摘要:

THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYEDITORIAL ADVISORY BOARD'Chairman: H. J. Cluley (Wembley)'L. S. Bark (Salford)R. Belcher (Birmingham)L. J. Bellamy, C.B.E. (Waltham Abbey)L. S. Birks (U.S.A.)E. Bishop (Exeter)L. R. P. Butler (South Africa)'R. M. Dagnall (Huntingdon)E. A. M. F. Dahmen (The Netherlands)A. C. Docherty (Billingham)D. Dyrssen (Sweden)J. Hoste (Belgium)H. M. N. H. Irving (Leeds)H. Kaiser (Germany)M. T. Kelley (U.S.A.)W. Kernula (Poland)'W. T. Elwell (Birmingham)'J. A. Hunter (Edinburgh)' G . F. Kirkbright (London)G. W. C. Milner (Harwell)G. H. Morrison (U.S.A.)'J. M. Ottaway (Glasgow)' G . E. Penketh (Billingham)E. Pungor (Hungary)D. I. Rees (London)'R. Sawyer (London)P. H. Scholes (Sheffield)*W.H. C. Shaw (Greenford)S. Siggia (U.S.A.)A. A. Smales, O.B.E. (Harwell)A. Walsh (Australia)T. S. West (Aberdeen)A. L. Wilson (Medmenham)P. Zuman (U.S.A.)'A. Townshend (Birmingham)' Members of the Board serving on The Analyst Publications CommitteeREGIONAL ADVISORY EDITORSDr. J. Aggett, Department of Chemistry, University of Auckland, Private Bag, Auckland, NEWProfessor G. Ghersini, Laboratori CISE, Casella Postale 3986, 201 00 Milano, ITALY.Professor L. Gierst, Universitb Libre de Bruxelles, Facult6 des Sciences, Avenue F.- D. Roosevelt 50,Professor R. Herrmann, Abteilung fur Med. Physik., 63 Giessen, Schlangenzahl 29, GERMANY.Professor Axel Johansson, lnstitutionen for analytisk kemi, Tekniska Hogskolan, Stockholm, 70,Professor W.E. A. McBryde, Dean of Faculty of Science, University of Waterloo, Waterloo, Ontario,Dr. W. Wayne Meinke, KMS Fusion Inc., 3941 Research Park Drive, P.O. Box 1567, Ann Arbor,Dr. 1. Rubes'ka, Geological Survey of Czechoslovakia, Kostelni 26, Praha 7, CZECHOSLOVAKIA.Professor K. Saito, Department of Chemistry, Tohoku University, Sendai, JAPAN.Dr. A. Strasheim, National Physical Research Laboratory, P.O. Box 395, Pretoria, SOUTH AFRICA.ZEALAND.Bruxelles, BELGIUM.SWEDEN.CANADA.Mich. 48106, U.S.A.Published by The Chemical SocietyEditorial: The Director of Publications, The Chemical Society, Burlington House,London, W1 V OBN. Telephone 01 -734 9864. Telex No. 268001.Advertisements: J. Arthur Cook, 9 Lloyd Square, London, WClX 9BA. Telephone 01 -837 631 5.Subscriptions (non-members) : The Chemical Society Publications Sales Office, Blackhorse Road,Letchworth, Herts., SG6 1 HN.Volume 101 No 11 99 February 1976CO The Chemical Society 197

ISSN:0003-2654    

DOI:10.1039/AN97601FX005

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

ANALAO 101 (1199) 73-144 (1976)ISSN 0003-2654February 1976THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYCONTENTS738691961031111221251281361 40141143ORIGINAL PAPERSAnalytical Optoacoustic Spectrometry. Part 1. Instrument Assembly andPerformance Characteristics-M. J. Adams, A. A. King and G. F. KirkbrightObservations on the Limitation Imposed by Interferences i n Flame Atomic-absorption Spectrometry at High Analyte Concentrations-M. s. Cresser andD. A. MacLeodAn Improved Digestion Method for the Extraction o f Mercury from Environ-mental Samples-Haig Agemian and A. s. Y. ChauThe Application of a Wide-slot Nitrous Oxide - Nitrogen - Acetylene Burner forthe Atomic-absorption Spectrophotometric Determination of Aluminium,Arsenic and Tin in Steels by the Single-pulse Nebulisation Technique-K. C.Thompson and R. G. GoddenThe Determination of Mobile Nitrogen in Steel Using an Ammonium lon-selectiveElectrode-J. B. Headridge and G. D. LongThe Determination o f Substituted Phenylurea Herbicides and Their Impurities inTechnical and Formulated Products by Use of Liquid Chromatography-J. A. Sidwell and J. H. A. RuzickaAn Improved Col u m n-c h romatog rap hic Quantitative lsolat ion o f Diosg en i n andYamogenin from Plant Crude Extracts Prior to Their Determination byInfrared Spectrophotometry-T. M. Jefferies and Roland HardmanRapid Sample Dissolution and Determination of Total Iron in Iron Ore, Sinter,Concentrates and Agglomerates-Om P. BhargavaSpectrophotometric Determination of Free Chlorine in Air-J. Gabbay, (the late)M. Davidson and A. E. DonagiA Colorimetric Method for the Determination of lsoprenaline Sulphate inPharmaceutical Preparations-R. B. Salama and H. A. El-ObeidCO M M U N I CAT1 0 N SA Rapid Method for Detecting Erythrosine in Canned Red Fruits-J. B. Adamsand R. ButlerX-ray Analysis of High-alumina Cement Concrete-C. Plowