 343.5 Sb 231.1 Se 196.1 Si 251.6 Sn 286.3 Sr 460.7 Te 214.3 Ti 365.3 Zn 213.8 T1 276.8 TABLE I SENSITIVITY DATA WITH L'vov FURNACE (FROM L'vovl6) Tube diameter1 mm 2.6 4.5 2-5 2.5 3.0 4.5 2.6 2.5 1.2 2.5 2.5 2.5 2.5 2.5 2.5 2-5 2.5 2.5 3.0 4.5 2-5 2.5 2-5 2.5 2-5 2.5 2.5 2-5 2.5 2.5 2-5 2.5 3-0 2.5 2.5 2.5 4.5 Argon pressure/ atm.2 1 2 2 1 6 2 2 1 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 4 Temperature1 "C 1800 2100 1700 2400 2200 2400 1800 2300 1500 2200 2200 1900 2100 2100 2100 700 1900 1800 1900 1800 2000 2500 2200 1900 2100 2300 1900 2300 2000 1600 2250 2000 2200 2000 2500 1800 1500 Measured amount/g 5 x 10-13 5 x 10-9 2.5 x 7 x 10-11 1 x 10-10 2.6 x 2.5 x 2.5 x 6 x 10-14 7.5 x 10-12 5 x 10-1' 6.6 X 10-l' 6.3 x 2.5 x 2.5 x 4.5 x 10-10 8 x 10-l2 6-3 X 10-l' 5 x 10-11 3 x 10-12 2.5 x 10-l2 5 x 10-11 2.5 x 10-1l 3 x 10-11 5 x 10-11 2.5 X 7-5 x 10-12 6.3 x 10-11 5 x 10-11 2 x 10-10 2.7 x 10-la 1 x 10-11 2 x 10-1' 7.6 x 10-l' 5 x 10-10 2.5 x 1 x 10-12 Sensitivity1 g per 1 per cent. absorption 1 x 10-12 1 x 10-12 2 x 10-10 6 x 10-l2 4 x 10-12 2 x 10-12 2 x 10-12 6 x 10-13 1 x 10-12 1 x 10-12 8 x 10-11 4 x 10-11 3 x 10-12 3 x 10-12 9 x 10-12 2 x 10-12 4 x 10-12 1 x 10-11 1 x 10-12 8 x 10-12 5 x 10-12 9 x 10-12 2 x 10-12 1 x 10-12 1 x 10-12 4 x 10-11 1 x 10-12 1 x 10-13 3 x 10-14 4 x 10-13 8 x 10-14 4 x 10-13 4 x 10-13 4 x 10-14 2 x 10-13 5 x 10-14 3 x 10-14 determination of In addition to the extensive use of the graphite tube furnace for the trace amounts of elements whose resonance lin& fie above 200 nm, L'vov,l6 in work by himself and Khartsyzov, has examined the possibility of its application to the determination of sulphur, phosphorus and iodine in the ultraviolet region between 170 and 190 nm.By using electrodeless discharge-lamp sources, the conventional argon-purged graphite furnace, lithium fluoride windows and lenses and a vacuum monochromator, sensitivities (for 1 per cent. absorption) of 3 .X 10-l2 g of phosphorus at 177.5 nm, 3 X g of iodine at 183-0 nm and 1 x 10-10 g of sulphur at 180.7 nm were attained with AAS. L'vov and co-workers13J5~17 have also made use of the favourable properties of the graphite furnace to study the Lorentz widths of resonance lines, the determination of absolute values of oscillator strengths and the estimation of atomic-diffusion coefficients.18September, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCE NCE SPECTROSCOPY 613 (id Fig.2. (i), Graphite cuvette assem- bly for AAS (from Massman'O); and (ii), for AFS (from Massmado). A, graphite tube or cuvette; and B, steel holders Massmanlg has described graphite cuvette devices for use in both AAS and AFS. The furnace arrangements used are shown in Fig. 2. The graphite tube used for AAS is 55 mm long, of 6.5 mm i.d. and 1.5-mm wall thickness. An orifice 2 mm in diameter is cut into the centre of the tube length so that the sample can be introduced with a micropipette. For AFS, a cup-shaped graphite cuvette, 40 mm long and of 6.5 mm i.d. and 16mm wall thick- ness, is used. The source radiation enters at the open top of the cuvette, and the fluorescence radiation is viewed through a slit cut into its wall.The sample is inserted into the fluorescence cuvette at its open top. Both the tube and cuvette can be heated to 2600 "C within a few seconds by using a power supply capable of supplying up to 400A. The absorption tube is purged with argon, but the optical path through the tube is open to the atmosphere. Emission from the fluorescence cuvette, which is similarly purged with argon, is viewed through a silica window in the side of the chamber. Modulated radiation from hollow-cathode lamp sources can be used for both techniques; so as to obtain high fluorescence intensities with these sources, however, they were operated for short periods of time at high current levels.Sample solution volumes of between 5 and 200 pl were used for AAS work and of 5 to 50 pl for AFS studies. Solid samples of up to 1 mg in weight were reported by Massman to be acceptable with his apparatus ; for larger samples, unacceptable background absorption effects were encountered. The same worker20 also corrected for non-selective background absorption in AAS by using a two-channel spectrometer and monitoring the absorption of a nearby elemental non-resonance or filler gas line from the source. Table 11 shows detection limits obtained by Massman for sixteen elements by atomic-absorption spectroscopy and nine ele- ments by atomic-fluorescence spectroscopy. Only for zinc and cadmium were clearly superior detection limits obtained with the latter technique, and for other elements the available source intensities were insufficient to permit low detection limits with it.The analytical precision achieved in AAS and AFS was similar when the dissolved solid content of sample solutions614 KIRKBRIGHT: APPLICATION OF NON-FLAME ATOM CELLS I N [Afia&jst, vol. 96 TABLE I1 DETECTION LIMITS BY AAS AND AFS WITH GRAPHITE CUVETTES (FROM MASSMANl') Element Zn Cd 2 2 Se Sb Fe Mn Pb Mg In Bi cu Na Linelnm 213.86 228.80 328-07 189.04 196.09 231.15 248.33 253.65 276.79 279.48 283.31 285.21 303.94 306.77 324-75 588.99 Detection limit by AASk 8 x 10-13 2 x 10-18 8 x 10-15 6 x 2 x 10-9 1 x 10-10 2 x 10-11 2 x 10-10 4 x 10-1' 8 x 10-12 1 x 10-11 2 x 10-10 2 x 10-10 1 x 10-11 7 x 10-12 5 x 10-13 Detection limit by AFSk 4 x 10-14 2.5 x 10-13 1-5 x - - 2 x 10-10 3 x 10-9 2 x 10-9 - - 3.5 x 10-11 3.5 x 10-12 4.5 x 10-10 - - - was 2 per cent.or less and not more than 100 and 30-pl sample volumes were taken, respec- tively. Standard deviations of between 4 and 12 per cent., depending on the matrix element and its concentration, were attained. Higher precision was achieved by use of an internal standard element whose absorption was measured simultaneously by using the two-channel spectrometer. A graphite furnace similar to that described by Massman has recently been introduced as an accessory to a commercially available atomic-absorption spectrophotometer,21 and its performance appears to be similar to that of the original device described by Massman. The determination of copper and strontium in milk, which is difficult by AAS with flames without sample pre-treatment, was chosen by Manning and Fernandez21 for preliminary studies to illustrate the application of the technique.The signal recorded during the analysis sequence in this determination is shown in Fig. 3. The atomic-absorption signal is recorded A J 3 Time - 60s 20s - L Fig. 3. Copper in milk with graphite furnace: A, dry- ing; B, charring; and C, atomic- absorption signal shown; 25-4 sample.September, 19713 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 61 5 TABLE I11 SENSITIVITY DATA FOR ELEMENTS OBTAINED WITH FURNACE OF WOODRIFF, STONE AND H E L D ~ ~ Element Ag Cd Zn Pb Li c u Mn Fe A1 Ni DY Ho Er Ca Mg Linelnm 328.1 228.8 213.9 283.3 670.8 324.7 279.5 285.2 248.3 309-2 232.0 421.2 410.4 400.8 422.7 Furnace temperature/"C 1265 1200 1200 1200 1200 1650 1200 1200 2200 1650 2200 2200 2200 2200 1200 Sensitivity / g per 1 per cent. absorption 8 x 9 x 10-12 9 x 10-12 1 x 10-11 1 x 10-11 9 x 10-1' 8 x 10-l1 8 x 10-l1 1 x 10-10 2 x 10-10 2 x 10-10 1 x 10-10 9 x 10-11 1 x 10-10 1 x 10-10 Detection limit/g* 1.2 x 10-11 1.3 x 10-l' 3.9 x 10-11 5.7 x 10-11 3 x 10-11 3.2 x 10-1O 2.7 x lo-'' 2.5 x 10-11 2.8 x lowlo 5.4 x 10-10 7.3 x 10-1' 2.2 x 10-10 3.3 x 10-10 3.7 x 10-11 5-3 x 10-10 * Amount producing absorbance value equal to twice experimental standard deviation of the blank.for a 25-pl sample containing 0.05 p.p.m. of copper, which corresponds to the normal con- centration in milk. This signal (C) is preceded by the non-specific absorption readings obtained at lower operating power during the evaporation of the water (A) and charring of the organic material (B) in the milk sample.I monc I 1/4-m Czerny-Turner xhromator 150-W focused arc L -. PMT H.V. supply Reversible counter 500 -HZ chopper Reference Control timer Arc welder Fig. 4. Schematic drawing of graphite cell system for atomie- Photon counting detection system used fluorescence spectroscopy. (from Winefordnerl) (PMT = photomultiplier)616 KIRKBRIGHT : APPLICATION OF NON-FLAME ATOM CELLS IN [Analyst, Vol. 96 Woodriff and c o - w o r k e r ~ ~ ~ 1 ~ ~ ~ ~ ~ developed a graphite tube furnace device for AAS in which a tube 150 mm in length and of 7 mm i.d., heated by current from an electric-arc welder, is used. The sample is introduced into the tube on a carbon cup inserted through a side-arm of the tube, or it may be nebulised and carried into the tube with an inert gas; Table I11 illustrates the sensitivities obtained with this apparatus for fifteen elements.No significant matrix effects were observed from the presence of relatively large amounts of aluminium, chromium, copper, iron, nickel, manganese, zinc and magnesium on the atomic-absorption signal observed for Winefordnerl has described the construction of a graphite cell system for AFS studies. This is shown schematically in Fig. 4. The graphite tube is heated by an electric-arc welder of 300-A output, and sample solutions are introduced into the cell with a hypodermic syringe and needle inserted through a rubber septum in the front of the cell housing.A low flow-rate of argon is used to provide a continuously flushed cell. Radiation from a 150-W concentrated xenon arc lamp, modulated at 500 Hz by a mechanical chopper, is used as excitation source. A photon counting system is used to record the fluorescence emission after its detection through a $-m grating monochromator and photomultiplier. Analytical results obtained with this system have not yet been published, but it is expected that the apparatus will provide a versatile tool for AFS. Headridge and Smith25 have reported the construction of a simple induction furnace for determining volatile elements in solutions and volatile matrices by AAS. The arrange- ment of graphite tube and side electrode for sample introduction is similar to that of the L'vov furnace.The furnace is inductively heated to 1350 "C for routine use, although the maximum attainable furnace temperature is 1900 "C. The authors describe the use of the furnace for the determination of cadmium in 5-mg samples of zinc-base alloys within the concentration range 5 to 400 pg g-l by using the cadmium 326.1 nm line, and in microlitre solution volumes, cadmium can be determined in the range 1 to 20 ng by using the cadmium 228.8 nm line. g of silver. FIL.4MENTS- Atomisation techniques in which a wire loop or a sample boat carrying the sample is introduced into the hot flame gases above the primary reaction zone of pre-mixed flames can be traced back to the original work of Runsen26 in flame-emission photometry; several devices of this type have been re-introduced in the past 5 years for use in AAS.Several non-flame devices of the same type, but in which an electrically heated filament or boat is used, have also been reported recently. In these open devices, in which the atomic vapour released from the filament passes into an unconfined volume in the absorption or fluorescence light path, transient analytical signals are obtained, and it may be difficult to achieve freedom from interference efTects when a large temperature gradient exists between the hot filament and the cooler volume above it in which atomic-absorption or atomic-fluorescence measure- ments are made. Because of the unconfined nature of the analytical cell volume in these devices, however, any tendency towards memory effects with some analyte elements may be minimised.U l f ~ a r s o n , ~ ~ Brandenburger and Bader28 and Brandenb~rger~~ were able to detect nano- gram amounts of mercury by collecting it as an amalgam on a wire and then heating the wire to vaporise the mercury into the optical path of an atomic-absorption spectrophotometer. Brandenburger and Bader30131 have also used this technique with a closed system in which the mercury vapour is confined within an absorption tube after vaporisation from the filament. The filament technique has also been applied to the AAS determination of cadmium, zinc, lead, tellurium, copper, silver, gold and platinum after electrolytic deposition on a wire filament. Bratzel, Dagnall and Winef~rdner~~ have described the application in AFS of a filament technique similar to that used by Brandenburger and Bader.In the device used by the former workers the sample solutions are vaporised from a platinum loop, which is positioned in a flowing stream of inert shielding gas. The vaporised analyte element is swept into the fluorescence light path where atomic fluorescence is excited by the electrodeless discharge tube; a d.c. electrometer system is used to measure the atomic-fluorescence signal. Typical absolute detection limits reported are cadmium, g ; mercury, g; and gallium, 10-7 g.September, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 617 A H 1 H Fig. 5. Filament atom reservoir (from Alder and West3') : A, base ; B, water-cooled electrodes; C, water link between electrodes ; D, laminar-flow box; E, inlet for shield gas; F, support stem for reservoir; G, transformer terminals; H, water inlet and outlet; and J, filament have reported the construction and use in AAS and AFS of a device in which a graphite filament 2 mm in diameter and about 20 mm in length, supported by water-cooled stainless-steel electrodes, can be heated to between 2000 and 2600 "C within 5 s by passing a current of about 100 A at 5 V through it.The small liquid samples (5 p1) are placed on a depression in the filament and the assembly is housed within a chamber with quartz windows; this chamber can be purged with argon. The original device was used for the detection of silver and magnesium by AAS and AFS ; 10-l0-g amounts of these elements were found to produce 1 per cent.absorption in AAS, while in AFS 10-l6g of magnesium and 3 x 10-11g of silver could be detected. An improved design of the device, in which inert-gas shielding of the filament eliminates the need for a closed chamber assembly, is shown in Fig. 5.34 Table IV shows the limits of detection and maximum amounts for a range of elements that can be determined by AFS with this type of assembly. Typical sensitivity values (for 1 per cent. absorption) and detection limits for AAS work with similar devices are given in Table V. As with other non-flame cells, e.g., the L'vov and Massman furnaces, in West and TABLE IV DETECTION LIMITS AND UPPER DETERMINATION LIMIT DATA IN AFS WITH FILAMENT ATOM RESERVOIR OF WEST et nl. Maximum determinable Reference Element Limit of detection/g* amountlg (West el al.) 1 x 10-12 2 x 10-9 35 3G 4 x 10-12 - Ag Au Bi 1 x 10-1' 1 x 10-8 35 Cd 1.5 x 10-13 - 34 co 2 x 10-1' 6 x 37 c u 1 x 10-12 4 x 1 0 - 9 37 Ga 5 x 10-11 1 x 10-8 35 1 x 10-12 1 x 10-9 35 5 x 10-12 2 x 1 0 - 9 37 Mg Mn Ni 5 x 10-12 5 x 10-9 3 7 Pb 1 x 10-1' 1.5 x 1 0 - 7 35 Sb 1 x 10-9 3 x 10-8 37 T1 5 x 10-11 2 x 10-9 35 Zn 2 x 10-14 4 x 10-1'' 35 * Signal: noise = 2.618 KIRKBRIGHT: APPLICATION OF NON-FLAME ATOM CELLS IN [ANaZySI, VOI.96 TABLE V AAS SENSITIVITY DATA OBTAINED WITH CARBON FILAMENT ATOM CELL OF WEST et al. Sensitivitylg per 1 per cent. Reference Element absorption Detection limitlg (West et al.) A1 c u Mn Ni Pb 6 x 10-lo 3-3 x 10-11 5 x 10-11 2.4 x lo-’’ 7 x 10-12 2 x 10-9 5 x 10-11 5 x 10-1’ 3 x 10-10 5 x 10-1’ 39 39 40 39 39 which pulse atomisation is effected, it is important that the response time of the electronic circuitry used is fast enough to permit accurate recording of the AAS or AFS signal versus time relationship.Under these conditions either peak height or integration methods of measurement can be used. Additionally, with the filament device, and in contrast to the L’vov and Massman furnaces, no further energy is available in the space above the filament immediately after the vaporisation pulse to prevent condensation of the atomic vapour. The decay of the atomic population is therefore promoted by condensation of analyte atoms with their own species or with the atoms or molecules formed from concomitant elements present in the sample. The effective lifetime of the free atoms may therefore be considerably shorter than when a furnace system is used. Hence it is necessary to view the atomic vapour in AAS or AFS close to the filament if inter-element effects are to be minimised.Fig. 6 illus- trates the manner in which the absorbance obtained for copper, nickel, lead and aluminium decreases as the height of observation above the filament is increased.39 Interference effects from volatile matrix elements have been observed with the carbon-filament device to be generally more serious than those observed from these elements in flames; interference effects from refractory oxide-forming elements, however, appear to be significantly less troublesome with this device than the effects observed in flame cells. This most probably arises from the fact that inter-element effects occur in the vapour phase above the filament on condensation of those (more volatile) elements which readily vaporise from it, whereas it is difficult to vaporise the more refractory elements from the filament. 5 10 0.04 0 5 10 0 5 10 Observation height above filament/mm Fig.6. Variation of absorbance with height of observation above filament for various elements (from Anderson, Johnson and Wests): (a) copper; (b) aluminium; (c) lead; and (d) nickelSeptember, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 619 Amos41 has described the use of a modified form of the carbon-filament device used by West et al.; a hole is drilled through the rod to form a sample cavity into which the small liquid sample can be introduced, and this cavity may permit the use of smaller sample volumes (0.5 to 1 pl) for AAS. Additionally, Amos describes the use of the filament device within a hydrogen-diffusion flame, which is produced by replacement of the argon or nitrogen shield gas by hydrogen.In operation, the glowing rod then ignites the hydrogen and a diffusion flame results. These modifications may give rise to improved atomisation, partic- ularly for less volatile elements, and tend to minimise inter-element effects. In a study of the determination of lead in biological materials, for example, Amos et ~ 1 . ~ ~ observed con- siderably less serious spectral and chemical interferences from other ions when the rod was surrounded by the hydrogen-diff usion flame. Vacuum/Gas Fig.7. Absorption chamber and filament assembly used by Donega and Burgess43 Donega and Burgess43 have assembled a filament device similar to that used by West and co-workers, but in which sample boats are cut from graphite sheet or tantalum or tungsten foil to be 50 mm long and 6 mm in width. The assembly used is shown in Fig. 7. The sample boat is heated electrically with a current of 30 to 50 A at 12 V, which is sufficient to heat the boat to about 2200 "C in less than 0.1 s. The two copper rods that support the sample boat are insulated and pass through the brass base-plate, which contains an O-ring seal and to which a quartz window is also attached. The filament assembly is enclosed in a quartz tube 50mm in diameter, which has an optical quartz window sealed on one end and an O-ring flange on the other.The chamber can be purged with inert gas and used at operating pressures between 1 and 760 torr. The device can be used with volumes of liquid samples as large as 50 to 100 p1. The mode of operation is similar to that used by West et al. for the carbon-filament atom cell. Table VI shows some sensitivities obtained. No results for the TABLE VI ABSOLUTE SENSITIVITIES REPORTED FOR FILAMENT TECHNIQUE OF DONEGA AND BURGESS43 Element Al Cr cu Mn Mo Ni Pt Si Na V Wavelength 309.3 389.4 324-8 279-5 313.3 341.6 265.9 251.6 589.0 318.4 Sensitivitylg 4 x 10-10 6 x 10-l2 3 x 10-12 3 x 10-12 3 x 10-0 3 x lo-@ 3 x 10-11 6 x 1O-Io 3 x 10-7 3 x 10-9 Operating conditions* 1 2 2 2 3 2 2 2 4 1 * 1. Tantalum boat, 300 torr hydrogen. 2.Tantalum boat, 300 torr argon. 3. Tungsten boat, 1 torr argon. 4. Graphite boat, 100 tom argon.620 KIRKBRIGHT: APPLICATION OF KON-FLAME ATOM CELLS IS [Analyst, Vol. 9G effect of matrix elements on the sensitivity have yet been reported. The method of operation at low pressure should result in low background absorption interference, in the same manner as that reported by Massman for the hot hollow-cathode technique, but this can be achieved only at the expense of short residence time in the effective absorption volume and the require- ment of a fast-response detector system. CATHODE-SPCTTERING CELLS- The sputtering action, which is responsible for the production of atomic vapour within a hollow-cathode lamp, can be exploited for the atomisation of samples for AAS.Thus Walsh and c o - w ~ r k e r s ~ ~ ~ ~ ~ constructed a sputtering ,chamber in which metal samples, machined to the shape of an open-ended hollow cathode, were clamped. The chamber was then evacuated, filled with argon at the required pressure, and a discharge initiated as in a conventional hollow-cathode source. The sputtering chamber was fitted with silica windows and could be placed in the light path of an atomic-absorption spectrophotometer. The determination of silver and phosphorus in copper, and silicon in aluminium and steel, with this device has been reported.46 A disadvantage of the technique is the need to prepare the cathode from the sample itself; this tends to restrict the technique to metal samples. In order to avoid this restriction, Goleb and Brodya7 evaporated sample solutions on to the inner wall of an aluminium hollow cathode, which was then used as part of a water-cooled de-mountable hollow-cathode lamp with a continuous flow low-pressure system. Only small sample volumes were required, and 1-pg amounts of sodium, calcium, magnesium, silicon and beryllium were detected ; considerable inter-element effects were observed.Goleb4* has also applied this device as an atom source for the isotopic analysis of uranium. Ivanov, Gusinsky and J e ~ i k o v ~ ~ have used a graphite hollow cathode in a sputtering cell; in this way they achieved some of the advantages of the graphite cuvette technique in addition to those gained from the sputter- ing cell technique. Sample solutions were evaporated directly on to the walls or on to a fine molybdenum wire, which was then placed along the central axis of the cathode.Massmanso has described the use of a hot hollow-cathode assembly for AAS determina- tions with solid samples. In this apparatus a graphite tube (30 mm in length and of 7 mm i d . ) is supported on a small cylindrical graphite electrode that passes through the wall of the cathode tube. This carrier electrode also holds the solid sample in a depression 2 mm in diameter and 3 mm deep bored out of its end; the dimensions of this depression can be varied, depending on the sample size. The cathode assembly is mounted on a 2-mm diameter molybdenum rod whose insulated base passes through the water-cooled base-plate of the metal housing of the assembly. This housing acts as the anode and is at ground potential.The compartment is fitted with quartz windows and can be purged with argon. The dis- charge is then operated at an argon pressure in the range 1 to 10 torr at a power of up to 1 kW. When hollow-cathode cells of this type are used as the atom cell in AAS, the cathode continuum radiation and emission from the sample itself might be expected to interfere seriously with the measurement of the absorption of the radiation from the primary source. With cool cathodes this interference is difficult to avoid, as the atomic vapour persists only while the discharge current is being applied. Massman pointed out that this difficulty can be overcome with the hot-cathode device, as under suitable operating conditions atomic vapour may persist after the discharge current has been allowed to fall to zero.When the graphite hollow-cathode discharge is operated with a 50-H~ half-wave current, therefore, the discharge, which creates an atomic vapour, is present for only one half of the operating period. During the alternate discharge-free half-period the atomic vapour may still persist, so that its absorp- tion of the primary source radiation can be measured without interference. In the instru- mental assembly described by Massman, the radiation from the hot-cathode atom cell itself is prevented from being recorded at the monochromator - detector assembly by the use of a rotating sector, which is driven in phase with the discharge current cycle and is placed between the cell and the monochromator slit. This sector transmits the primary source radiation only during a prescribed period (just less than one half cycle) while the discharge current is at zero. The primary hollow-cathode lamp radiation is modulated at 450 Hz by a second rotating sector placed between the source and the graphite cathode cell.When the peak-absorption signal is measured, linear calibration is possible only over a narrow working range, e.g., 1 to 10 p.p.m. of silver in a 30-mg sample of lead. When signal integration is used it is possible to achieve improved linear range. Thus, for example, 1 to 300 p.p.m. ofSeptember, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 62 1 silver in lead then yields a linear working graph. In the analysis of lead, zinc or aluminium samples (30 mg) coefficients of variation of 8 to 15 per cent.were obtained. In the deter- mination of silver in lead with integration of the absorption signal a standard deviation of 4 per cent. was achieved. The hot graphite hollow cathode cannot be heated above about 2000 OC, so that only those samples whose evaporation proceeds rapidly below this temperature can be analysed. The method is more suitable for the determination of relatively low concentrations of volatile elements in milligram amounts of samples such as lead, zinc or aluminium than for the determination of small absolute amounts of analyte element. This situation results from the short residence time of the atomic vapour in the absorption volume at the low pressure used. The vaporisation rate of samples in which metals such as iron, nickel and copper are the major elements is too slow to allow their effective analysis by this technique.When these elements are present as impurities in more volatile matrices, however, the evaporation rate may be sufficiently high to permit their determination. Table VII shows the detection limits obtainable with this device for the determination of various elements in 30-mg samples of relatively volatile matrices. These values are compared with the detection limits obtained for the same samples by emission spectrography with a hot hollow cathode and a medium quartz ~pectrograph.~~ Even with large sample weights (100 mg) of lead, zinc, antimony or aluminium, less than 1 per cent. interference is encountered from background absorption at wavelengths greater than 220 nm.From this point of view the hot graphite hollow cathode operated at low pressure may be superior to graphite furnaces operated at atmospheric pressure for large samples. TABLE VII DETECTION LIMITS WITH HOT HOLLOW CATHODE FOR DETERMINATION OF VARIOUS ELEMENTS I N SOLID METAL SAMPLES (FROM lf.4SSMAN50*”) Element 2 Zn c u Cd w Mn Cr AAS method Sample -n 30 mg Linelnm 1imitlp.p.m. 328.0 0.3 231.1 50 213.9 0.05 Lead 324-7 2.8 228.8 0 01 Zinc 285.2 1.2 279.5 2-2 Aluminium 357.9 10 Atomic-emission method , Detectioh Linelnm 1imitlp.p.m. 328.0 0-1 252-85 1.0 213.9 0.5 324.7 0.3 228.8 0.2 - - OTHER NON-FLAME CELLS- The use of the d.c. arc for sample atomisation in AAS has been described by several Robins0n~~9~6 has reported that 20 per cent.absorption was obtained for aluminium by using a spark to atomise the sample; the solution was nebulised in the con- ventional fashion and the aerosol was passed between two electrodes between which the spark was struck. There have been several reports of the use of RF or microwave plasma sources to atomise samples for AAS. Wendt and F a ~ s e l ~ ’ , ~ ~ used an induction-coupled plasma together with a triple-glass optical system to obtain large absorbance signals for AAS. Friend and Diefenderfc~~~ investigated the possibility of using a plasma jet for the deter- mination of refractory elements. The RF plasma has recently been applied to AAS by three groups of These plasma systems, some of which have previously been used extensively for analytical emission spectroscopy, show great promise for AAS.The high energy available should permit considerable freedom from matrix interference effects, and the possibility of transferring samples to the plasma with low inert-gas flow-rates should result in high atomic concentration for the analyte element and high sensitivity. Greater control over the chemical environment can be achieved with plasma than with flames, and consequently the lifetime of free ground-state atoms may be greater in these atom cells. The use of lasers offers the possibility of direct atomisation for the examination of solid surfaces by AAS. The heat produced at the solid when a focused laser beam strikes it may vaporise the solid material over a small area of the This technique has been622 [Analyst, Vol.96 used in spark-emission spectroscopy; a spark is passed across electrodes placed just above the surface so that it passes through the plume of vapour produced by the laser. In AAS this direct sampling and atomisation method should provide for low-emission background and high sensitivity for small solid samples. In the laser microprobe - BAS assembly devised by Mossotti, Laqua and HagenahG2 a pulsed laser system is used to produce rapid heating and vaporisation from the sample surface. The atomic absorption in the transient atomic vapour produced is then measured with a fast-response system. Venghiattisa6 has proposed a technique for the direct conversion of solid samples into atomic vapour, in which the powdered sample is mixed with a solid-propellent powder; the mixture is then compressed and ignited.The atomic vapour of the analyte element in the gases of the “flame” that is produced then passes into the absorption light path of a con- ventional atomic-absorption spectrophotometer. Calibration can be performed by the similar technique of mixing standard samples with the propellent mixture. The technique has been applied to the determination of trace elements such as gold, silver, copper, lead and zinc in ores. Although it may be difficult to achieve better than 5 per cent. precision without very careful sample preparation, this technique may find considerable application in the field or “on-site” where the transportation or use of gas cylinders, solutions and chemicals is inconvenient. CONCLUSION KIRKBRIGHT : APPLICATION OF NON-FLAME ATOM CELLS IN The development of non-flame atom cells during the past few years has shown these devices to provide capabilities complementary to those of flame cells.High absolute sensi- tivity is attainable, which permits very small samples to be examined effectively by AAS. The analytical selectivity is high for these applications. Provided that suitable sources and detector systems are available, the use of non-flame cells should be of considerable benefit in AFS, with which the low emissive background obtainable in suitably designed non-flame cell systems should assist the attainment of high sensitivity. The results for precision obtained with many of the devices reviewed here are not as satisfactory as those obtainable with flame systems; when both the stability and response time of the detector as well as the cell design are optimised, however, it is the reproducibility of the sample transfer to the cell that may at present limit the precision attainable.Efforts to improve the sample transfer repro- ducibility are warranted only if the 5 to 10 per cent. coefficients of variation typically attained in normal working are unsatisfactory. For very small samples containing traces of analyte element, coefficients of variation of about 5 to 10 per cent. may often be acceptable. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. REFERENCES Winefordner, J. D., in Dagnall, R. &I., and Kirkbright, G. F., Editors, “Atomic Absorption Spec- Plenary Lectures presented a t the International Atomic Absorption Spectroscopy Mislan, J.P., Paper presented a t 7th Conference on Analytical Chemistry in Nuclear Technology, Hudson, R. D., Phys. Rev., 1964, 135, 1212. TTidale, G. L., Space Science Laboratory Aerospace Operation, General Electric T.I.S. Report, Choong, S. P., and Loong-Seng, W., Nature, 1964, 204, 276. Tomkins, F. S., and Ercoli, B., APPZ. Optics, 1967, 6, 1299. King, A. S., Astrophys. J., 1908, 27, 353. L’vov, €3. V., J . Engng Phys., 1969, 2 (2)’ 44. -, Ibid., 1959, 2 (ll), 56. -, Spectrochim. Acta, 1961, 17, 761. -, Zav. Lab., 1962, 28, 931. -, Optikn Spektvosk., 1966, 19, 507. L’vov, B. V., and Pljush, G. V., “Proceedings of the 15th Symposium on Spectroscopy,” Minsk, 1963, Volume 2, p. 159. Academy of Sciences, U.S.S.R., 1964.L’vov, B. V., and Lebedev, G. G., Zh. Prikl. Spectrosk., 1967, 7, 264. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” Adam Hilger, London, 1970. - , Spectrochim. Acta, 1969, 24B, 53. L‘vov, B. V., and Lebedev, G. G., “Symposium on Theoretical Spectroscopy,” Yerevan, 1966. Nikolayev, G. I., and Aleskovskii, V. B., Zh. Tekh. FZz., 1964, 34, 753. Massman, H., Spectrochinz. Acta, 1968, 23B, 215. - , Z . annlyt. Chem., 1967, 225, 203. Manning, D. C., and Fernandez, F., Atomic Absorption Newsletter, 1970, 9, 65. Woodriff, R. , Stone, R. IT., and Held, A. AT., Aihpl. Spectrosc., 1968, 22, 408. troscopy. Conference,” Sheffield, 1969, Butterworths & Co. (Publishers) Ltd., London, 1970, p. 37. Gatlinberg , Tennessee, 1963. K605D330, 1960.September, 19711 ATOMIC-ABSORPTION AND ATOMIC-FLUORESCENCE SPECTROSCOPY 62 3 23.24. 35. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 55. 63. 5 1. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. Woodriff, R., and Ramelow, G., Spectvochim. Acfa, 1968, 23B, 665. Woodriff. R., and Stone, R. W., A$@. Optics, 196S, 7, 1337. Headridge, J. B., and Smith, D. R., Talanta, 1971, 18, 247. Bunsen, R., Justus Liebigs Annln Chem., 1859, 111, 257. Ulfvarson, U., Acta Chem. Scand., 1967, 21, 641. Brandenburger, H., and Bader, H., Helv. Chim. Acta, 1967, 50, 1409. Brandenburger, H., Chimia, 1968, 22, 449. Brandenburger, IL, and Bader, H., Atornu Absorption Newsletter, 1967, 6, 101. -- , Ibzd., 1968, 7, 53. BratLel, AT. P., Dagnall, R. M., and Winefordner, J. D., Ibid., 1969, 48, 197. West, T. S., and Williams, X. K., Analytica Chim. Acta, 1969, 45, 27. Alder, J . F., and West, T. S., Ibzd., 1970, 51, 365. Anderson, R. G., Maines, I. S., and West, T. S., Ibid., 1970, 51, 355. Aggett, J., and West, T. S., Ibid., 1971, in the press. Alger, D., Anderson, R. G., Maines, I. S., and West, T. S., lbzd., 1971, in the press. Aggett, J., and West, T. S., Ibid., 1971, in the press. Anderson, R. G., Johnson, H. N., and West, T. S., Ibzd., 1971, in the press. Ebdon, L., Kirkbright, G. F., and West, T. S., Ibid., 1971, in the press. Amos, M. D., Amer. Lab., 1970, August, 33. Amos, M. D., Bennett, P. A., Brodie, K. G., Lung, P. W. Y., and Mstousek, J. P., AnaZyt. Chern., Donega, H. M., and Burgess, T. E., Ibid., 1970, 42, 1521. Jones, W. G., and Walsh, A., Sfiectvochim. Acta, 1960, 16, 349. Gatehouse, B. M., and Walsh, A., Ibid., 1960, 16, 602. Walsh, A., “Proceedings of the Xth Colloquium on Spectroscopy, International,” Hilger, 1962, Goleb, J. A,, and Brody, J. I<., Aizalytica Chin+. Acta, 1963, 28, 457. Goleb, J. A, Analyt. Clteni., 1963, 35, 1978. Ivanov, N. P., Gusinsky, M. N., and Jesikov, A. D., Zh. Analit. K l h i w z . , 1965, 20, 1133. Massman, Ii., Spectrochzm. Acta, 1970, 25B. 393. - , “Proceedings of the IXth Colloquium on Spectroscopv, International,” Hilger, 1961, MarinkoviC, &I., BojoviC, B., and PesiC, D., “Proceedings of thc XIIIth International Colloquium Belyaev, U. I., Ivantsov, L. M., Karyakin, A. V., Phi, P. H., and Shemet, V. V., Zh. Analit. Khim., Icantor, T., and Erdey, L., Spectvochim. Acta, 1969, 24B, 283. Robinson, J. W., Analytica Chim. Acta, 1962, 27, 465. -, Ind. Chemist, 1962, 38, 336 and 362. Wendt, R. I<., and Fassel, 17. A., Analyt. Chew., 1965, 37, 920. -- , Ibid., 1966, 38, 337. FrieAd, K. E., and Diefenderfer, A. J., Ibid., 1966, 38, 1763. Veillon, C., and Margoshes, M., Spcctrochirn. Acta, 1968, 23B, 503. Greenfield, S., Smith, P. B., Breeze, A. E., and Chilton, N. 31. D., Analytica Chim. Acta, 1968, Mossotti, V. G., Laqua, K., and Hagenah, W. D., Spedrochim. Acta, 1967, 23B, 197. Rasberry, S. D., Schribner, B. F., and Margoshes, M., AFP. Ofhcs, 1967, 6, 81 and 87. Karyakin, A. V., and Kaigorodov, V. A., Zh. Amzlzt. Klzim., 196S, 23, 930. T’enghiattis, A. A., Atomic Absorption Newsletter, 1967, 6, 19. 1971, 43, 211. p. 127. Volume 11, p. 170. on Spectroscopy,” Hilger, London, 1967, p. 1181. 1968, 23, 508. 41, 385. Received April 26th. 1971 Accepted June loth, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600609

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

624 Analyst, September, 1971, Vol. 96, pp. 624-630 SLEPT: A Simple Computer Language for Examining Data Recorded on Punched Paper Tape BY C. R. BOSWELL* (-4 nalytical Research and Development Unit, Atomic Energy Research Establishment, Harwell, Didcot, Berks.) A versatile language has been developed for the computer processing of paper tape output from multi-channel analysers. It is intended for use in laboratories in which the work load is too varied to justify writing a program, or suite of programs, dedicated to a specific set of operations. The language consists of a set of commands and associated information, which is read from cards and executed by a program written in FORTRAN IV. The program has a modular structure so that new commands can easily be incor- porated into the language as required.hfANY nuclear experiments result in data, such as pulse height spectra, which are stored in multi-channel analysers and subsequently punched on to paper tape. In many instances the data required from the resultant tapes are few, e.g., peak locations and areas, or the area under selected peaks. A large number of versatile and comprehensive programs have been written for digital computers for processing such information. For example, Wangen and Isenhourl have described the semi-quantitative analysis of mixed y-ray spectra by computerised learning machines, while Hoffman and Wainerdi2 have developed a computer program for least squares resolution of y-ray spectra by using half-life information as well as y-ray energies. Dams and Adams3 have developed a program by using about 2000 precise y-ray energies of 250 radioisotopes for the computer-assisted identification of individual y-ray emitters in complex spectra.Black4 and Connelly and Black5 have investigated the use of cross-correlation techniques for the identification of structure in y-ray, X-ray and neutron spectra, and for the calculation of peak areas. Several other programs are also available for the reduction of y-ray ~pectra,~,~98 and Yuleg has given an extensive discussion and review of the type of calculation carried out on pulse height spectra. Many of the computer programs described in the literature have been developed for specific systems in which the tasks they are required to undertake are well defined. Also, they are used extensively, so that the effort expended in writing such programs can be readily justified. In many laboratories, however, particularly those conducting research and de- velopment, the number of tasks performed is varied, with little common activity between tasks.For example, one laboratory may be carrying out fast-neutron and thermal-neutron activation analysis, charged-particle activation analysis , prompt y-ray studies, multi-scaling exercises, etc. Under such circumstances the data may consist of a relatively small number of spectra that have been obtained by using a set of unique instrumental settings. The intensity of effort on any one exercise may, therefore, not justify a large expenditure of pro- gramming time and expertise dedicated to each of the specific studies.In this type of situ- ation, most of the information is often accumulated in multi-channel analysers operating in either the pulse height analyser or time-scaling mode, and recorded on punched cards, magnetic tape or punched paper tape before being further processed by an off-line digital computer. SLEPT is a program written in FORTRAN IV and implemented on an IBM 360/75 computer. It has a simple language structure and has been developed for the off-line * Present address : c/o Computer Unit, Massey University, Palmerston North, New Zealand. 0 SAC; Crown Copyright Reserved.BOSWELL 625 processing of information collected in multi-channel analysers and stored on punched paper tape. It consists of a series of commands that enables relatively simple operations to be carried out on the data to be manipulated.TABLE I VOCABULARY OF SLEPT Command TITLE .. EDIT .. DROP .. CHANNEL DIGIT .. FIND .. SKIP .. BGD .. DIFF .. BEDIT .. BDROP .. BDIFF .. STANDARD SMOOTH .. SUB .. AREA .. COVELL .. PEAKS . . LIST .. PUNCH .. PLOT .. SCALE .. LINE . . POINT .. END .. Blank .. .. .. .. .. .. .. .. .. .. .. .. .. .. . . .. .. .. . . .. .. .. .. .. .. . . .. Information Title associated with spectrum Number of channels between channel Number of channels a t beginning of Number of channels in spectrum identifiers spectrum to be dropped Maximum number of digits in chan- Number of spectrum required nel content Number of spectra to be skipped Title associated with background - Number of channels between channel identifiers Number of channels at beginning of background spectrum to be dropped - Two numbers the ratio of which gives Number giving number of channels the scaling factor to be used for smoothing - Two numbers identifying two chan- nels Two numbers giving approximate peak position and number of chan- nels either side of peak Number giving the number of chan- nels used for smoothing Number giving channels per line - New title (if desired) New title (if desired) - Anything Function Converts image of paper tape into Drops channel identifiers included in Drops the first “n” channels integer values paper tape output Allows the first “n” only channels of Controls number of digits per channel a spectrum to be read in Permits jumping to any specified Omits the specified number of spectra Reads in a spectrum to be used as Calculates first forward differences ,4s for EDIT but operates on spec- trum read in by using BGD As for DROP but operates on spec- trum read in by using BGD spectrum before executing next TITLE a background As for DIFF but operates on spec- Adjusts the contents of all channels Smooths raw data trum read in by using BGD by a scaling factor Subtracts background spectrum from spectrum under consideration Sums the contents of all channels between and including those speci- fied in information field Finds exact peak maximum and obtains Cove11 area Smooths the raw data and locates Lists channel contents on line printer Punches channel contents on to Plots data on plotter and line printer Plots data, with scale expansion, on Specifies line plotting on plotter Specifies point plotting on plotter Indicates end of data (optional) If first four columns are blank treats remainder of the card as comment peaks and valleys in spectrum cards plotter and line printer GENERAL FEATURES- Because of the diverse nature of the type of information stored on the paper tape and the variety of processing required, a simple language structure has been evolved for processing the data. The paper tapes are read into the computer, and the codes corresponding to each626 BOSWELL: SLEPT-A SIMPLE COMPUTER LANGUAGE FOR [Analyst, Vol.96 character on the tape written sequentially on to magnetic discs. All that is required of the operator who needs to manipulate the stored data is that he should submit a series of punched control cards containing commands referring to the data manipulation required.The control cards have been designed so as to have maximum clarity and as little format control as possible, and the control commands have been kept as close as possible to the actual words used to describe the operations performed. The card is split into two fields, the command being stored in the first eight columns and all other required information, in free format, in the remaining seventy-two columns of the card. When the command is executed by the program the data in the information field are processed as specified by the command. For example, when the command is that which requires a new spectrum to be read in, the data in the information field are taken as the title of the associated spectrum.If, on the other hand, the command is to find the area within a specified region of the spectrum, the data in the information field will give either the limits of integration or the approximate position of the peak, together with the number of channels either side of the peak over which integration is required, depending on the command given. Cards, each containing one command with the associated requisite information, are separately and sequentially processed until all have been treated. Because of the combination of commands resembling the opera- tions specified, the restriction of one command to each card and the sequential processing of such cards, the input to the program thus has the appearance of a simple programming language. This has the advantage, among others, that it is very simple to learn, and involves the use of the minimum number of mnemonics and rules.Restriction of one command and its associated information to each card and the sequential processing of cards - commands enable a program of treatment given to any set of spectra to be easily and logically written and clearly followed. The program has been designed in a modular form so that, as new techniques for treating the data are required, suitable sub-routines can be added to the system and corresponding new commands added to the vocabulary of the language. CURRENT STRUCTURE- A list of the commands currently implemented in the language, together with the associated information and an outline of the corresponding action, is given in Table I.The commands can be categorised into three groups: (a) input of data and associated editing; ( b ) data processing; and (c) output of data; and will be discussed according to this classification. Input apzd editing-Several analysers are available in this laboratory giving data output on to eight-hole paper tape in ASCII code under a number of different formats. These include either five or six digits per channel with, in some instances, the channel number at regular intervals for channel identification. In addition, output from one analyser is often produced in such a way that the sum of all previous channel contents is punched, rather than specific channel contents. To allow for the different number of digits per channel a command DIGIT is provided with the appropriate number in the information field to specify this value.Unless otherwise specified there is a value of six digits per channel automatically incorporated in the program. The command TITLE, on execution, reads the image of the paper tape stored on disc and converts it into a series of integer numbers corresponding to channel contents (and channel numbers if these are punched on to the tape). The data in the information field of a command card are assumed to be an alphanumerical identification of the spectrum, and are printed on the line printer as the command is executed. If channel identification is punched on to the tape, these values will be included in the data obtained with TITLE and can be eliminated with EDIT command. For the latter, the integer value in the information field specifies the number of channel contents punched between channel identifiers.Data punched on to tape in the integration mode can be restored to channel contents (after EDIT-ing, if necessary) by obtaining the forward differences with the command DIFF. Channels containing unwanted information a t the beginning of the spectrum can be deleted with the DROP command. In this event, the integer in the information field gives the number of channels to be dropped. A background spectrum can be read in and stored separately by using BGD, which is effectively a TITLE command but with reserved storage for the resultant spectrum for subsequent background subtraction exercises on other spectra; BEDIT, BDROP and BDIFF are EDIT, DROP and DIFF commands, which operate only on the background spectrum.The command FIND is an instruction to prepare to start input on the nth spectrum in the sequence of spectra stored on the paper tape, where n is an integer in the informationSeptember, 19711 EXAMINING DATA RECORDED ON PUNCHED PAPER TAPE 627 field. In normal use a new TITLE card causes the immediately subsequent spectrum on the tape to be read in. The use of the FIND command enables the spectra to be read in and treated in any desired order. In a similar manner, SKIP causes n spectra to be skipped before executing the next TITLE command, where n is an integer in the information field. Data processing-Typical operations carried out on the resultant spectrum are peak location, the calculation of the number of counts between specified channels with background correction, determination of the area under a peak by using the method of Covell,lo spectrum smoothing, the subtraction of a background spectrum and the multiplying of all the channel contents of a spectrum by a constant factor to allow standardisation to a normalised flux, live time, etc.TABLE I1 EXAMPLE OF A SLEPT PROGRAM Tape 4CW6 Spectra of irradiated samples with decay curves taken 10 minutes after irradiation TITLE EDIT 8 LINE PLOT LIST COVELL 1767 TITLE EDIT 8 DROP 3 POINT PLOT PUNCH LIST CHANNEL 256 TITLE EDIT 8 LIST LINE PLOT COVELL 1767 CHANNEL 512 NO. 1, 5 MIN. IRR., 2 MIN. DELAY, 1 MIN. COUNT NO. 1, MULTISCALING, 1 SEC. STEPS NO. 2, 1 MIN. IRR., 1 MIN. DECAY, 5 MIN. COUNT TITLE EDIT 8 NO. 2, MULTISCALING, 5 SEC.STEPS DROP 3 POINT PLOT LIST PUNCH END A control card with the command SUB causes a 1 : 1 subtraction of a background spectrum, which is previously read in with a BGD command, from the spectrum currently being manipu- lated. An AREA control card instructs the program to calculate the total number of counts in the region bounded by the two channel numbers in the information field. The command SMOOTH permits smoothing of the spectrum, by quadratic convolution,11~12 the number of points involved in the smoothing being given by the integer in the information field. The PEAKS command produces a listing of the channel numbers at which peaks and valleys occur in the spectrum. The criterion used is the change in sign of the first differential of the smoothed data and a statistical test for a peak.The number in the information field specifies the number of channels over which smoothing takes place. The area under a peak with a local maximum near channel “n” and “2m + 1” channels wide is obtained with the COVELL instruction, in which n and m are specified in the information field. The peak maximum does not need to be specified accurately as the program searches for the local maximum in the region. The command STANDARD permits normalisation to a given value of neutron flux, beam current, counting time, etc. Two values are provided in the information field, the first being the value of the parameter for the spectrum (actual flux, etc.) and the second the desired value. All channel contents are multiplied by the ratio of these numbers.BOSWELL: SLEPT-A SIMPLE COMPUTER NO.1 5 H I N I R R , 2 MIN- DELAYS I MIN. COUNT MAXIMUR COUNT I N llNY CHANNEL - 3903 NUMBER OF COUNTS PER LINE I N PbCE PLOT - 78 NUMBER OF CHANNELS I N SPECTRUM - 512 NUMBER OF CHINNElS PER POINT I N PLOT - 4 ZCRE 946 @OSbiELL POINT 2 3QOO LOO 98 96 94 92 * 9 0 8 8 86 84 82 80 78 76 72 7 0 6 8 66 64 62 60 5 0 C 56 0 54 U 52 N 5 0 T 48 S 46 44 42 40 3 8 3 6 34 32 30 28 26 24 22 20 18 16 14 12 10 8 . 6 4 2 74 * * * * 4 * k * * ** * * * * * +** *. *.* ** +*+* 4 * * + LANGUAGE FOR [ArtaZyyst, Vol. 96 0 10 2P 30 40 5 0 60 70 8C 90 100 110 120 130 140 1sc 160 170 180 19F 20@ 210 220 230 240 COVFLL RETPOD C 390 3 2745 1636 l C Q 3 1036 P 0 8 714 t e 2 642 E t 9 612 62 2 716 P54 1'25 24C4 399 14 18 9 8 4 4 77 r 333 3 159 2488 1546 1105 905 769 737 6 l h 65R =93 6 1 8 693 771 846 445 22A 2.544 2 3 8 17 19 13 6 5 6 165 3743 2619 1571 1051 095 790 722 6h7 6 3 1 0 591 7R6 703 R30 4 20 2 807 14R 13 7 14 7 7 279 r 8 4 3 3531 2466 1600 978 896 71 0 bh5 6 4 8 543 567 R13 R1 I 335 341 2bR7 96 19 1 3 14 7 10 12 7 1 3 729 1293 3526 7757 1537 1000 841 68b 723 685 602 571 625 907 719 803 209 489 2466 60 14 19 5 4 1n 1 3 1 7 9 6 3610 2169 1 4 2 5 1010 81 5 689 1 3 8 b64 57 5 61 6 63 3 880 74 0 84 1 24 7 71 R 2130 62 1 9 17 1 4 7 l o 17 2144 3383 1893 1373 971 858 664 671 5 0 5 588 663 908 745 726 232 991 I 6 5 5 2 7 17 1R 6 19 5 8 703 3083 3097 1899 I 7 8 7 912 864 660 696 675 65R 517 5 79 87 1 76'2 664 2 19 1 3 5 1 1230 24 1 3 I h 10 5 3 17 COUNTS I N REGION C€TWFEN CHANNELS.166 AVD 180 TOTAL COUNTS = 2645R BACKGROUND = 9723 BY TRAPELflIOAL INTFGRATION DIFFERENCE = 16735 Fig.1. Partial listing of results of first LIST command and that produced by COVELL command 3526 3507 1893 1879 1169 1 1 4 1 969 965 855 8 25 690 6 52 625 659 6 5 7 6 34 584 5 85 587 5 89 6 19 651 821 R I D 747 813 6 2 3 5 93 197 708 1724 7077 931 611 19 18 1 9 22 11 15 10 0 I 1 1 3 h 0 1 8 16 3007 2067 the subsequentSeptember, 19711 EXAMINING DATA RECORDED ox PUSCHED PAPER TAPE 629 Output of data-The spectrum obtained after manipulation with any of the above com- mands can be printed on the line printer by using the command LIST with the number of channels printed per line given in the information field. Alternatively, it can be stored on punched cards for further processing with the control card PUNCH, while PLOT enables graphs of the data to be produced on both the line printer and an off-line plotter. Plotting with scale expansion of the ordinate is accomplished with SCALE. If the information field is blank in both PLOT and SCALE commands, then that from the TITLE card is reproduced on the graph for identification.Otherwise the content of the PLOT or SCALE control card information field is included as a title in the graph. Two plotting modes are available and are chosen by the appropriate command, vix., POINT or LINE. Unless specified, line plotting is normally used and the selected mode is held until it is re-specified. Two other control cards are used: END specifies the end of a batch of cards and is optional because if it is not included in the deck, the sub-routine, which reads in the control cards, creates a dummy end card.A card that has the first four columns blank is assumed to be a comment card, and any data stored in the other 76 columns are reproduced in the control card listing but are not processed. A set of SLEPT cards is read and stored by the computer under the control of the main sub-routine of the program. Each card is then examined by a controlling sub-routine that checks the command against a dictionary of valid commands and calls the appropriate sub- routine. The content of the information field is transferred to this sub-routine as an array of literal data. If numerical values are required from this array they are extracted by the use of two further sub-routines, which effectively enable this array to be read, in free format, as a dummy data card.On completion of the processing specified by the command, control returns to the controlling sub-routine for processing of the next control card. In this example, the paper tape input consisted of a series of y-ray spectra, each separated by the An example of the type of SLEPT program possible is shown in Table 11. FRRnE NO I NO. I 5 WIN IRR. 2 WIN. DLLRT. 1 I I I Y . CWNt YO00 3500 3000 2500 2000 1500 I000 500 +' l:oo' ' I IU-' zmo j 3000 ' ' ' ' YODO ' t - 1 Fig. 2. Graph produced on the off-line plotter by PLOT command corresponding to results shown in Fig. 1630 BOSWELL results obtained from a multi-scaling operation. Graphs of the former, obtained by using the line mode, were required as was the Covell area of a peak under channel 176.Point graphs of the multi-scaling operations were required after the contents of the first three channels had been dropped. The multi-scaled data were also required to be punched on to cards for further processing. All of the paper tape output had channel identifiers, after each 8-channel content, which had to be removed by editing. Only the first 256 channels of the last spectrum were required to be processed. As 512 channels were recorded in the multi-scale mode, it was necessary to re-set the number of channels read in after the CHANNEL command had set the number of channels to 256. Listings of all data were required. Some of the output corresponding to the SLEPT program shown in Table I1 is given in Figs.1 and 2. A partial listing of the results of the first LIST command and that produced by the subsequent COVELL command is given in Fig. 1. Fig. 2 is the graph produced on the off -line plotter by the corresponding PLOT command. DISCUSSION and is listing 1 . 2. 3. 4. 5. 6 . 7 . 8. 9. 10. 11. 12. One of the criteria in designing the program was to maintain its modular structure with as little interaction as possible between various commands. The program therefore consisted of a controlling sub-program and a modular set of sub-routines, all of which are virtually independent of each other. After constructing the spectrum with the TITLE, EDIT and DROP commands, the channel contents are stored in two separate arrays, one of which is used for data manipulation as in SMOOTH, etc., while the other is effectively a “read only” store that can be used for restoring the channel contents after such manipulation.By design, all of the commands implemented in SLEPT perform relatively simple opera- tions, but the structure of the program is such that any new features can be added to the language with considerable ease. It enables a wide variety of tasks to be performed on different types of information stored on paper tape in a number of formats. The program, together with library plotting sub-routines, currently occupies 220 K bytes and, apart from the plotting routines, requires no special facilities. In the version used here the image of the paper tape is stored on discs, but modification of the appropriate sub-routines will enable it to be adapted to any other system. The program with the corresponding SLEPT language can therefore be used on any system (provided the core storage is available) capable of handling IBM FORTRAN IV. The use of SLEPT is particularly appropriate in laboratories that produce a wide variety of data which require simple non-routine processing. A SLEPT program is easy to write self-explanatory, and the language can be learnt and used in a very short time. A of the program is available from the author on request. REFERENCES Wangen, L. E., and Isenhour, T. L., Analyt. Chew., 1970, 42, 737. Hoffman, B. W., Wainerdi, R. E., J . Radioanalyt. Chem., 1970, 4, 375. Dams, R., and Adams, F., Ibid., 1970, 4, 311. Black, W. W., Nucl. Instrum. Meth., 1969, 71, 317. Connelly, A. L., and Black, W. W., Ibid., 1970, 82, 141. Mariscotti, M. A., Ibid., 1967, 50, 309. Helmer, R. G., Heath, R. L., Putnam, M., and Gipson, D. H., Ibid., 1967, 57, 46. Barnes, V., IEEE Trans. Nucl. Sci., 1968, NS 15-3, 437. Yule, H. P., i n DeVoe, J. R., Editor, “Modern Trends in Activation Analysis,” N.B.S. Special Publication 312, 1969, pp. 1155-1204. Covell, D. F., Analyt. Chem., 1959, 31, 1785. Savitzky, A., and Golay, M. J. E., Ibid., 1964, 36, 1627. Gorman, D. J., UCRG199U3, June, 1970. Received January 27th, 1971 Accepted April 22nd, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600624

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

Analyst, September, 1971, Vol. 96, PP. 631-639 63 1 The Atomic-emission Spectroscopy of Rhenium in the Nitrous Oxide - Acetylene Flame BY R. SMITH AND A. E. LAWSON (Imperial Chemical Industries Ltd., Petrochemicals Division, Billingham, Teesside) Rhenium can be determined by atomic-emission spectroscopy by use of a pre-mixed nitrous oxide - acetylene flame supported on a 6-cm slot burner. The limits of detection were 0.7 pg ml-l (346.1 nm) and 1.5 pg ml-l (488.9 nm) ; analytical working curves were linear for rhenium concentrations below 200 p g ml-1 a t both wavelengths. Spectral interference from palla- dium, nickel, rhodium, cobalt and large amounts of lanthanum or phosphate occurred a t 346.1 nm and from large amounts of aluminium a t 488.9 nm. A number of elements gave rise to chemical interference, but this was elimi- nated by the addition of sulphuric or phosphoric acid.A monochromator giving a spectral band pass of the order of 0-1 nm must be used for the determination, and background corrections must be made either by wave- length scanning or by measurement at the peak and at an adjacent wavelength. DURING the past decade, the use of atomic-emission spectroscopy in flames has been neglected compared with the rapid growth of atomic-absorption spectroscopy. However, in recent years there has been a renewal of interest in atomic-emission spectroscopy, mainly as a result of investigations with high-temperature pre-mixed flames and more sophisticated optical instrumentation than had been used previously. Pickett and Koirtyohann,l by using a nitrous oxide - acetylene flame supported on a conventional slot burner, have achieved detection limits for many elements that are comparable with those obtained by atomic-absorption spectroscopy.The advantages of pre-mixed, high-temperature flames and monochromators of moderately high dispersion (band passes of the order of 0-1 nm) have been emphasised by these authors,lS2 and atomic-emission spectroscopy should shortly emerge as a comple- mentary technique to atomic-absorption spectroscopy, particularly for the non-routine analysis of uncommon elements. Atomic emission has been observed from rhenium in the oxygen - hydrogen turbulent flame,3 the inner cone of the air - acetylene flame,4 and in the pre-mixed oxygen - acetylene flame.5f6p7 Schrenk, Lehman and Neufelds have carried out an investigation into the atomic- absorption characteristics of this element in oxygen - acetylene flames, but almost no informa- tion is available concerning the formation of rhenium molecular species in flames.Pickett and Koirtyohannl have reported a detection limit of 0.2 pg ml-1 of rhenium by atomic-emission spectroscopy with a nitrous -oxide - acetylene flame; this is substantially lower than could be obtained by atomic-absorption spectroscopy. However, no systematic examination of this element has yet been carried out. APPARATUS- Results were obtained by using a Techtron AA-5 atomic-absorption spectrophotometer with a mechanical chopper accessory for atomic-emission measurements. The 0.5-m Ebert monochromator with an f 10 aperture and a 50 x 50-mm2 grating with 638 lines mm-1 gave a reciprocal linear dispersion at the exit slit of 3.3 nm mm-l in the first order.An HTV R213 (Hamamatsu TV Co) photomultiplier was used as a detector, operating at 325 to 800 V. The signal was fed to a 285-Hz phase-sensitive amplifier, then to a potentiometric chart recorder (Servoscribe RE511, Smiths Industries Ltd.). A geared motor drive was made for the monochromator, which enabled wavelength scanning to be carried out at 0.6 nm min-l. Gas flow meters on the spectrophotometer were calibrated at 20 "C for the gases used against a wet EXPERIMENTAL A conventional 6-cm slot, grooved Techtron burner was used throughout. 0 SAC and the authors.632 gas flow meter that had recently been standardised.are given in Table I. SMITH AND LAWSON : ATOMIC-EMISSION SPECTROSCOPY OF [Analyst, Vol. 96 Instrument operating parameters REAGENTS- Rhenium solutions were made by dissolving ammonium perrhenate (Johnson Matthey Chemicals Ltd.) in de-ionised distilled water. All other solutions were prepared from analyti- cal-reagent grade materials when commercially available, otherwise general-reagent grade chemicals were used. TABLE I INSTRUMENT OPERATING PARAMETERS FOR OPTIMUM SIGNAL-TO-NOISE RATIO Wavelength/nm . . .. .. .. .. Slit width/pm . . .. .. .. .. Spectral band passlnm . . .. .. .. Observation height/mm . . .. .. .. Acetylene flow-rate/l min-l* . . .. .. Nitrous oxide flow-rate/l min-l* .. .. Solution uptake ratelm1 min-1 . . .. .. Burner . . .. .. .. .. ..346.1 or 488-9 50 0.17 6-cm slot 10 3.9 7.0 7.2 * Fuel fiow-rates refer to volumes at 20 O C and atmospheric pressure. The height of the red reaction zone in the flame was about 12 to 13 mm. RESULTS AND DISCUSSION EXPERIMENTAL CONDITIONS- No emission was detected at wavelengths of 346-1 or 488.9 nm from 500 pg ml-1 rhenium solutions when using air - acetylene or nitrous oxide - hydrogen flames supported on the 6-cm slot burner and covering a wide range of fuel-to-oxidant ratios. A nitrogen-separated nitrous oxide - acetylene flame, supported on a circular slot b ~ r n e r , ~ gave a limit of detection of 8 pg ml-l of rhenium at 346.1 nm with a solution uptake rate of 7-2 ml min-1. As the detection limit was 14 pg ml-l without separation, it appeared that this poor performance was a result of the relative incompatibility of this burner configuration with the optics used, and that the separated nitrous oxide - acetylene flame, supported on a slot burner, might be worth investigating.Atomic emission was observed at the wavelengths listed in Table I1 when using a pre-mixed nitrous oxide - acetylene flame supported on a conventional atomic- absorption slot burner, All further work was carried out with this system. TABLE I1 LIMITS OF DETECTION AND INTENSITIES OF RHENIUM EMISSION LINES Wavelength/ for rhenium/ r & Detection limit Uncorrected intensity Corrected intensity nm pg ml-l Peak* Background Peak* Background 345.2 4 34 18 35 18 346.1 0.7 124 19 124 19 346.5 2.5 78 19 77 19 488.9 1-5 128 28 75 17 527.6 4 43 32 28 21 * The peak signal is measured from the level of the flame background.The background signal is measured from zero. The principal rhenium emission wavelengths are 346.1 and 488.9 nm, the former of which gives the lowest detection limits while the latter gives greater freedom from spectral inter- ference. Rhenium gives maximum emission intensity in a fuel-rich flame that lacks luminosity caused by incandescent carbon particles. The nitrous oxide flow-rate was fixed at 7-0 1 min-1, giving a solution uptake rate of 7.2 ml min-I. The acetylene flow-rate for maximum intensity was 4-0 1 min-1, giving a red reaction zone in the flame approximately 15 mm high. The optimum signal-to-noise ratio, however, was obtained by using a slightly lower acetylene flow-rate (3.9 1 min-I), giving a red zone 12 to 13 mm high when measurements were carriedSeptember, 19711 633 out at 346.1 nm.For measurements carried out at 488.9 nm either 3.9 1 min-1 or 4.0 1 min-1 gave equal signal-to-noise ratios. The emission signal and the background noise level were particularly dependent on the acetylene flow-rate. The flame background at the lower rhenium wavelengths was a continuum interposed between the strong NH band system at 336 nm and the strong CN violet system at 359 nm (Fig. 1). The background at 488.9 nm arises from a number of weak C, bands and the Q branch band head of the CH system at 489.0 nm. A difference of 0.1 1 min-l in the acetylene flow-rate can affect the signal-to-noise ratio by about 10 per cent. at both 346.1 nm and 488.9 nm.RHENIUM IN THE NITROUS OXIDE - ACETYLENE FLAME d 300 350 Wave lengt h/n m Fig. 1. Emission spectrum of rhenium at 345.2, 346.1 and 346.5 nm Maximum emission intensity was obtained if measurements were carried out within the red reaction zone of the nitrous oxide - acetylene flame (approximately 8 mm above the upper surface of the burner). However, the flame background was high and the noise level prohibitive in these circumstances. The maximum signal-to-noise ratio for 346.1 nm and 488-9 nm was found when measurements were carried out at a height 9 to 11 mm above the burner. The optical axis of the instrument is then just below the upper tip of the red reaction zone of the flame. A slit width of 50pm (spectral band pass 0.17nm) was found to give the maximum signal-to-noise ratio at 346.1 nm and 488.9 nm and represented a reasonable compromise between signal intensity and spectral resolution.A band pass of 0.8 nm resulted in a reduction of the signal-to-noise ratio by approximately 50 per cent. at the 20 pg ml-l concentration level.634 SMITH AND LAWSON : ATOMIC-EMISSION SPECTROSCOPY OF [Analyst, vol. 96 SENSITIVITY- Limits of detection, obtained under the conditions listed in Table I, are given in Table 11. The detection limit is defined as the concentration of rhenium in aqueous solution that gives a signal-to-noise ratio of unity, the noise being the peak-to-peak noise of the flame background when de-ionised water is aspirated. Four pairs of measurements of background and test solution intensity were taken for each detection limit given by using the maximum damping available on the instrument and carrying out the measurements with a rhenium concentration no greater than five times the limit of detection.Intensity measurements of the rhenium emission lines (in arbitrary units) were also made (Table 11) and were corrected for the spectral response factors of the monochromator and photomultiplier, obtained from the instrument manufacturer's data. Limits of detection for rhenium by atomic absorption have been reported by SlavinlO as being 1.5 pg ml-l and by Shrenk et aZ.* as 1 to 2 pg ml-l. A similar detection limit should be obtainable on the instrument used for this investigation. In contrast, the detection limit by atomic emission is unquestionably lower: 0.7 pg ml-l reported in this paper and 0.2 pg ml-1 by Pickett and Koirtyohann.lS2 Detection limits by atomic emission can, in principle, be reduced by the use of more sophisticated instrumentation.In particular, the use of a high-dispersion monochromator, narrower band width amplifier and more sensitive detector would be expected to give sub- stantial improvements in signal-to-noise ratios. Corresponding reductions in atomic-absorp- tion detection limits could not easily be obtained.2 Analytical working curves were found to be linear up to 200 pg ml-l of rhenium and only slightly curved up to 1 mg ml-1 at both 346.1 nm and 488.9 nm. SPECTRAL INTERFERENCES- If the practice of carrying out measurements at the emission line wavelength only is followed, it is not possible to correct for increases in signal caused by molecular band emission, Fig.2. Spectral line interferences a t the Re 346.1 nm wavelength: (a), de-ionised water only; (b), 20 pg ml-l rhenium; (c), 20pg ml-l rhenium, 200pg ml-1 molybdenum; (d), 20 pg ml-l rhenium, 200 pg ml-l cobalt ; (e) 20 pg ml-1 rhenium, 200 pg ml-l rhodium; (f), 20 pg ml-1 rhenium, 200 pg ml-1 palladium; and (g), 20 pg ml-1 rhenium, 200 pg ml-1 nickelSeptember, 19711 RHENIUM IN THE NITROUS OXIDE - ACETYLENE FLAME 635 light scatter within the monochromator, or partially resolved atomic-line emission from concomitant elements. If the analyst scans the characteristic emission line, or makes measure- ments a t an adjacent "background" wavelength, continuum interference from broad molecular bands or scattered light can be eliminated entirelyq2 Further, the presence of partially resolved atomic-line interference is immediately indicated and it is frequently possible to make useful measurements if the lines are moderately resolved.Alternatively, measurements can be made at another wavelength. Solutions of the following metals (rhenium free) were aspirated into the flame and scans made over the wavelength ranges 345.6 nm to 346.6 nm and 488.4 nm to 489.6 nm; con- centrations (Jug ml-l) are given in brackets:Ag (5000), A1 (10 000), As (5000), B (lOOO), Ba (lOOO), Ca (lOOO), Cd (lOOO), Co (lOOO), Cr (lOOO), Cs (lOOOO), Fe (lOOO), In (lOOO), I r (lOOO), K (10 000), La (10 000), Li (10 000), Mg (10 000), Mn (lOOO), Mo (500), Na (10 OOO), Ni (lOOO), 0 s (lOOO), Pb (lOOOO), Pd (lOOO), Pt (2000), Rh (1250), Ru (loo), Sb (lOOO), Si (2000), Sn (SOOO), V (lOOO), W (500) and Zn (5000).Also, the following acids were examined after one 100-fold dilution of the concentrated acid : acetic, hydrochloric, hydrofluoric, nitric, perchloric, phosphoric and sulphuric. The amplification was adjusted so that 30 Jug ml-l of rhenium gave a full-scale recorder deflection. The flame background spectrum obtained on aspirating de-ionised water is a featureless horizontal line between 345.6 nm and 346.6 nm. Important spectral lines in this range occur at the cobalt 346.28 nm, cobalt 345.52 nm, palladium 346.08 nm, nickel 346.17 nm, nickel 345.85 nm and rhodium 346.20 nm wavelengths; in addition, a line at about 345-5 &- 0.2 nm caused by molybdenum was encountered.This line was approximately one twentieth of the intensity of the rhenium 346-1 nm line (for equal concentrations by weight) and could not be readily identified as a genuine molybdenum atomic line, MOO molecular band or even as a probable impurity. However, the line was well resolved from the rhenium emission at 346.1 nm and caused no interference. Fig. 2 shows the interference from atomic lines on the rhenium 346.1 nm atomic emission. Palladium 346.08 nm line is the most serious spectral interference as it is indistinguishable from the 346-05 nm rhenium line even on scanning. However, it is almost exactly one tenth of the intensity of the rhenium line. 20 (Ug Water 20pgml-' Re ~ ~ ~ p g r n i - ~ La + ml-' Re I !Opgml-' Re + 1% H3P04 Fig.3. Weak spectral inter- ferences a t the Re 346.1 nm wave- length: (a), de-ionised water only; ( b ) , 20 pg ml-l rhenium; ( G ) , 20 pg ml-1 rhenium, lo4 pg ml-1 lanthanum; and (d), 20 pg ml-l rhenium, 1 per cent. phosphoric acid636 SMITH AND LAWSON: ATOMIC-EMISSION SPECTROSCOPY OF [A?ZdySt, VOl. 96 The two nickel lines at 346.17 nm and 345.85 nm were completely resolved from one another but were of comparable intensity to the rhenium emission and could not be resolved from this line. It is essential to use the rhenium 488.9 nm line if nickel is present in any sample. Partial resolution of the rhenium 346-1 nm and rhodium 346.2 nm lines was possible and was sufficient to indicate the presence of rhodium in a sample and the need to use an alternative rhenium line.The rhodium line is approximately one tenth as intense as the rhenium emission. Resolution of the cobalt 346.3 nm and rhenium 346.1 nm lines was almost complete and the 345.5 nm line was completely resolved (Fig. 2 ) . Useful analyses can therefore be obtained even in the presence of a considerable excess of cobalt, especially if narrow slits are used. Lanthanum exhibits a pair of weak lines (Fig. 3), or perhaps part of a band system, which is about two hundred times less intense than the emission from an equal concentration of rhenium. Phosphoric acid gives rise to a series of weak PO molecular bands in this region also (Fig. 3). The latter are so weak as to cause concern only when large amounts of phos- phoric acid are present in the analyte and in these circumstances this would usually be known and an approximately equal amount of the acid could be added to the standards.Several elements gave rise to slight increases in the general background level of the spectrum (e.g., 10 mg ml-l amounts of alkali metals) but these could easily be overcome by scanning and would not present difficulties, even in the determination of less than 5 ,ug ml-l of rhenium. 20pg ml-’ Re I O4 pg m I-’ A1 + 20pg ml-’ Re Fig. 4. Emission characteristics of con- comitant elements a t the Re 488.9 nm wave- length: (a), de-ionised water only; (!), 20 pg ml-1 rhenium; (c), lo4 pg ml-1 alumm- ium; (d), 20 p g ml-1 rhenium, lo4 pg ml-l aluminium; and (e), lo3 p g ml-l vanadium Between 488.4 nm and 489.6 nm, the flame background spectrum consists of two small peaks on either side of the rhenium 488.9 nm wavelength.Spectral interferences are prac- tically insignificant at this wavelength ; only large amounts of concomitant elements cause any difficulty. Aluminium (10 mg ml-l) gives a substantially increased background (Fig. 4) and a peak height (equivalent to that of about 8 pg ml-1 of rhenium) that is probably part of the A10 band system, with band head at 488.84 nm. Line interference from second-orderSeptember, 19711 RHENIUM IN THE NITROUS OXIDE - ACETYLESE FLAME 637 spectral interferences is unlikely as no strong lines are present in the 244nm region. Of the elements listed, vanadium (1 mg mV), lanthanum (10 mg ml-l) and molybdenum (10 mg ml-l) gave unusually high continuum backgrounds, which do not present problems a t the 20 pg ml-l rhenium level.Other elements give smaller over-all increases in the level of the background emission. Wavelength tables indicate that gadolinium may give rise to spectral interference at 489.2nm from the GdO molecular band head, but as this band system is degraded to the red end of the spectrum, the band may be completely resolved from the rhenium emission. A detailed study of rare earth spectral interferences was, however, considered to be outside the scope of this work. CHEMICAL INTERFERENCES- The determination of rhenium by atomic absorption in oxygen - acetylene flames8 was shown to be prone to chemical interference (ie., inter-element effects caused by the formation of involatile species in the flame), and the study carried out here indicates similar behaviour in the pre-mixed nitrous oxide - acetylene flame.The effects of common acids (addition of 1 per cent. of the concentrated acid) on the emission of 20 pg ml-l of rhenium were first investigated. Phosphoric and hydrochloric acids gave decreased signals of 76 and 88 per cent., respectively, relative to aqueous am- monium perrhenate solution. Acetic, hydrofluoric, nitric and sulphuric acids gave no inter- ference. During further studies on the interference effects of cations, measurements were carried out in 1 per cent. nitric acid solution in an attempt to overcome the effects of con- comitant anions. This was successful when tests were made with sodium chloride and sodium nitrate, neither of which caused interference in the presence of 1 per cent.nitric acid. TABLE I11 CHEMICAL INTERFERENCES IN THE DETERMINATION OF RHENIUM Added element* Re only . . Ag . . A1 . . As . . B .. Ba . . Ca . . co * . Cr . . cs .. c u .. Fe . . In . . Ir . . K .. Li . . Mg . . Mn . . Mo . . Na . . Ni . . 0 s . . Pb .. Pd . . Pt . . Rh .. Ru . . Sb .. Si . . Sn . . Sr .. v .. w .. Zn . . .. * . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 1 per cent. of nitric acidt 100 100 70 102 84 0 13 89 85 100 101 99 101 101 100 55 45 83 99 103 93 101 96 99 100 100 100 46 101 100 2 100 56 101 1 per cent. of hydrofluoric acidt 100 75 100 18 94 89 79 - - - - 1 per cent.of sulphuric acidt 100 100 100 98 100 100 103 - - - - 102 100 99 100 100 100 - - - - 97 101 - 1 per cent. of phosphoric acidt 63 101 100 99 99 100 98 - - - - 101 101 99 99 98 100 - - - - 101 100 - - * Effects of 200 pg ml-l of added element were investigated on 20 pg ml-l of rhenium. t 1 per cent. of concentrated acid.63 8 SMITH AND LAWSON : ATOMIC-EMISSION SPECTROSCOPY OF [Autdyst, Vol. 96 Cationic interferences were investigated by using 200 pg ml-l of the foreign element, 20 pg ml-1 of rhenium and 1 per cent. nitric acid. Blanks containing no rhenium were also measured and the blank value was subtracted from the total signal. In practice, blanks of this kind would not be available, but the blank value could be obtained at an adjacent background wavelength or by scanning.The results of this study are summarised in Table 111; comparable effects were obtained at 346-1 nm and 488-9 nm. Where spectral interferences were known to occur, measurements were carried out at an alternative wavelength. A definite interference was assumed to be present when the intensity of the rhenium emission was changed by an amount greater than &4 per cent. relative to the pure solution. The presence of lead represented the only borderline case (96 per cent. signal intensity) and was assumed to give rise to a genuine interference effect. It must be stressed that this type of interference will occur in both atomic-absorption and atomic-emission spectroscopy, and that the magnitude of the effect will be similar for measurements carried out by the two techniques on any particular instrument.The results given in Table I11 are similar to those obtained by Schrenk et aLS by using atomic absorption in turbulent oxygen - acetylene flames. Calcium was founds to reduce the rhenium absorption seriously, but the effect was not quite so marked as indicated in Table IV; the interference from manganese, however, was much more serious in the oxygen - acetylene flame. In addition, Schrenk et aL8 reported interference from iron, potassium and molybdenum, none of which was encountered in this work. TABLE IV EFFECTS OF TYPICAL RELEASING AGENTS ON THE INTERFERENCE OF CALCIUM (200 pg ml-l) ON RHENIUM (20 pg ml-l) Rhenium Releasing agent emission intensity* Background intensityt None .. .. .. .. .. 10 per cent.EDTA (at pH 11) . . 0.1 per cent. Si (as Na,SiO,) . . 0.1 per cent. A1 (as Al,SO,) 10 per cent. La (as LaCI,) . . .. 5 per cent. H,PO, . . .. .. 5 per cent. HF . . .. .. 5 per cent. H,SO, . . .. .. .. 11 45 12 102 50 116 109 100 0 43 11 38 120 17 0 0 * After correction for background from releasing agent and calcium. t The emission intensity of the releasing agent a t 346.1 nm. As the alkaline earth elements represent the most significant chemical interferences efforts were concentrated on the elimination of effects arising from the presence of these metals. Various releasing agents were examined; in particular, those which were known to form involatile alkali-metal species were chosen (e.g. , aluminium, silicate, sulphuric acid and phosphoric acid) as well as some well known releasing agents for the alkali metals (e.g., am- moniacal solution of EDTA, lanthanum).The effects of the releasing agents are summarised in Table IV. Apart from the alkaline EDTA solution, almost all of the compounds examined considerably suppressed the calcium interference. However, many of the releasing agents (e.g. , aluminium, lanthanum, silicate) gave rise to an inconveniently high background emission. This could be compensated for fairly easily but such a compensation was unnecessary as dilute hydrofluoric, sulphuric and phosphoric acids were equally effective as releasing agents for calcium and did not give rise to increased flame background. The dilute solutions of mineral acids containing 1 per cent. of concentrated acid were then investigated for use as general releasing agents for all of the elements that interfere in the rhenium determination.These results are also included in Table 111. Hydrofluoric acid, although acting as a releasing agent for the effect of calcium on rhenium, allowed many other elements to interfere, including strontium and barium. The interference of boron (as borate) was removed and the effects of aluminium and magnesium were reduced. Sulphuric and phosphoric acids (added as 1 per cent. of concentrated acid), however, acted as releasing agents for all of the interfering metals examined and are the most convenient releasing agents for general use. With phosphoric acid the rhenium emission is considerably depressed (Table 111) but this effect is overcome when small amounts of a third element are present.September, 19711 RHENIUM IN THE NITROUS OXIDE - ACETYLENE FLAME 639 At the 20 ,ug ml-l of rhenium level, the excess of the third element should be about three times (expressed on a molar concentration basis) that of rhenium.However, this appears to be a fairly complex ternary system and for analytical purposes it is sufficient to be aware of its existence and to take the necessary precautions. The small spectral interference of phosphoric acid (Fig. 3) at 346.1 nm is also a disadvantage. Sulphuric acid is the recom- mended releasing agent, but the choice may well be made on the basis of the most suitable acid for the dissolution of the sample. CONCLUSIONS The determination of rhenium can be conveniently carried out by atomic-emission spectroscopy in the pre-mixed nitrous oxide - acetylene flame if a monochromator of moderate dispersion is available. Measurements carried out at the recommended wavelength of 346.1 nm are subject to spectral interference from cobalt, nickel, palladium and rhodium, as well as from very large amounts of lanthanum or phosphate.The alternative wavelength of 488.9 nm is subject to spectral interference only from large concentrations of aluminium but the emission detection limits are higher. The use of scanning or measurement at the wavelength peak and an adjacent wavelength was found to be necessary to compensate for variations in the background level of the sample relative to standards. Chemical interferences are many, but can easily be prevented by addition of sulphuric or phosphoric acid. The same chemical interferences will be present in the corresponding determination by atomic-absorption spectroscopy, although the emission method gives lower detection limits. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. REFERENCES Pickett, E. E., and Koirtyohann, S. R., Spectrochim. Acta, 1968, 23B, 235. -- , Analyt. Chem., 1969, 41 (14), 28A. Gilblrt, P. T., Beckman Instrum. Bull., 1961, 7538. Mavrodineanu, R., and Boiteux, H., “Flame Spectroscopy,” John Wiley & Sons, New York, 1965. Fassel, V. A,, Myers, R. B., and Kniseley, R. N., Spectrochim. Acta, 1963, 19, 1187. Kniseley, R. N., D’Silva, A. P., and Fassel, V. A., Analyt. Chem., 1963, 35, 910. Fassel, V. A., and Golightly, D. W., Ibid., 1967, 39, 466. Schrenk, W. G., Lehman, D. A,, and Neufeld, L., Appl. Spectrosc., 1966, 20, 389. Kirkbright, G. F., Sargent, M., and West, T. S., Talanta, 1969, 16, 245. Slavin, W., “Atomic Absorption Spectroscopy, ” Interscience Publishers Inc., New York, 1968. Received January 5th, 1971 Accepted April 16th, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600631

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

640 Analyst, September, 1971, Vol. 96, $p. 640-655 A Method for the Chemical Analysis of Magnesites and Dolomites* BY H. BENNETT AND R. A. REED (The Bvitish Ceramic Research Association, Queens Road, Penkhull, Stoke-on-Trent, S T 4 7LQ) This paper includes a detailed description of a method for the analysis of magnesites and dolomites; the method has been accepted (in principle) by the British Standards Institution as a standard method. Determinations include SiO, (gravimetric) , TiO,, Fe,O,, Cr,O,, MnO (colorimetric) and A1,0,, CaO and MgO (complexometric) . Notes on the development of the method and tables of co-operative results obtained by the Refractories Working Group of the Analysis Committee are included. THIS method, a schematic diagram of which is shown in Fig. 1, has been developed by the Refractories Working Group of the British Ceramic Research Association Analysis Committee, and will, after approval by the Analysis Committee, be submitted to B.S.I.for acceptance as a standard method. Co-operative results obtained by the Working Group on two magnesites and two dolomites are given so that the capabilities of the method can be assessed. The method has been under development for several years and in this time the Working Group have attempted a considerable number of procedures, meeting with little success in many instances. This work has not been reported in detail, but some indication of the Group's activities has been given in the form of notes on the development of the method. METHOD The method covers the determination of loss on ignition and of SiO,, Fez?,, TiO,, MnO, Cr203, Al,O,, CaO and MgO in magnesites and dolomites suitable for industnal refractories.In general, the method is not sufficiently sensitive for the analysis of samples of special high purity MgO. PRINCIPLE OF THE METHOD LOSS ON IGNITION- The loss on ignition is determined at 1025 & 25 "C. DETERMINATION OF OXIDES OF SILICON, IRON(III), TITANIUM(IV), MANGANESE(II), CHROMIUM(III), ALUMINIUM(III), CALCIUM AND MAGNESIUM- The sample is decomposed with hydrochloric acid and the silica separated by coagulation and filtration. The crucible containing the silica is ignited and weighed before and after treatment with hydrofluoric and sulphuric acids. The silica remaining in the filtrate is subsequently determined by a spectrophotometric method based on the formation of molyb- denum blue in a portion of the solution used for the determination of Fe203, TiO,, MnO, Cr203, A1,0,, CaO and MgO.The residue from the silica is fused with sodium carbonate and boric acid and the cold melt dissolved in the filtrate from the silica. After dilution to a standard volume aliquots are used for the determination of iron(II1) oxide spectrophotometrically with 1,lO-phenan- throline, titaniuni(1V) oxide with hydrogen peroxide, manganese(I1) oxide with potassium periodate and chromium(II1) oxide with diphenylcarbazide for very low chromium contents and as the chromium(II1) - EDTA complex for materials containing more than about 0.1 per cent. of chromium(II1) oxide. For the aluminium(II1) oxide determination, oxides of elements in the ammonia group of the scheme for systematic analysis in a measured volume of the solution are precipitated with ammonia solution t o remove most of the magnesium * A Report prepared by the Refractories Working Group of the Analysis Committee of the British Ceramic Research Association. 0 SAC and the authors.Sample 1 (19) Colour Sample 2 (59) Main analysis Dissolve in HCI I I I Coagulate with polyethylene oxide Loss on ignition Precipit? Filjer I I I 1 I I Molybdenum I Hydrogen blve 1 ,lO-Phen~nthroline peroxide F i It rate t o volume I Potassium periodate < 0.i % > 0.1% c r 2 0 3 cr2 0 3 I Diphenyl carbazide EDTA R2O3 Sepirate from MgO (Ammon la precipitation) I Fez03 and Ti02 (Solvent extraction) EDTA titrimetric I I A1203 Fig.1. Schematic diagram of the analysis of magnesites and dolomites Separate Fe, Mn (Solvent Separate extraction) M P (precipitation) DCTA titrimetric I Titrimetric I CaO MgO W M w M M U642 [Analyst, Vol. 96 oxide. The precipitate is dissolved in hydrochloric acid and the iron(II1) and titanium(1V) ions are removed by a cupferron - chloroform solvent extraction. The determination is completed titrimetrically with ethylenediaminetetraacetic acid (EDTA) and zinc, with dithizone as indicator. This titration includes the amount of EDTA consumed by the Cr203, which must be allowed for. The use of diaminocyclohexanetetraacetic acid (DCTA) in place of EDTA enables A1203 to be determined directly.1 Lime is determined on a measured volume of the solution to which triethanolamine is added to complex interfering elements.A known volume of standard 1,2-bis-(2-aminoethoxy)ethanetetraacetic acid (EGTA), sufficient to complex all the calcium, is added, and the magnesium is then precipitated with potassium hydroxide. The excess of EGTA is titrated with a calcium solution, Calcein being used as indicator. Magnesium oxide is determined titrimetrically on a measured volume of the solution after removal of iron(III), manganese(I1) and most of the titanium(1V) ions by a sodium diethyl- dithiocarbamate - chloroform solvent extraction. Aluminium(II1) and any remaining titan- ium ions are complexed with triethanolamine and the magnesium is titrated with DCTA in a strongly ammoniacal solution containing ammonium chloride.If more than 1 per cent. of Al,Op is present, both it and TiO, are removed by a buffered cupferron - chloroform solvent extraction. The titration also includes the titration for lime, which must be allowed for. BENNETT AND REED: A METHOD FOR THE CHEMICAL DETERMINATION OF ALKALI-METAL OXIDES- If a simple filter flame photometer is used the alkali-metal oxides can be determined by the method described by Eardley and Reed2 under the heading “High lime materials.” REAGENTS- Unless otherwise stated, all reagents must be of analytical-reagent grade when available, and distilled water must be used throughout the analysis. The following reagents are required. Ammonium chloride. Boric acid. Calcein indicator-Mix, by grinding together, 0.2 g of Calcein, 0.12 g of thymolphthalein Potassium periodate.Silica-A quartz or sand of known purity, not less than 99 per cent. Sodium carbonate, anhydrous. Solochrome black 6B (also known as Eriochrome blue-black B)-Mix, by grinding together, Thy molphtlzalein. Acetic acid, glacial. Acetone. Ammonia solution, sp.gr. 0.88. Ammonium acetate solution, approximately 10 per cent.-Dilute 140 ml of glacial acetic acid to 1700 ml with water and add 140 ml of the ammonia solution. Mix, cool and adjust to pH 6.0 to 6.5. Ammonium acetate buffer (for aluminium(III) oxide determinations)-Add 120 ml of glacial acetic acid to 500 ml of water, followed by 74 ml of the ammonia solution. Cool the mixture, dilute it to 1 litre and mix. Ammonium acetate buffer (for magnesima oxide determinations)-Dilute 5 ml of the am- monia solution to 100 ml with distilled water and add 30 ml of glacial acetic acid.Adjust the solution to pH 3.8 by using a pH meter and then dilute to 200 ml with distilled water. Ammonium cerium(IV) nitrate solution, 1 eer cent. w/v-Dissolve 2.5 g of ammonium cerium(1V) nitrate in about 200 ml of water, cautiously add 7 ml of sulphuric acid (sp.gr. 1-84), cool, dilute to 250ml and mix. Ammonium molybdate solution, 8 per cent. w/w-Discard after 4 weeks, or earlier if any appreciable deposit is observed. Ammonium nitrate solution, approximately 1 per cent.-Dilute 10 ml of nitric acid (sp.gr. 1.42) to about 200 ml. Add ammonia solution (1 + 1) until the solution is slightly alkaline to bromophenol blue, then cool and dilute it to 1 litre. Byornophenol blue solution-Grind and dissolve 0.1 g of bromophenol blue in 1.5 ml of sodium hydroxide solution (0.4 per cent.w/v), dilute the solution to 100 ml with water and mix. and 20 g of potassium chloride. 0.5 g of Solochrome black 6B and 20 g of sodium chloride.September, 19711 ANALYSIS OF MAGNESITES AND DOLOMITES 643 Chloroform, B.P. grade. Cupferron, 6 per cent. w/v-This solution must be freshly prepared. Saturated 2,4-dinitrophenol solution-Dissolve 0-1 g of 2,4-dinitrophenol in 100 ml of hot Diphenylcarbaxide, 1 per cent. w/v solution-Dissolve 0.1 g of diphenylcarbazide in 10 ml Dithixone solution-Dissolve 0.025g of dithizone in 100ml of 95 per cent. ethanol. Ethanol, 95 per cent.-Industrial methylated spirit, 95 per cent, Ethylenediaminetetraacetic acid, disodium salt dihydrate (EDTA)-A 5 per cent.wlv solution. Hydrochloric acid, sp.gr. 1.18. HydroJ7uoric acid, 40 per cent. w/w. Hydrogen peroxide, 20 volume. Hydroxylammonium chloride, 10 per cent. wlv. Mag$ok, approximately 2 per cent. w/v-The resin is a thick viscous liquid and it is there- fore most convenient to transfer a drop to a beaker and weigh the amount taken. Sufficient water is then added to make up a solution approximately 2 per cent. w/v. The resin is available from Ridsdale & Co. Ltd., Newham Hall, Newby, Middlesbrough. Magnesium sulphate solution-Dissolve 6.1 1 g of magnesium sulphate heptahydrate in water, filter and dilute to 500 ml (50 ml = about 0.1 g of MgO). Naphthol green B, 0.1 per cent. wlv. Nitric acid, sp.gr. 1.42. 1,lO-Phenanthroline hydrate, 1 per cent.w/v-Prepare enough solution for immediate use at a concentration of 0.1 g per 10 ml of acetic acid (1 + 1). Phosphoric acid, sp.gr. 1-75. Polyethylene oxide, 0.25 per cent. w/v-Add 0.5 g of polyethylene oxide slowly to 200 ml of water with stirring, preferably on a mechanical stirrer, until dissolved. Discard this solution after 2 weeks. Union Carbide Polyox resins WSR-35, WSR-N-80, WSR-205, WSR-N-750 and WSR-N-3000 are suitable as sources of polyethylene oxide. water, allow to cool and filter. of acetone. This solution must be freshly prepared. Potassium hydroxide solution, 25 per cent. w/v. Sodium axide solution, 2 per cent. w/v. Sodium diethyldithiocarbamate, 10 per cent. w/v-This solution must be freshly prepared. Sodium sulphite, 5 per cent. w/v-This solution must be freshly prepared.Tin(I1) chloride, 1 per cent. w/v solution-Dissolve, by warming, 1 g of tin(I1) chloride in 1.5 ml of the hydrochloric acid. Cool the solution and dilute it to 100 ml. This solution should not be kept for more than 24 hours. Sulphuric acid, sp.gr. 1.84. Sul~hurous acid-Saturate 250 ml of water with sulphur dioxide. Tyiethanolamine. STANDARD SOLUTIONS- Calcium solution, 0.05 M-Dissolve 5.0045 g of calcium carbonate (dried at 150 "C) in a slight excess of hydrochloric acid (1 + 4). Boil the solution to expel carbon dioxide, cool it and dilute to 1 litre. Chromium solution A (0.5 mg ml-l of Cr,O,)-Dissolve 0.9677 g of potassium dichromate (dried at 150 "C) in water and dilute to 1 litre. Chzromium solution B (0.025 mg mZ-l of Cr,O,)-Dilute 25 ml of the chromium solution A to 500ml.Solution of 1,2-diaminocyclohexanetetraacetic acid (DCTA) , about 0.05 M-Dissolve 18.2175 g of DCTA in 500 ml of water by the progressive addition of the minimum amount of potassium hydroxide solution (25 per cent. wlv). Dilute to 1 litre. Standardise this solution against standard magnesium solution. Ethylenediaminetetraacetic acid, disodium salt dihydrate (EDTA) , 0.05 M-Dissolve 18.6125 g of the salt in warm water, filter if necessary, cool and dilute to 1 litre. Standardise the resulting solution against the standard zinc solution, dithizone being used as indicator. 1,2-bis(aminoethoxy) ethanetetraacetic acid (EGTA) , about 0.05 M-Dissolve 19-0174 g in 500 ml of water by the progressive addition of the minimum amount of potassium hydroxide solution (25 per cent.wlv). Standardise the solution against standard calcium solution. Dilute to 1 litre.644 BENXETT AND REED: A METHOD FOR THE CHEMICAL [Analyst, Vol. 96 Iyon solution A (0.1 mg ml-l of Fe,O,)-Dissolve 0-0699 g of the oxide-free metal wire in a slight excess of hydrochloric acid (1 -k 9); add 5 ml of hydrogen peroxide and boil for 15 minutes. Cool the solution and dilute it to 1 litre. Iron solution B (0.01 mg ml-l of Fe,O,)-Dilute 50 ml of the iron solution A to 500 ml. Magnesium solution (5 mg ml-1 of Mg0)-Dissolve 3.016 g of the oxide-free metal in a slight excess of hydrochloric acid (1 + 9), cool and dilute to 1 litre. Manganese solution A (0.1 mg ml-1 of Mn0)-Dilute the calculated volume of previously standardised (about 0.1 N) potassium permanganate to 1 litre.(With exactly 0.1 N potassium permanganate, 70.5 ml are required.) Manganese solution B (0.01 mg ml-1 of Mn0)-Dilute 50 ml of the manganese solution A to 500 ml. Potassium perrnanganate solution, about 0.1 N-Dissolve about 3-2 g of potassium per- manganate in 1 litre of water to obtain an approximately 0.1 N solution; boil for 5 minutes, filter through a sintered-glass crucible of porosity grade No. 4 and cool. Standardise the solution against sodium oxalate and store it in an amber-glass bottle. Silica solution A (approximately 0.5 mg ml-l of SiOJ-Fuse 06000 g of pure silica (>99 per cent. purity) with 5 g of anhydrous sodium carbonate in a platinum crucible. Cool the mixture and dissolve it in water in a polythene beaker.Cool again and dilute the solution to 1 litre. This solution will remain stable for at least 6 months. Silica solution B (approximately 0.020 mg ml-1 of 50,)-Dilute 20 ml of the silica solution A to 500ml. Titanium solution A (1.0 mg ml-1 of Ti0,)-Ignite some pure TiO, and then fuse 1.000 g with l o g of potassium pyrosulphate. Allow to cool and dissolve the melt, at a low tem- perature to prevent hydrolysis, in 200 ml of water to which 20 ml of the sulphuric acid have been cautiously added. Cool again and dilute the solution to 1 litre. Titanium solution B (0.04 mg ml-1 of Ti0,)-Dilute 20 ml of the titanium solution A to 500ml. Zinc solution, 0.05 M-Dissolve 3.2685 g of the oxide-free metal in 50 ml of hydrochloric acid (1 + 4), cover the beaker with a watch-glass and allow to stand overnight on a steam- bath.Cool and dilute to 1 litre (1 ml = 2-55 mg of Al,O,). STANDARDISATIONS- EGTA against calcium-Transfer 50 ml of the magnesium sulphate solution to a 250-ml calibrated flask and add 50.0 ml of the EGTA solution. Dilute to 150 ml, add potassium hydroxide solution (25 per cent. w/v) until no further precipitation occurs and then add 10 ml in excess, followed by 10 ml of Magflok solution. Dilute the mixture to 250 ml, shake and allow it to stand for about 10 minutes to settle. Filter it through a dry 125-mm Whatman No. 541 filter-paper into a dry beaker. Pipette 200 ml of the filtrate into a 500-ml conical flask and add 15 ml of potassium hydroxide solution. Titrate with standard calcium solution ( O - O ~ M ) , with screened Calcein as indicator, to the first appearance of green fluorescence (1 ml of 0.05 M EGTA = 2.804 mg of CaO).DCTA against magnesium-Transfer 50.0 ml of the standard magnesium solution to a 500-ml conical flask. Add 100 ml of the DCTA solution followed by 2 g of ammonium chloride and 25 ml of ammonia solution. Titrate with the DCTA solution, with Solochrome black 6B as indicator, from red, through purple, until the last change in colour to a clear ice blue (1 ml of 0-05 M DCTA = 2.016 mg of MgO). EDTA against zinc-Transfer 50.0ml of the EDTA solution to a 500-ml conical flask and add 5 to 6 drops of hydrochloric acid. Add a few drops of bromophenol blue solution and then add ammonium acetate buffer solution (for aluminium(II1) oxide determination) until the indicator turns blue, followed by an excess of 10 ml.Add a volume of the ethanol equal to the total volume of existing solution, followed by 1 to 2 ml of dithizone solution, and titrate with the standard zinc solution from green to the first appearance of a permanent pink colour (1 ml of 0.05 M EDTA = 2.55 mg of Al,O,). BLANK DETERMINATIONS- Blank determinations should be carried out on all reagents in accordance with the general scheme of analysis. When carrying out the blank determination for lime, it is necessary to add 50 ml of the magnesium sulphate solution before the addition of the standard EGTA solution.September, 19711 ANALYSIS OF MAGSESITES AND DOLOMITES 645 PREPARATION OF SAMPLE- The sample prepared for analysis should be ground to pass a 125-pm B.S.test sieve. A non-metallic (e.g., 120-mesh nylon bolting cloth) sieve is preferable. Many of these materials are comparatively soft and may be ground in an agate mortar without appreciable contamination. In a few instances the material may be sufficiently hard to be contaminated and it is then necessary to prepare two samples, one ground in an iron mortar (most of the iron in the sample being removed with a magnet), and the other sample ground in an agate mortar. The first sample is used for the main analysis and the second for the determination of iron(II1) oxide, the results obtained on the first sample being corrected for iron contamination. EXPERIMENTAL DETAILS OF THE METHOD DETERMINATION OF LOSS ON IGNITION- Weigh 1.000 g of the finely ground sample, previously dried at 110 “C, into a platinum crucible.Cover the crucible almost completely with a lid and start the ignition over a low mushroom flame, slowly increasing the temperature to full heat over a period of about 20 minutes, after which the crucible is transferred to a furnace a t 1000 “C for 30 minutes. Remove the crucible from the furnace, cover it completely with a lid and weigh as soon as possible. DETERMINATION OF OXIDES OF SILICON, IRON(III), TITANIUM(IV), MANGANESE(II), CHROMIUM(III), Decomposition of the sample-Weigh 5.000 g of the finely ground sample, previously dried at 110 “C, and transfer to a 250-ml beaker. Add 25 ml of water and 40 ml of hydrochloric acid and cover the beaker with a clock-glass. Then transfer it to a sand-bath and boil the mixture for 30 minutes.Determination of the main silica-Allow the beaker and contents to cool and rinse the clock-glass with water into the solution. Add a Whatman accelerator tablet and stir to break up the pulp, then add, with stirring, 2 to 3 ml of polyethylene oxide solution and allow to stand for 5 minutes. Filter the solution through a 110-mm Whatman No. 42 filter- paper and transfer the silica to the filter with hot dilute hydrochloric acid (1 + 19), scrubbing the beaker with a rubber-tipped glass rod. Wash the precipitate six times with hot dilute hydrochloric acid (1 + 19) and then with hot water until it is free from chlorides. Reserve the filtrate and washings. Transfer the paper and precipitate to an ignited and weighed platinum crucible.Ignite them at a low temperature until the precipitate is free from carbonaceous matter and then heat in a muffle furnace at 1200 “C to constant weight, 30 minutes being normally sufficient. Moisten the contents of the cold crucible with water, add 5 drops of sulphuric acid (I + 1) and about 10 ml of hydrofluoric acid. Evaporate the mixture to dryness on a sand- bath in a fume cupboard. For the evaporation, the crucible and contents should be heated from below as the use of top heating alone, as with a radiant heater, may result in incomplete elimination of silica by the hydrofluoric acid. Heat the crucible and residue, cautiously at first, over a gas flame and finally for 5 minutes in a furnace at 1200 “C, cool and weigh. If the residue weighs more than 30mg repeat the treatment with sulphuric and hydrofluoric acids to ensure that all of the silica is removed.The difference between the two weights represents the “gravimetric” silica. PREPARATION OF A SOLUTION FOR THE DETERMINATION OF THE RESIDUAL OXIDES- Fuse the residue from the hydrofluoric acid treatment of the “gravimetric” silica with 2 g of sodium carbonate and 0.4g of boric acid. Place the crucible containing the cooled melt, plus the lid, into the filtrate and washings from the main silica. When the melt has dissolved remove the crucible and lid, scrubbing them with a rubber-tipped glass rod. Cool, dilute the solution to 500 ml in a calibrated flask and mix. This solution is referred to as the stock solution. Determination of residual siZica-Transfer 5.0 ml of the stock solution to a 100-ml cali- brated flask (A) and add 15ml of water.Add 2 drops of 2,4-dinitrophenol indicator and dilute ammonia solution (1 + 1) dropwise until the indicator turns yellow (note the amount of ammonia solution used), then add 5 ml of dilute hydrochloric acid (1 + 4). ALUMINIUM(III) , CALCIUM AND MAGNESIUM-646 BENNETT AND REED: A METHOD FOR THE CHEMICAL [An@ly!yst, VOl. 96 To another 100-ml calibrated flask (B), add 20ml of water and the same amount of ammonia solution (1 + 1) as was used to neutralise the aliquot in flask A. Add 2 drops of 2,4-dinitrophenol indicator followed by dilute hydrochloric acid (1 + 4) until the solution is neutral and then 5 ml in excess. To both flasks add 6ml of ammonium molybdate (8 per cent. w/v) and stand them for 5 to 10 minutes at a temperature of not less than 20 “C and not more than 30 “C.Then add, with swirling, 45 ml of dilute hydrochloric acid (1 + 1) and leave to stand for 10 minutes. Add 10 ml of tin(I1) chloride solution (1 per cent. w/v), dilute to 100 ml and mix. The deep yellow - brown colour that appears on addition of the tin(I1) chloride is quite normal and does not interfere at the wavelength used. Measure the optical density of the solution in flask A against the solution in flask B in 10-mm cells at 800 nm, or by using a colour filter (Ilford 609) in a suitable instrument. The colour is stable for between 5 and 30 minutes after the addition of the tin(I1) chloride solution. Determine the silica content of the solution by reference to a calibration graph, then add the figure for residual silica content to that obtained for the “gravimetric” silica to obtain the total silica content.Determination of iron(1II) oxide-This determination is for total iron expressed as iron(II1) oxide. Any iron normally present in the sample should have been oxidised to the iron(II1) state during the determination of loss on ignition. The recording of total iron as iron(II1) oxide takes account of this and results in correct analysis totals. Dilute 50.0 ml of the stock solution to 250 ml in a calibrated flask and mix. This solution is referred to as the dilute stock solution and is also used for the determination of magnesium oxide. Transfer 5.00 ml of the dilute stock solution to a 100-ml calibrated flask.Add 2 ml of hydroxylammonium chloride solution (10 per cent. w/v) , 5 ml of 1,lO-phenanthroline solution (1 per cent. w/v) and 2 ml of ammonium acetate solution (about 10 per cent.). Allow the solution t o stand for 15 minutes, dilute to 100 ml and mix. Measure the optical density of the solution against water in 10-mm cells at 510nm, or by using a colour filter (Ilford 603) in a suitable instrument. The colour is stable for 15 to 75 minutes after addition of the ammonium acetate solution. Determine the iron(II1) oxide content of the solution by reference to a calibration graph. Determination of titani.um(IV) oxide-If the sample has a high chromium(II1) oxide content giving a very yellow solution, the chromium in the control solution must be reduced by the addition of 2 ml of hydroxylammonium chloride solution (10 per cent.w/v). Transfer 40.0 ml of the stock solution to each of two 100-ml calibrated flasks, A and B. To each flask add 10 ml of phosphoric acid (2 + 3) and, to flask A only, add 10 ml of hydrogen peroxide solution. Dilute the solution in each flask to 100 ml and mix. Measure the optical density of the solution in flask A against the solution in flask B in 40-mm cells at 398 nm, or by using a colour filter (Ilford 601) in a suitable instrument. The colour is stable from 5 minutes to 24 hours after addition of the hydrogen peroxide solution. Determine the titanium(1V) content of the solution by reference to a calibration graph. Determination of manganese(II) oxide-Transfer 10-0 ml of the stock solution to a 250-ml beaker.Add 10 ml of dilute sulphuric acid (1 + l), 10 ml of dilute nitric acid (1 + 1) and evaporate to strong fumes to remove chlorides. Add 20 ml of nitric acid, 10 ml of dilute phosphoric acid (1 + 9) and about 50 ml of water. Boil the solution to remove any nitrous fumes, filter it through a Whatman No. 40 filter-paper, then add about 0.2 g of potassium periodate. Boil until the pink colour develops and then for a further 2 minutes. Transfer the beaker to a steam-bath and keep it hot for 10 minutes. Allow the solution to cool and transfer it to a 100-ml calibrated flask. Dilute the solution in the flask to 100 ml and mix. Measure the optical density of the solution against water in 40-mm cells at 524nm, or by using a colour filter (Ilford 604) in a suitable instrument.Determine the manganese(I1) oxide content of the solution by reference to a calibration graph. Determination of chromium(III) oxide (a) by diphenylcarbaxide method (for Cr,O, contents of up to approximately 0.1 per cent.)-Transfer 10.0 ml of the stock solution to a 100-ml beaker, add 5 ml of dilute sulphuric acid (1 + 9) and 5 ml of nitric acid and evaporate the mixture to dryness. To the dry residue add 2 ml of dilute sulphuric acid (1 + 9) and about 15 ml of water. Warm to dissolve as much of the residue as possible, then filter, if necessary, through a Whatman No. 42 filter-paper and wash the residue with warm water. Evaporate the solution and washings to about 20 ml, add 2 ml of ammonium cerium(1V) nitrate solutionSeptember, 19711 ANALYSIS OF MAGNESITES AND DOLOMITES 647 (1 per cent.w/v) and allow to stand on a steam-bath for 25 minutes. Cool to 10 "C and add sodium azide solution (2 per cent. w/v), dropwise, to destroy the colour of the excess of cerium(1V) ions. Transfer the solution to a 100-ml calibrated flask containing 3 ml of dilute sulphuric acid (1 + 9) and dilute to about 90 ml. Add 2 ml of diphenylcarbazide solution (1 per cent. w/v), dilute to 100ml and mix. Allow to stand for 5 minutes. Measure the optical density of the solution against water in 10-mm cells at 540nm, or by using a colour filter (Ilford 605) in a suitable instrument. Determine the chromium(II1) oxide content of the solution by reference to a calibration graph. Determination of chromium(III) oxide (b) by EDTA method (for Cr,O, contents above approximately 0.1 per cent.)-Transfer 50.0 ml of the stock solution to a 500-ml separating funnel (a Squibb type is recommended).Add 20ml of chloroform and 10ml of sodium diethyldithiocarbamate solution (10 per cent. w/v), stopper the funnel and shake it vigorously. Release the pressure in the funnel by carefully removing the stopper and rinse the stopper and neck of the funnel with water. Allow the layers to separate and withdraw the chloroform layer. Then wash the aqueous solution with 10-ml portions of chloroform until the chloroform layer is colourless (at least three washes are required) and discard the chloroform extracts. Transfer the aqueous phase to a 400-ml beaker, boil off any traces of chloroform and cool to room temperature.Add 20 drops of hydrochloric acid, followed by 10 ml of sodium sulphite solution, with stirring, and boil for 5 minutes. Cool to room temperature, then add 10 ml of EDTA solution followed by ammonia solution, dropwise, until the first appearance of a slight permanent precipitate. Dissolve this precipitate by adding 20 drops of dilute acetic acid (1 + 1) and dilute to about 200ml. Heat the solution to boiling and boil for 10 to 15 minutes, cool, dilute to 250 ml in a calibrated flask and mix. Measure the optical density of the solution against water in 40-mm cells at 550 nm, or by using a colour filter (Ilford 605) in a suitable instrument. Determine the chromium(II1) oxide content of the solution by reference to a calibration graph. Determination of aZuminium(1II) oxide-Transfer 100.0 ml of the stock solution to a 400-ml beaker.Add 10 ml of sulphurous acid and boil off the excess sulphur dioxide. Cool slightly, add 5 ml of nitric acid and boil for 15 minutes. Cool to about 80 "C, add 5 g of ammonium chloride and stir to dissolve, then add dilute ammonia solution (1 + l ) , with stirring, until the solution is just alkaline to bromophenol blue. Boil off the slight excess of ammonia. Allow the solution to stand for 5 minutes for the precipitate to settle, then filter it through a Whatman No. 541 filter-paper. Rinse the beaker with hot, slightly ammoniacal, ammonium nitrate solution and pour the washings through the filter. Wash the precipitate well with more hot, slightly ammoniacal, ammonium nitrate solution, and discard the filtrate and washings.Place the precipitation beaker under the funnel and dissolve the precipitate through the filter with 40ml of hot, dilute hydrochloric acid (1 + 1). Wash the paper thoroughly with hot water and discard it. Cool the solution and transfer it to a separating funnel. The volume at this stage should be about 100 ml. Add 20 ml of chloroform and 10 ml of cupferron solution (6 per cent. w/v). Stopper the funnel and shake it vigorously, releasing the pressure in the funnel by carefully removing the stopper. Rinse the stopper and neck of the funnel with water, allow the layers to separate and withdraw the chloroform layer. Confirm that extraction is complete by checking that the addition of a few drops of cupferron solution does not produce a permanent coloured precipitate.Add 10-ml portions of chloroform and repeat the extraction until the chloroform layer is colourless. At this point wash the stem of the funnel inside and out with chloroform. Discard the chloroform extracts and transfer the aqueous solution to a 50 0-ml conical flask. Add a few drops of bromophenol blue solution and then add ammonia solution drop- wise until the solution is just alkaline. Re-acidify it quickly with hydrochloric acid and add 5 to 6 drops in excess. Add 10.0 ml of standard EDTA solution (0.05 M) ; this is sufficient for 2.5 per cent. of alumina. Then add ammonium acetate buffer solution (for alumina determination) until the indicator turns blue, followed by 10 ml in excess. Boil the solution for 10 minutes and cool.Add a volume of 95 per cent. ethanol equal to the total volume of the solution, followed by 1 to 2 ml of dithizone solution and titrate with the standard zinc solution (0.05 M), which is run in from a semi-micro or similar burette, from green to the648 BENNETT AND REED: A METHOD FOR THE CHEMICAL [Analyst, Vol. 96 first appearance of a permanent pink colour. (The pinkish tinge that sometimes appears in the solution after the addition of the dithizone and the purple colour due to the chromium - EDTA complex can both be screened out by the dropwise addition of naphthol green B solution.) If the EDTA solution is not exactly 0.05 M, calculate the equivalent volume of exactly 0.05 M solution. If V ml is the volume of 0.05 M EDTA solution and v ml is the volume of zinc solution used in the back-titration, then the percentage of aluminium(II1) oxide is 0.255 (V - v).This figure must be corrected for any chromium(II1) oxide in the sample. Multiply the percentage of chromium(II1) oxide content by 0.667 and deduct this figure from the percentage of aluminium (I I I) oxide. Determination of lime-Transfer a 50-ml aliquot of the stock solution to a 250-ml cali- brated flask. Add 5 ml of dilute triethanolamine solution (1 + l ) , a slight excess (10 ml for magnesites, or as appropriate for dolomites) of EGTA solution (approximately 0.05 M) and dilute to 150ml. Add potassium hydroxide solution (25 per cent. w/v) until no further precipitation takes place and then add 10 ml in excess, followed by 10 ml of Magflok solution.Dilute to 250 ml, shake and allow the mixture to stand for about 10 minutes to settle. Filter it through a 150-mni dry Whatman No. 541 filter-paper into a dry beaker. Then pipette 200 ml of the filtrate into a 500-ml conical flask and add 15 ml of potassium hydroxide solution followed by about 0.03 g of screened Calcein indicator and titrate with standard calcium solution (0.05~) until the first appearance of a green fluorescence. The titration is best carried out in good daylight, but direct sunlight should be avoided. If the EGTA solution is not exactly 0.05 M, calculate the equivalent volume of exactly 0.05 M EGTA solution. If V ml of 0.05 M EGTA are taken and v ml of calcium solution (0.05 M) are used for back-titration, then the percentage of lime = 0-701 x (4/5 V - v).Correct for the blank determination. Detemzination of magnesium oxide-Transfer 100.0 ml of the dilute stock solution [see Determination of iron(II1) oxide] to a 500-ml separating funnel. Add dilute ammonia solution (1 + 1) dropwise, until the solution is slightly alkaline to bromophenol blue. Just re-acidify with dilute hydrochloric acid (1 + 3) and then add 4 ml in excess. Add 20 ml of chloroform and 10 ml of sodium diethyldithiocarbamate solution (10 per cent. w/v), then stopper the funnel and shake it vigorously. Release the pressure in the funnel by carefully removing the stopper and rinse the stopper and neck of the funnel with water. Allow the layers to separate and withdraw the chloroform layer. (If an emulsion has formed, it will be necessary to add a few drops of hydrochloric acid and re-shake the mixture.) Add 10-ml portions of chloroform and 5-ml portions of diethyldithiocarbamate solution, repeating the extraction after each pair of additions until the additions no longer cause a coloured precipitate to form. Finally, wash the aqueous phase three times with 10-ml portions of chloroform. If the sample contains more than approximately 1 per cent.of aluminium(II1) oxide it will be necessary to carry out a buffered cupferron - chloroform solvent extraction as follows. Add dilute ammonia solution (1 + 1) dropwise until the solution is just alkaline to bromophenol blue. Just re-acidify with dilute hydrochloric acid (1 + 9) and add 20 ml of the ammonium acetate buffer solution. Add 20 ml of chloroform and 10 ml of cupferron solution (6 per cent. w/v), stopper the funnel and shake it vigorously.Release the pressure in the funnel by carefully removing the stopper and rinse the stopper and neck of the funnel with water. Allow the layers to separate and withdraw the chloroform layer. Repeat the extraction with a further 10ml of cupferron solution and finally wash the aqueous phase three times with 10-ml portions of chloroform. This separation will remove aluminium and titanium. Transfer the aqueous phase from either the diethyldithiocarbamate separation or, if aluminium(II1) oxide has been removed, from the buffered cupferron separation to a 500-ml conical flask and boil off any traces of chloroform. Cool, add 2 g of ammonium chloride and 2 ml of dilute triethanolamine solution (1 + 1) with swirling, followed by an appropriate, known amount of about 0.05 M DCTA solution (e.g., 80 ml for magnesites or 40 ml for dolo- mites).This addition is made to complex most of the magnesium before the solution is made alkaline, so that the tendency for magnesium hydroxide to precipitate is greatly reduced. Then, add 30 ml of ammonia solution followed by 5 ml of hydroxylammonium chloride solution. Correct for the blank determination. This separation will remove iron and manganese.September, 19711 ANALYSIS OF MAGNESITES AND DOLOMITES 649 Add approximately 0.015 g of Solochrome black 6B indicator and titrate the solution with the standard DCTA solution from red, through purple, until the last change in colour to a clear ice blue.This titration also includes the titration for lime in the sample, and this must be determined by the method given. If the DCTA solution is not exactly 0-05 M, calculate the equivalent volume of exactly 0.05 M DCTA solution. Then, if V is the total amount of 0-05 M DCTA added and z, is the calculated amount of 0.05 M EGTA required to react with the lime in a 0-2-g sample, (V - v ) x 1.008 = percentage of magnesium oxide in the sample. REPORTING OF RESULTS- Results should be reported with respect to either the dried material (dried to constant weight at 110 "C), or the ignited (at 1000 "C) material (calculated to zero loss on ignition). CALIBRATIONS FOR COLORIMETRIC METHODS- Residual silica-Transfer 0, 2.0, 4.0, 6.0, 8.0 and 10.0-ml portions of the standard silica solution B to 100-ml calibrated flasks and add 20, 18, 16, 14, 12 and 10 ml of water, respec- tively.This will give a calibration for 0 to 0.4 per cent. of SiO,. Add to each flask, with swirling, 5 ml of dilute hydrochloric acid (1 + 4) followed by 6 ml of ammonium molybdate solution and leave them to stand for 5 to 10 minutes at a temperature of not less than 20 "C and not greater than 30 "C. Then add, with swirling, 45 ml of dilute hydrochloric acid (1 + 1) and leave to stand for 10 minutes. Add 10ml of tin(I1) chloride solution (1 per cent. w/v), dilute the solution in each flask to 100ml and mix. Measure the optical densities of the solutions against the zero solution in 10-mm cells at 800nm, or by using a colour filter (Ilford 609) in a suitable instrument. The colour is stable for 5 to 30 minutes after the addition of the tin(I1) chloride solution.Prepare a calibration graph from the optical densities. Iro%(III) oxide-Transfer 0, 10.0, 20.0, 25.0, 40.0 and 50-0-ml portions of the standard iron solution B to 100-ml calibrated flasks. This will give a calibration graph for 0 to 5 per cent. of Fe,O,. Add to each flask 2ml of hydroxylammonium chloride solution, 5ml of 1,lO-phenanthroline solution and 2 ml of ammonium acetate solution (approximately 10 per cent.). Allow to stand for 15 minutes, dilute the solution in each flask to 100 ml and mix. Measure the optical densities of the solutions against water in 10-mm cells at 510nm, or by using a colour filter (Ilford 603) in a suitable instrument. The colour is stable for 15 to 75 minutes after the addition of the ammonium acetate solution.From the optical densities prepare a calibration graph. Titanium(IV) oxide-Transfer in duplicate 0, 10.0, 20.0, 25.0, 40.0 and 50.0-ml portions of the standard titanium solution B to 100-ml calibrated flasks. This will give a calibration graph for 0 to 0.5 per cent. of TiO,. Add to each flask 1Oml of dilute phosphoric acid (2 + 3) and to one of each pair only add 10 ml of hydrogen peroxide solution. Dilute the solution in each flask to 100ml and mix. Measure the optical densities of the pertitanic acid solutions against the appropriate control solution in 40-mm cells at 398 nm, or by using a colour filter (Ilford 601) in a suitable instrument. The colour is stable for 5 minutes to 24 hours after the addition of the hydrogen peroxide solution.From the optical densities prepare a calibration graph. Manganese(1I) oxide-Transfer 0, 10.0, 20-0, 25.0, 40.0 and 50.0-mZ portions of the standard manganese solution B to 100-ml calibrated flasks. This will give a calibration graph for 0 to 0.5 per cent. of MnO. Dilute the solution in each flask to 100 ml and mix. Measure the optical densities of the solutions against water in 40-mm cells at 524 nm, or by using a colour filter (Ilford 604) in a suitable instrument. From the optical densities prepare a calibration graph. Ch.romium(IIl) oxide-For method (a), the diphenylcarbazide method, transfer 0, 1.0, 2.0, 3.0, 4.0 and 5-0-ml portions of the standard chromium solution B to 100-ml calibrated flasks.This will give a calibration graph for 0 to 0.125 per cent. of Cr,O,. Add to each flask 5 ml of dilute sulphuric acid (1 + 9) and dilute to about 90 ml. Add to each 2 ml of diphenylcarbazide solution (1 per cent. w/v), dilute to 100ml and mix. Allow to stand for 5 minutes. Measure the optical densities of the solutions against water in 10-mm cells at 540nm, or by using a colour filter (Ilford 605) in a suitable instrument. From the optical densities prepare a calibration graph.650 BENNETT AND REED: A METHOD FOR THE CHEMICAL TABLE I RESULTS ON MAGNESITE SAMPLE AN 31 (IGNITED AT 1000 "C) [Analyst, Vol. 96 Labora- SiO,, TiO,, Fe,O,, Alto3, Cr,O,, MnO, tory per cent. per cent. per cent. per cent. per cent. per cent. A B C D E F G Mean Standard deviation 2.50 0.03 2.48 0.03 2.50 0.03 2.43 0-05 2-46 0-05 2.49 0.04 2-48 0.04 2.49 0.04 2.48 0.04 2.46 0.039 2.46 0.037 2.45 0.039 2.49 0.03 2-50 0.04 2.51 0.03 2-51 0.03 2.5 1 0.03 2.52 0.03 2.48 0.03 2-52 0.03 2.48 0.02 2.49 0.04 0.02 0.01 1.75 0.84 0.07 1-76 0.85 0.07 1.75 0.86 0.07 1.83 0.88 0.07 1.77 0.89 0.07 1.83 0.88 0.07 1.74 1.73 1.73 1.76 1.75 1-76 1.79 1.81 1-76 0.87 0.86 0.86 0.85 0.87 0.85 0.84 0.85 0.84 0.07 0.08 0.07 0-066 0.067 0.067 0.08 0.07 0.08 1.76 0.84 0.08 1.76 0.84 0.08 1.75 0.84 0.08 1-80 0.84 0.07 1-78 0.84 0.06 1.77 0.82 0.07 1.77 0.85 0.07 0.03 0-02 0.01 0.09 0.09 0.09 0.08 0.075 0.08 0.11 0.12 0.11 0-087 0.087 0.087 0.10 0.12 0.10 0.09 0-09 0.09 0.12 0.12 0.11 0.10 0.01 CaO, per cent.2.31 2.30 2-32 2.30 2.31 2.30 2.33 2.32 2.34 2.31 2.35 2-31 2-32 2.3 1 2.34 2.34 2.31 2.31 2-28 2.33 2.30 MgO, per cent.92.1 1 91.97 91.97 92.04 92.04 92.27 92-62 92.48 92.78 92.30 92-46 92.52 92.17 92-03 92.17 92.04 92.17 92.17 92-14 91-80 92.03 2.32 92-20 0.02 0.24 Total,* per cent. 99-70 99.55 99.59 99.68 99-67 99-96 100.26 100.12 100.41 99.88 100.09 100.09 99.82 99-59 99.83 99-70 99.79 99.79 99.76 99.48 99.60 * The alkali-metal oxides in this sample total about 0.05 per cent. and have not been included in the total. TABLE I1 RESULTS ON MAGNESITE SAMPLE AN 32 (IGNITED AT 1000 "c) Labora- SiO,, TiO,, Fe203, A1,03, Cr,03, hgno, tory per cent. per cent. per cent. per cent. per cent. per cent. A B C D E F G Mean Standard deviation 0.95 0.93 0.94 0.97 0.97 0.96 0.95 0.94 0.95 0.98 0.98 0.98 0.99 0.97 1.00 1.01 0.99 1-01 1.00 0.97 0.98 0-97 0.02 0.01 0.02 0.02 0.026 0.026 0-026 0.02 0.03 0-02 0.029 0.027 0.027 0.03 0.02 0.02 0.02 0-02 0.02 0.03 0.02 0.02 0.02 0.01 5.49 0.99 5.49 0.99 5.49 0.99 5.41 0.91 5.49 0.93 5.49 0.93 5.31 1-05 5-43 1.09 5-33 1.02 5.40 0.98 5-40 0.99 5.45 0.99 5-46 1.01 5-42 0.99 5.45 1.00 5.42 1-01 5.38 1.01 5.42 1.00 5.45 0.96 5-50 0.98 5-50 0.98 5.44 0.99 0.05 0.04 0.76 0.11 0.76 0.11 0.76 0.11 0.76 0.08 0-74 0.09 0.76 0.075 0.74 0.77 0.77 0.80 0.77 0.77 0.72 0.72 0.68 0.74 0.74 0.75 0.72 0.74 0.74 0.75 0.16 0.16 0.15 0.109 0.109 0.109 0.13 0.12 0-12 0.13 0.13 0.14 0.13 0.12 0.12 0.12 0.03 0.02 CaO, per cent.1.79 1-78 1-77 1.78 1-79 1.78 1.78 1.76 1.74 1.81 1-79 1.80 1.70 1.76 1-70 1.78 1.75 1.75 1-73 1.70 1.70 1.76 w@, per cent. 89.60 89.68 89.55 89.91 89.80 89.76 89.59 89.41 89.36 89-50 89.45 89-37 89-47 89.47 89-60 89.76 89-75 89.81 89.36 89.47 89.23 89.57 0-04 0.18 Total, * per cent. 99.70 99-76 99-63 99.85 99.84 99.79 99.60 99-59 99.34 99.61 99-52 99.50 99.51 99.47 99.55 99.87 99.77 99.90 99.38 99-50 99.27 * The alkali-metal oxides in this sample total about 0.04 per cent.and have not been included in the total.September, 19711 ANALYSIS OF MAGNESITES AND DOLOMITES 651 For method (b), the EDTA method, transfer 0, 5.0, 10.0, 15.0 and 20.0-ml portions of the standard chromium solution A to 400-ml beakers and add 50, 45, 40, 35 and 30ml of water, respectively. This will give a calibration graph for 0 to 2 per cent. of Cr,O,. Add to each 20 drops of hydrochloric acid, followed by 10 ml of sodium sulphite solution (5 per cent.w/v) with stirring, and boil for 5 minutes. Cool to room temperature, add 10ml of EDTA solution (5 per cent. w/v) followed by ammonia solution dropwise, until the first appearance of a permanent precipitate. Dissolve this precipitate by adding 20 drops of dilute acetic acid (1 + 1) and dilute to about 200ml. Heat the solutions to boiling and boil for 10 to 15 minutes, cool, dilute to 250 ml in calibrated flasks and mix. Measure the optical densities of the solutions against water in 40-mm cells at 550nm, or by using a colour filter (Ilford 605) in a suitable instrument. From the optical densities prepare a calibration graph. RESULTS The results obtained by the Refractories Working Group are shown in Tables I to IV. The samples are two magnesites (Tables I and 11) and two dolomites (Tables I11 and IV); the dolomite in Table 111, B.C.S.368, is a standard sample. Labora- tory A B C D E F G H I Mean Standard deviation TABLE IIr RESULTS ON DOLOMITE SAMPLE B.C.S. 368 (DRIED AT 110 "c) SO,, TiO,, Fe,O,, Al,O,, Cr,O,, MnO, CaO, MgO, Loss, per cent. per cent. per cent. per cent. per cent. per cent. per cent. per cent. per cent. 0.90 <0-01 0-92 <0-01 0.93 <0*01 0.92 0.005 0-91 0-005 0.93 0-005 0.93 <0.01 0.95 <0.01 0.94 <0.01 0.89 0.006 0.90 0.006 0.91 0.006 0.90 <0-01 0.93 <0.01 0.90 <0*01 0.92 0.005 0.94 0.005 0.93 0.005 0.92 0.01 0.93 0.01 0.92 0-01 0.89 0.004 0.92 0.004 0.90 0.004 0.90 0.003 0.89 0.002 0.91 0.002 0.92 0.004 0.01, 0.003 0.22 0-22 0.22 0.21 0.22 0.21 0.22 0.22 0.2 1 0.22 0.22 0.22 0.24 0.24 0.24 0-24 0.24 0.25 0-23 0.22 0.23 0-23 0.23 0.21 0.24 0.21 0.25 0.23 0.01, 0-17 0.17 0.18 0.17 0.17 0-17 0.15 0.15 0.15 0.16 0.17 0.18 0-14 0.14 0.13 0.16 0.16 0.16 0.18 0.17 0.18 0.17 0.17 0.17 0.19 0.16 0.20 0.17 0.01, <0.01 <0-01 <0*01 0.009 0.010 0.009 <0*01 < 0.01 <0*01 0.006 0,006 0.006 0.010 0-008 0.009 0.013 0.012 0.013 0.01 0.01 0.01 0.006 0.006 0-006 0.01 0.005 ' 0.01 0.007 0.004 0.04 0.04 0.04 0.058 0.055 0.060 0.06 0.06 0.05 0.06 0.06 0.06 0-06 0.06 0.06 0-06 0.06 0.06 0.03 0.04 0.04 0.06 0.06 0-06 0.07 0.06 0.06 0.055 0.010 30.66 30.68 30.68 30-75 30.77 30.75 30.81 30-85 30.83 30.70 30.69 30.69 30.70 30.84 30.84 30.88 30.84 30.84 30.91 30.84 30.80 30.91 30-81 30-81 30-52 30.66 30.66 30.77 0.09, 20.97 20.97 20.92 20.93 20.86 20.97 20.67 20.69 20.77 20.84 20.90 20.80 21.27 21.27 21.08 20.76 20.84 20.85 20.30 20-40 20.40 20.88 20.88 20.88 21-04 21.14 21-04 20.86 0.23 DEVELOPMENT OF THE METHOD AND DISCUSSION OF RESULTS LOSS ON IGNITION- 46.78 46-8 1 46-82 46.80 46.80 46.70 46.77 46.77 46-77 46-68 46.67 46.67 46.59 46.60 46.50 46.67 46.67 46.67 46.86 46-82 46.86 46.50 46.38 46.46 46.54 46.59 46.52 46-68 0-13 Total, per cent.99.74 99.81 99.79 99.86 99.81 99.81 99.61 99.69 99.72 99.57 99.63 99.85 99.91 100.09 99.76 99-67 99.72 99-73 99.45 99.44 99-45 99.65 99.46 99.50 99.51 99-72 99.65 No difficulty was experienced with this determination as it was appreciated that when large volumes of carbon dioxide were to be evolved, care would be needed during the warming- up stage.It was also realised that the crucibles used would need to have well fitting lids so as to minimise re-carbonation during cooling. The results of analyses shown in Tables I to IV are calculated with respect to ignited material, with the exception of the dolomite (B.C.S. 368), which was a raw sample. As the652 Labora- tory A B C D F G H I Mean Standard deviation BENNETT AND REED: A METHOD FOR THE CHEMICAL TABLE IV RESULTS ON DOLOMITE SAMPLE AN 34 (IGNITED AT 1000 "C) [Analyst, Vol. 96 SiO,, per cent. 1-04 1.04 1.06 0.97 1.02 0.99 1-05 1-05 1.05 1-04 1.01 1.04 1.03 1-03 1.04 1-02 1.03 1.03 0.95 0.95 0.96 1.02 1.04 1-02 1.02 0.03, TiO,, per cent. 0.01 0.01 0.01 0-005 0.005 0.005 0.02 0.03 0.03 0.026 0.026 0.026 0.01 0.01 0.01 0.01 0.02 0.01 0.015 0.015 0-015 0.025 0-020 0.025 0,016 0.008 FezO3, per cent.1-04 1-06 1.04 1.14 1.14 1.13 1.06 1.06 1-05 0.98 0.98 0.99 1.04 1.04 1.03 1.08 1.07 1.08 1-04 1.04 1.00 1.06 1.06 1.06 1.05 0.04, 0.38 0.36 0.38 0.38 0.38 0.39 0-37 0.39 0.39 0.36 0.38 0.36 0.38 0.37 0.37 0.38 0.37 0.37 0.40 0.40 0.39 0.39 0.40 0.39 0.38 0.01, 0.01 0.01 0.01 <0*01 <0*01 < 0.01 0.01 0.0 1 0.01 0.009 0.009 0.009 0.01 0.01 0.01 0.0 1 0.01 0.0 1 0.006 0.006 0.008 t 0 . 0 1 (0.01 <om01 0.007 0.004 MnO, per cent. 0.13 0.15 0.15 0.10 0.10 0.10 0.16 0.16 0.16 0.15 0.14 0.14 0.14 0-14 0.14 0.14 0.14 0.13 0.16 0.16 0.16 0-15 0.16 0.16 0.14 0.02 CaO, per cent. 57.06 57.13 57.06 57-14 57.12 57.26 57.09 57.17 57.15 57.29 57.31 57.49 56.95 56.99 56.97 57.25 57.40 57.25 57-17 57.04 57.17 57.21 57.27 57.14 57.17 0-13 BlgO, per cent.39.92 39.99 39.99 39.80 39.72 39-76 40.19 40.08 40.26 40.03 39.80 39.71 39.90 39.87 39.94 39.90 39.76 39.90 40.1 1 39.92 39.74 40.35 40-22 40.22 39.96 0-19 Total, per cent. 99.59 99-75 99.70 99.54 99.49 99.64 99.95 99.95 100.10 99-90 99-66 99.77 99.46 99.46 99.51 99.79 99-80 99.78 99.86 99.54 99.45 100.21 100.17 100.02 samples (particularly dolomite) are unstable when exposed to the atmosphere after grinding the calculation of results on an ignited basis is the only reasonable method of comparing analytical results. Only when the materials are in their natural (unfired) state can comparisons be made on a dried (at 110 "C) basis. DETERMINATION OF SILICA- In the method originally proposed3 the sample was decomposed with hydrochloric acid and the insoluble residue fused with sodium carbonate.After this treatment, silica was determined gravimetrically by means of a single dehydration. This process was very time consuming and attempts were made, for routine use, to eliminate the fusion stage. It was found that solutions of samples decomposed with hydrochloric acid, even if evaporated to dryness and re-dissolved in dilute hydrochloric acid, were very slow to filter ; substitution of perchloric acid for the hydrochloric acid, and evaporation, first to fumes and then for a further 30 minutes, instead of dehydration, improved the speed of filtration considerably. Co-operative results with perchloric acid, together with a l-g sample of magnesite, resulted in standard deviations of about 0-08 per cent.As the industrial demand was for an accuracy of at least 50.1 per cent. these results were clearly of little value. The errors were greatly minimised by increasing the size of the sample to 5g, at which level the standard deviation dropped to 0.01 to 0-03 per cent. (coefficient of variation about 1 per cent.). The main disadvantage of the method was the need to use about 40ml of perchloric acid (60 per cent. w/v) and to heat it to fumes in a glass beaker for at least 30 minutes. The use of perchloric acid in such large amounts introduced an element of danger which, although acceptable from necessity, wits nevertheless not desirable. An additional factor was the cost of the perchloric acid which, when used in a duplicate determination, was 40p. After the discovery that polyethylene oxide coagulated silica in acidic ~olution,~ it seemed obvious to investigate the effect of this reagent on the solutions resulting from hydrochloric acid decomposition of magnesites and dolomites.This treatment proved effectiveSeptember, 197 11 ANALYSIS OF MAGSESITES AND DOLOMITES 653 and the residues could be quickly filtered off, provided slight changes were made to the volumes of liquids so as to keep the total volume at the coagulation stage to a reasonable minimum. Perchloric acid attack yielded much smaller residues than those obtained with hydrochloric acid so that there seemed to be a risk of retention of sulphur trioxide after the hydrofluoric acid - sulphuric acid treatment, thus giving rise to low silica figures. In fact, it was demonstrated that the figures obtained for silica when using hydrochloric acid were within 0.01 per cent.of those obtained with perchloric acid decomposition. The use of a coagulant did, in fact, have one disadvantage. It proved impossible to use the yellow molybdosilicate colour for the determination of the residual silica because of clouding of the solution. However, reversion to the molybdenum-blue method overcame this difficulty. DETERMINATION OF IRON(II1) OXIDE- No difficulty was experienced in adapting the 1,lO-phenanthroline method to magnesites or dolomites. To maintain accuracy it was necessary to prepare a preliminary dilution so as to take a reasonably sized aliquot. This dilution served a second purpose in that it enabled a large aliquot to be taken for the determination of magnesium oxide.DETERMINATION OF TITANIUM(1V) OXIDE- No difficulty was experienced in adapting the hydrogen peroxide method. The low level of titanium(1V) oxide in these classes of materials makes this method rather insensitive but the present technological demands of the industry do not require a high order of accuracy, rather a check to ensure that a reasonably low level is maintained. Thus there seemed t o be no reason to introduce more sensitive and, probably, more complicated procedures. DETERMINATION OF MANGANESE(II) OXIDE- The periodate method offered no difficulties in its application to magnesites and dolomites. It was anticipated that some difficulty might be caused by the low solubility of calcium sulphate and it had been intended to replace the sulphuric acid used for removing chlorides by perchloric acid.In fact, no difficulty was encountered in this respect. Measurement against a portion of the same solution reduced with sodium nitrite would probably improve the results. However, the manganese content of these materials is not significant in practice and is usually determined only to ensure that it is not unusually high. DETERMINATION OF CHRQMIUM(III) OXIDE- The normal diphenylcarbazide procedure was first adopted for this determination but it was found that, at above about 0.1 per cent., the results tended to be low. For this reason the EDTA spectrophotometric method, similar to that used for chrome-bearing material^,^ was adopted. This proved to be sufficiently sensitive for amounts of chromium(II1) oxide greater than about 0.1 per cent.Therefore both methods are included so as to ensure full coverage. DETERMINATION OF ALUMINIUM(III) OXIDE- Several techniques were tried in the early stages of development of the present method, vix., gravimetric determination with 8-hydroxyquinoline, spectrophotometric determination with the same reagent and direct titration with EDTA. All of these methods resulted in high figures caused by the interference of magnesium, so it became clear that a separation from magnesium was necessary. Separation was accomplished by precipitation with am- monia solution, which was conducted carefully to ensure complete recovery of aluminium even at the expense of not removing all the magnesium. Thus, if adequate care is taken, the method is not open to the normal objections when crude precipitation is recommended when it is the filtrate that is of prime importance.No attempt is made to oxidise the manganese before the precipitation so that it will escape and need not be corrected for in the EDTA titration. Iron and titanium are, of course, removed in the usual way with a cupferron- chloroform solvent extraction. Chromium, which may be present in moderate amounts (up to 0-75 per cent. in isolated cases), will also be titrated. As it is not possible to determine easily the amount of chromium in the solution after titration, it is necessary either to note its presence or ensure its absence. As it is far more difficult to ensure its absence, it is reduced The results show a relatively high coefficient of variation.654 [Arcalyst, vol.96 and precipitated quantitatively with the alumina. In a recent paper1 DCTA was used as a replacement for EDTA. The advantage of DCTA lies in the fact that aluminium is com- plexed in cold solution, whereas chromium is not. Thus, the interference from chromium is eliminated and so are the need and the desirability to reduce the chromium because it is best to allow as much as possible to pass into the filtrate in the form of chromate. In addition, the determination is more rapid in that there is no need to boil and cool the solution. The use of DCTA entails no change in the method as described here except that, to avoid indicator fading, about 2 g of hydroxylammonium chloride must be added to the solution before adding the dithizone.The results for aluminium(II1) oxide show standard deviations for low contents (less than 0.4 per cent.) of about 0.01 per cent. At the higher level of 0.8 to 0.9 per cent. the standard deviation increases to about 0.02 per cent. in the absence of Cr,O, and 0.04 per cent. when 0.75 per cent. of Cr,O, is present. The latter may serve to indicate the increased errors due both to greater problems with the actual titration and also to the errors in the determination of the chromium. Thus the use of DCTA could be expected to improve the accuracy of the aluminium(II1) oxide determination. DETERMINATI ON OF LIME- The method used for the determination of lime is similar to an earlier routine method6 wherein excess of EDTA was added to complex the lime and the magnesium was then precipitated with alkali-metal hydroxide solution.This determination, together with the determination of magnesium oxide, gave the Working Group a large number of problems. Several techniques were tried with various indicators but each gave rise to difficulties in one or more laboratories. Potentiometric methods of determining the end-points were also tried, but without success. In the end it was necessary to revert to a method similar to the original but which included a filtration and the use of screened Calcein as indicator. The speed of this filtration was increased by the use of a partially hydrolysed polyacrylamide coagulant that has been named Magflok. The substitution of EGTA for EDTA greatly improved the end-point because the superior stability of the calcium complex over that of magnesium is more apparent with EGTA than with EDTA.The results on the magnesites are very satisfactory; it will be noted that the sample containing 0.75 per cent. of Cr,O, again gives poorer results. This could be caused by background coloration from chromium in the titration solution. The results on the dolomites showed standard deviations (with reference to the ignited material) of 0-13 and 0.18, which is about the level normally obtained in co-operative work on ceramic materials between laboratories for major content determinations. DETERMINATION OF MAGNESIUM OXIDE- Difficulties similar to those with lime were experienced with the end-point for magnesium oxide and, once again, the Working Group tried a number of methods before developing the present procedure.In order to achieve the accuracy desired by the refractories industry it was necessary to take a large aliquot; this was done by using the dilute stock solution made for the iron determination. In the presence of large amounts of magnesium oxide end-points almost inevitably tend to drag and it was therefore desirable to use the most suitable indicator. This precluded the use of the more usual methylthymol blue and the Working Group turned to the Solochrome dyes, preferring, in the end, Solochrome black 6B (blue-black B) to the more usual Solochrome black T. These indicators are easily destroyed and interfering elements such as iron and manganese must be removed. End-points were again shown t o be sharper when EDTA was replaced, this time by DCTA.Since completing the results in the table, it has been demonstrated that aluminium(II1) oxide tends to interfere in the determination of magnesium oxide (see Table V). For accurate work, therefore, if the aluminium(II1) oxide content is greater than about 0.50 per cent., it is desirable to include the aluminium(II1) oxide separation. For normal work, however, it is possible to allow the presence of up to 1 per cent. of aluminium(II1) oxide, as the level of interference is still not greater than the experimental error. The inter- ference was not considered large enough to warrant repeating the analyses reported in the tables, The results are, of course, inferior to those for lime as the magnesium oxide content BENNETT AND REED: A METHOD FOR THE CHEMICALSeptember, 19711 ANALYSIS OF MAGNESITES AND DOLOMITES 655 TABLE V INTERFERENCE OF ALUMINIUM(III) OXIDE IN THE DETERMINATION OF MAGNESIUM OXIDE WITH DCTA Aluminium(II1) oxide added, per cent. 0.0 0-25 0.50 0.75 1.0 2.0 3.0 Magnesium oxide added, per cent. 101.1 (= 202 mg) 101.1 101.1 101.1 101.1 101.1 101.1 Magnesium oxide found, per cent. 101.1 101.1 101.1 101-25 101-25 102.1 103.0 is obtained by difference so that the errors in the lime determination may add to those in the magnesium oxide titration. This will be most significant in the case of the dolomites. The lime and magnesium oxide contents are usually determined on dolomites to ensure that the CaO - MgO ratio is not abnormal, and in the case of magnesites, prior to the develop- ment of this method, a magnesium oxide figure was normally obtained by the difference from 100 per cent. after subtracting the minor constituents. It is therefore clear that both sets of results are, in practice, sufficiently accurate. We thank Dr. N. F. Astbury, Director of Research of the British Ceramic Research Association, for permission to publish this paper. The work was carried out by certain members of the Refractories Working Group, the personnel of which was as follows- H. Bennett (Convener) W. C. Coppins . . J. Davey .. .. H. W. H. Pollitt . . B. Fletcher .. F. C. Gilbert . . A. A. Lea . . E. W. Orrell .. J. Sanderson . . C. E. A. Shanahan N. F. C. Shelton . . R. F. Statham . . P. Wilburn .. A. K. Wright . . R. A. Reed (Secretary) P. Hopkins .. .. D. Shireby .. . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. The British Ceramic Research Association Buyeau of Analysed Samples Ltd. Downan Long (Steel) Ltd. Associated Portland Cement Manufacturers Ltd. Pickford, Holland and Co. Ltd. Steetley Organization Research Department The Steel Co. of Wales Ltd. Pilkington Brothers Ltd. The Carborundum Co. Ltd. Consett Iron Co. Ltd. Stewarts and Lloyds Ltd. General Refractories Ltd. The United Steel Cos. Ltd. Samuel Fox and Co. Ltd. Simon Engineering Ltd. Lysaghts Scunthorpe Works Ltd. The British Ceramic Research Association REFERENCES 1. 2. 3. 4. 5. 6. Bennett, H., and Reed, R. A., Analyst, 1970, 95, 541. Eardley, R. P., and Reed, R. A., Ibid., 1971, 96, in the press. Bennett, H., and Hawley, W. G., “Methods of Silicate Analysis,’’ Academic Press, London and Bennett, H., and Reed, R. A., Analyst, 1967, 92, 466. Bennett, H., and Marshall, K., Ibid., 1963, 88, 877. Thompson, H. V., Mayer, A., Padget, G., Chirnside, R. C., and Bennett, H., Trans. BY. Ceram. Received July 31st, 1969 Amended March 24th, 1971 Accepted April 27t12, 1971 New York, 1965, p. 212. SOL, 1959, 58, 353.

ISSN:0003-2654    

DOI:10.1039/AN9719600640

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

656 Analyst, September, 1971, Vol. 96, $9. 656-658 Acid as a Fluorescent Indicator for the Complexometric Titration of Calcium plus Magnesium BY R. L. CLEMENTS, J. I. READ AND G. A. SERGEANT (Department of Trade and Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) 4-[Bis(carboxymethyl)aminomethyl]-3-hydroxy-2-naphthoic acid, known also as l-dicarboxymethylaminomethyl-2-hydroxy-3-naphthoic acid, a spectro- fluorirnetric reagent for beryllium, has been found to be an effective fluori- metric indicator for the complexometric titration of calcium plus magnesium when used in conjunction with a suitable fluorimetric titrimeter. A procedure is described in which the indicator is used in the titrirnetric determination of magnesium in silicate rocks.THE synthesis of 4- [bis(carboxymethyl)aminomethyl]-3-hydroxy-2-naphthoic acid, originally named as l-dicarboxymethylaminomethyl-2-hydroxy-3-naphthoic acid (DHNA), and its use as a spectrofluorimetric reagent for beryllium have been described by BudWnskf and West,l who noted that the reagent forms fluorescent complexes with other metals, e.g., it gives a green fluorescence with calcium and blue fluorescence with magnesium. In this paper DHNA is proposed as a fluorescent indicator for the complexometric titration of calcium plus magnesium, and a procedure is described for its application in silicate analysis. At pH 10 the fluorescence of the calcium and magnesium complexes of DHNA is progressively quenched by the addition of 1,2-diaminocyclohexane-NNN~N~-tetraacetic acid (DCTA), and the disappearance of this fluorescence occurs precisely at the equivalence point.Residual fluorescence of the reagent renders visual detection of the end-point rather difficult, although its detection is possible in a darkened room. The end-point can, however, be observed pre- cisely under normal lighting conditions with the aid of a suitable fluorimetric titrimeter. An instrument of simple construction has been described elsewhere2 and was used in the present work. METHOD APPARATUS- The simple fluorimetric titrimeter described by Clements and Sergeant2 is suitable. An air-bath is also used. This consists of a gas-ring on which stands an open-ended The PTFE basin is placed Vitreosil cylinder, 230mm high and 125 mm in diameter.over a small flame on a silica triangle supported by the cylinder. REAGENTS- water. Dilute nitric acid (1 + 1)-Mix equal volumes of concentrated nitric acid (spgr. 1.42) and Perchloric acid, 60 per cent. w/w. Hydrofluoric acid, 40 per cent. w/w. Dilute sulphuric acid, approximately 20 N-Add 1 volume of concentrated sulphuric acid (sp.gr. 1-84) cautiously, with stirring, to sufficient water to bring the final volume, when cooled, to 2 volumes. Triethanolamine solution (1 + 1)-Dilute 1 volume of triethanolamine with an equal volume of water. Bufer solution, pH 10-Dissolve 67.5 g of ammonium chloride in water, add 570 ml of concentrated ammonia solution (sp.gr. 0*88), and dilute the mixture to 1 litre with water. l-Dicarboxymethylaminomethyl-2-hydroxy-3-na~hthoic acid (DHNA) indicator-Prepare a mixture of the reagent with sodium chloride containing 0-5 per cent.w/w of the reagent. Magnesium standard solution-Dissolve 0-603 g of clean magnesium ribbon in about 100 ml of water containing 10 ml of perchloric acid, and dilute the solution to 1 litre with water. This solution contains the equivalent of 1 mg ml-l of magnesium oxide. 0 SAC; Crown Copyright Reserved.CLEMENTS, READ ASD SERGEANT 657 DCTA solution, ap$roxi.lnately 0.01 M-Dissolve 3.3 g of 1,2-diaminocyclohexane- NNN'N'-tetraacetic acid in about 200 ml of water by making small additions of M sodium hydroxide solution. Adjust the pH of the solution to about 10 with acetic acid, and dilute to 1 litre with water. Standardise the solution against standard magnesium solution by titration as described in the procedure given below.PROCEDURE FOR THE DETERMINATION OF MAGNESIUM IN SILICATE ROCKS- Transfer 1.0000 g of powdered rock to a platinum or PTFE basin. Add about 20 ml of water followed by 2 ml of dilute nitric acid, 5 ml of perchloric acid and 20 ml of hydro- fluoric acid, and allow the rock to digest overnight or longer at room temperature. Complete the decomposition of the sample by evaporating the mixture to dryness on an air-bath, and repeat the evaporation nearly to dryness with three further portions of perchloric acid, to the first of which 2 ml of dilute sulphuric acid have been added. Finally, add 5 ml of per- chloric acid and about 50 ml of water, cover the basin and heat it on a water-bath or hot-plate to dissolve the residue.Dilute the solution to 200 ml at room temperature in a graduated flask. Add, by pipette, to the titration vessel 10ml of the rock solution followed by about 80 ml of water. Transfer the vessel to the titrimeter and apply magnetic stirring, then add 5 ml of triethanolamine solution, 10 ml of buffer solution and about 30 mg of DHNA indicator. Titrate calcium plus magnesium with DCTA solution under filtered ultraviolet illumination (365 nm) by using the fluorimetric titrimeter. A blue gelatine filter is placed over the photo- cell to improve discrimination; for this purpose the blue component of the Ilford Spectrum Blue-Green 603 filter has been found to be effective. The approach to the titration end-point is characterised by a stepwise reduction in the meter reading for each drop of titrant added, and although the final steps may be fairly small, the end-point is well defined and stable.The magnesium present is calculated from the titration value after making the appropriate deduction for calcium, which may be determined by fluorimetric titration of a further aliquot of solution with ethylene glycol bis(aminoethy1)tetraacetic acid (EGTA) at pH 13 in the presence of triethanolamine, with Calcein as indi~ator.~ RESULTS AND DISCUSSION In this laboratory the fluorimetric titration of calcium plus magnesium, with DHNA as indicator, has now largely replaced the visual titration with Eriochrome black T indicator. The advantages claimed are the elimination of subjective judgment in the assessment of the end-point, and that it is not necessary to convert iron into ferrocyanide in order to mask it sufficiently to prevent interaction with the indicator.A reducing agent such as hydroxyl- amine is, therefore, not added, and the use of cyanide is required only if significant amounts of elements such as cobalt and nickel are known to be present, which need to be masked. In addition, large amounts of manganese do not interfere except to the extent that the triethanolamine complex reduces the light transmittance of the solution. Of other inter- ferences, barium and strontium have been found to be partially titrated in the presence of calcium and magnesium, but do not themselves form fluorescent complexes with DHNA. COMPARATIVE Material analysed Granite, G-1 . . Diabase, W-1 .. Lamprophyre . . Rhyolitic tuff . . Chloritic tuff . , Spilitic tuff . . Perido ti te . . Norite . . .. TABLE I FIGURES FOR MAGNESIUM DETERMINATIONS ON ROCK SAMPLES MgO, per cent. By method with DHNA indicator f A % Calcium found r Operator 1 Operator 2 Other figures CaO, per cent. A > (average) .. 0.35 0.33 0.38" 1.32 .. 6-68 6.64 6*62* 10.84 .. 0.39 0.40 0-42t 0.37 .. 1.98 1.95 2*08t 0.85 - 5.12 .. 10.75 10.82 .. 6.46 6-50 6.54t 4-34 .. 30.89 30.71 30.90t 1.11 .. 18-05 17.98 17.93t 4.69 * From the compilation by M. Flei~cher.~ 7 Titration with Eriochrome black T indicator,658 CLEMENTS, READ AND SERGEANT In Table I are shown some results of magnesium determinations obtained for several This paper is published by permission of the Government Chemist and the Director of silicate rocks by using the procedure described above. the Institute of Geological Sciences. REFERENCES 1. 2. 3. 4. BudMnskg, B., and West, T. S., Analytica Chim. Ada, 1968, 42, 455. Clements, R. L., and Sergeant, G. A., Lab. Pract., 1970, 19, 813. Pribil, R., and Vesely, V., Chemist Analyst, 1966, 55, 82. Fleischer, M., Geochim. Cosmochim. Acta, 1969, 33, 65. Received March 30th, 1971 Accepted May 3rd, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600656

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

Analyst, September, 1971, Vol. 96, pp. 659-664 659 Use of the Halphen Reaction for the Determination of the Cyclopropenoid Content of Lipids BY T. W. HAMMONDS, J. A. CORNELIUS AND L. TAN (Foreign and Commonwealth Ofice (Overseas Development Administration), Tropical Products Institute, 56/62 Gray’s I n n Road, London, W.C.1) An application to cottonseed oils of a quantitative version of the Halphen test for the determination of cyclopropenoid material has been published by other workers, but for other oils containing higher levels of cyclopropenoids, although the absorption at the 495 nm peak is linearly related to the con- centration of each oil examined, the relationship differs among the oils. How- ever, transmethylation of oil before applying the Halphen reaction has been found t o give results that are in better agreement with titration with hydrogen bromide for oils with widely differing cyclopropenoid content.The use of pressurised capsules for carrying out the reaction with reduced loss of solvent has proved advantageous, as flatter peaks are obtained when optical absorp- tion is plotted against time. The application of the modified technique to oils containing a wide range of concentrations of total cyclopropenoid material in the component fatty acids is described and discussed. THE Halphen reaction1 is used either as a qualitative test for cottonseed oil or to detect the presence of this oil in other vegetable oil^,^,^ but its usefulness as a specific test is limited by the positive reactions also given by seed oil from certain other species of the order Mal~ales.~ The nature of the red colour developed has not been fully elucidated, but positive Halphen tests have been associated with the occurrence of cyclopropenoid fatty acid5y6y7 and proportional increases in absorbances, measured at 505 nm, have been reported for mixtures of sterculic acid in corn The relationship between cyclopropenoid contents determined by hydrogen bromide titration and absorbances obtained by using the Halphen reaction has also been r e p ~ r t e d .~ Colorimetric methods involving the use of the Halphen reaction have been used for the determinatiori of cyclopropenoid material in cottonseed oil or meal, with crude cottonseed oil or cyclopropenoid-containing oils of unstated origin as standards.9Jo The 0.400 0.360 0 - 40 80 120 160 200 240280 320 360 400 440 480 Concentration of cyclopropenoid rnaterial/pg per 25 rnl of solvent Fig.1. Values with Sterculia foetida oil and its methyl esters by using Bailey and co-workers’ method : 0, oil; and e, methyl esters. (In all figures, open points represent the oils and closed points the methyl esters) 0 SAC; Crown Copyright Reserved.660 HAMMONDS et al: USE OF T H E HALPHEN REACTION FOR T H E [,4WaL'yst, VOl. 96 determination of the cyclopropenoid content of rat liver lipid by using a modified Halphen test has also been reported, with methyl sterculate used as a standard.ll A variety of reaction conditions has been used to produce the Halphen response from oils or methyl esters, including heating for 2.5 hours at 110 0C,9J0 or for 1 hour at 48 "C followed by heating for 45 minutes at 108 "C or for 15 minutes at 45 "C followed by heating for 5 minutes at 95 "C and finally for 1 hour at 105 "C.l1 Pentanol has been replaced by pyridine8,ll or by butan01,~~lO and morpholine has been used to improve colour ~ t a b i l i t y .~ ~ ~ ~ 0.400 0.360 0.320 0.280 0.240 0.200 0.1 60 0.120 0.080 0.040 40 80 120 160200 240 280 320 360 400 440 480 520 560 600 Concentration of cyclopropenoid material/pg per 25 ml of solvent Fig. 2. Values with Bornbacopsis glabra oil and its methyl esters by using Bailey and co-workers' method: 0, oil; and m, methyl esters EXPERIMENTAL By using apparatus and conditions similar to those described by Bailey, Pittman, Magne and Skau,g and Levi, Reilich and O'Neill,lO we have founa that absorbances produced by oils of cyclopropenoid content determined by hydrogen bromide titration did not correlate with the absorbances obtained with corresponding concentrations of their methyl esters.This deviation was found to be greatest between Sterculia foetida oil (57 per cent. of cyclo- propenoids) and its methyl ester (Fig. 1). For Bombacopsis glabra oil (34 per cent. of cyclopro- penoids) and kapok (Ceiba pentandra) oil (14.5 per cent. of cyclopropenoids), the deviation was present to a lesser extent (Figs. 2 and 3). Absorbances produced by the methyl esters of different cyclopropenoid-containing oils were found to have good correlation with cyclo- propenoid contents determined by hydrogen bromide titration. These results indicated that the colour development of the Halphen reaction is influenced by the presence of cyclopro- penoid glycerides, and that the use of methyl esters (Fig.4) instead of the parent cyclopro- penoid oils (Fig. 5) results in improved absorbance correlation. 0.200 0.160 0.120 - - - I I I I I I I I I 40 80 120 160 200 240 280 320 360 400 Concentration of cyclopropenoid material/pg per 25 mi of solvent Fig. 3. Values with kapok seed oil and its methyl esters by using Bailey and co-workers' method: A, oil; and A, methyl estersSeptember, 19711 DETERMINATION OF THE CYCLOPROPENOID CONTENT OF LIPIDS 661 0.040 0.34, I \j -a@? Po 1 I I I I I I I - 0.30 I t 5 0.26 0.22 Q, 0.18 u c.’ 0 = A 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 Concentration of cyclopropenoid material/pg per 25 ml of solvent Fig.4. Values with methyl esters by using Bailey and co- workers’ method : o, Sterculia foetida oil esters ; w, Bombacopsis glabra oil esters; and A, kapok seed oil esters 0.360 I 0.320 t 0 0.240 Q- z 0.200 0 A 0.1 10 0.1 00 0.090 $ 0.080 $ 0.07ot t t 0.060 0.050 0.040 0.030 0.020 20 40 60 80 100 120140160 180 Ti me/mi nutes Fig. 6. Colour development and degra- dation: c7 , oil; and v, methyl esters662 HAMMONDS et d: USE OF THE HALPHEN REACTION FOR THE [AfldySt, vol. 96 470 480 490 500 510 520 530 540 550 560 570 Waveiength/nm Fig. 7. Halphen colour with methyl esters: - , with morpholine; and - - - -, without morpholine VARIATION IN COLOUR DEVELOPMENT- Although glycerides were found to reach maximum absorbance at 495 nm after 2 to 2$ hours' heating, methyl esters under the same conditions reached maximum absorbance with about 60 to 90 minutes' heating.The colour became less intense if heating was continued for longer than 100 minutes (Fig. 6). However, the maximum absorbance of methyl esters, which occurred at about 75 minutes, was found not to be reproducible in a specific time. Moreover, the carbon disulphide con- centration was found to have an effect on colour development. By allowing the carbon disulphide to boil off quickly at 110 "C, by using open test-tubes, less intense absorbances were produced, while with the use of a water condenser to retain completely the carbon disulphide the reaction temperature was lowered and the colour development unduly slowed down.The use of a small air condenser filled with 5-mm diameter glass spheres to give a fractionation effect was found to improve colour stability. The most reproducible colour development, however, was achieved by using securely stoppered bottles, similar to those described in the Cottonseed Oil Test, British Standard 684 : 1958,2 with which stabilised colours could be obtained with 90 minutes' heating at 110 "C. 0080 r 0.040 1 4 0.030 ; 0.020 a 0010 20 40 60 80 100 120 140 160 180 200 Tim e/m i n u tes Fig. 8. Colour stability test on methyl esters with morpholine by using universal bottlesSeptember, 19711 DETERMINATION OF THE CYCLOPROPENOID CONTENT OF LIPIDS 663 The addition of morpholine to the reaction mixture was found not to have a noticeable effect upon the Halphen colour spectrum of methyl esters between 485 and 565 nm, but colour stability was improved by its use (Figs.7 and 8). METHOD APPARATUS- Oil-bath-This was regulated at 110 & 1 "C. Universal bottle-Approximately 20-ml capacity with stout glass walls and metal screw- cap fitted with a rubber seal. REAGENTS- distilled analytical-reagent grade methanol. sulphuric acid (spgr. 1.84) in 1 litre of methanol. Sodium methoxide solution, 0.4 N-Dissolve 9.2 g of freshly cut sodium in 1 litre of re- Sulphuric acid - methanol solzdion, 0.5 N-Dissolve 14 ml of analytical-reagent grade Petroleum spirit, boiling range 40 to 60 "C-Free from aromatic compounds. Sodium sulphate, anhydrous-Analytical-reagent grade. Batanol-Analytical-reagent grade. Morpholine solution, 4 per cent.-Dissolve 4 g of general-purpose grade morpholine in 100 ml of analytical-reagent grade butanol.Sulphur solution, 1 per cent. in carbon disulphide-Dissolve 1 g of recrystallised sulphur in 100 ml of analytical-reagent grade carbon disulphide. PREPARATION OF METHYL ESTER- Cyclopropenoid-containing oils were extracted by macerating seeds at room temperature with petroleum spirit, filtering the mixture and removing the solvent with the minimum of heating by means of a rotary evaporator. Methyl esters were prepared by using sodium methoxide in methanol solution12 in the proportion of 25 ml of solution to 1 g of fat. The flask containing the mixture was warmed by immersing it briefly in a hot water bath at 80 "C, and the solution was shaken continuously until a single phase was obtained.The solution was neutralised with 0.5 N sulphuric acid in methanol, petroleum spirit was added and then distilled water to partition the phases. The petroleum spirit phase was separated and the aqueous layer extracted with a small amount of petroleum spirit, which was combined with the first extract. The combined petroleum spirit extracts were washed with water and dried with anhydrous sodium sulphate. The solvent was then removed completely by using a rotary evaporator with a minimum of heating and the esters were stored in a refrigerator. TITRATION OF CYCLOPROPENOIDS- Suitable amounts of methyl esters were weighed and the cyclopropenoid contents were determined at 60 "C by using the Durbetaki titration with hydrogen bromide as described in American Oil Chemists' Society Tentative Method Cd 9-57, but with the glacial acetic acid replaced with 5 per cent.glacial acetic acid in benzene in the preparation of the 0.1 N hydrogen bromide s01ution.l~ The cyclopropenoid contents were expressed as methyl sterculate (molecular weight 308). PREPARATION OF STANDARD SOLUTION- An appropriate amount of the methyl esters was weighed into a 100-ml calibrated flask to produce a stock solution in butanol containing approximately 0.5 mg of cyclopropenoid material per ml of solution. Aliquots of 5 or 10ml were taken and diluted to 100ml with butanol in a calibrated flask. The solutions were stored in a refrigerator. HALPHEN REACTION- The rubber seal of the universal bottle cap was protected by a disc cut from aluminium- foil sheet with a cork borer of appropriate diameter.Suitable aliquots of cyclopropenoid methyl ester standard solution were introduced by pipette into the bottle, and made up to 5 ml with butanol. Next, 0.1 ml of the 4 per cent. solution of morpholine in butanol was added and the contents were shaken before adding 1.0 ml of a solution of 1 per cent. of sulphur in664 HAMMONDS, CORNELIUS AND TAN carbon disulphide. A solution blank with the cyclopropenoid esters omitted was prepared as reference. The caps were securely tightened to prevent loss of solvent, and the bottles were immersed to one third of their depth for 90 minutes in the oil-bath at 110 “C. At the end of the reaction time, the bottles were removed from the bath and allowed to cool at room temperature for about 15 minutes.The solution was transferred to a 25-ml calibrated flask and made up to volume with butanol. The absorbance of the solution was measured at 495 nm in a 1-cm cell, with the blank as reference. RESULTS AND DISCUSSION The results are shown in Fig. 9. The Halphen reaction given by the methyl esters, rather than by the original oils, was found to produce better correlation with cyclopropenoid content determined by titration with hydrogen bromide. The use of a securely stoppered bottle for the Halphen reaction resulted in improved methyl ester colour stability and increased absorbance intensity. The consequent enhance- ment of the sensitivity and improved reliability of the Halphen reaction extends its usefulness for the determination of low cyclopropenoid concentrations or in the analysis of limited amounts of samples. Further, the reaction is valid for cyclopropenoid-containing oils derived from a variety of seed species.0.34 0.30 - E C 0.26 - m cn * 0.22 a 0.18 - - - - +-’ - m - I I I I I I I I I I I 20 40 60 80 100 120 140 160 180 200 220240 260 Concentration of cyclopropenoid rnaterial/pg per 25 ml of solvent Fig. 9. Values with methyl esters by using universal bottle: 0, Sterczklia foetida oil esters; H, Bombacopsis glabra oil esters; and A, kapok seed oil esters We thank Miss G. Felber for experimental assistance, and Mr. G. Shone of Kingston Polytechnic for supplying samples of Sterculia foetida seeds. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. REFERENCES Halphen, G., J . Pharm., Anvers, 1897, 6, 390. British Standards Institution, “Cotton Seed Oil Test,” British Standard 684 : 1958, p. 95. American Oil Chemists’ Society, “Halphen Test,” A.O.C.S. Official Method Cb 1-25. Carter, F. L., and Frampton, V. L., Chem. Rev., 1964, 64, 513. Dijkstra, G., and Duin, H. J., Nature, 1955, 176, 71. Shenstone, F. S., and Vickery, J. R., Poult. Sci., 1959, 38, 1055. Deutschman, A. J., jun., and Klaus, I. S., Analyt. Chem., 1960, 32, 1809. Bailey, A. V., Pittman, R. A,, Magne, F. C., and Skau, E. L., J . Amer. Oil Chem. SOC., 1965,42, 422. Levi, R. S., Reilich, H. G., and O’Neill, H. J., Ibid., 1967, 44, 249. Sheehan, E. T., “Effects of Ingestion of Cyclopropenoid Fatty Acids on Reproduction in the Luddy, F. E., Barford, R. A,, and Riemenschneider, R. W., J . Amer. Oil Chem. SOC., 1960,37, 447. Feuge, R. O., Zarins, Z., White, J. L., and Holmes, R. L., Ibid., 1967, 44, 548. Received January 8tk, 1971 Accepted April 15th, 1971 8 , Nature, 1961, 190, 168. -- Female Rat,” University of Arizona Thesis, 1967 (University Microfilms Inc. 67-12187).

ISSN:0003-2654    

DOI:10.1039/AN9719600659

出版商:RSC    

年代:1971

数据来源: RSC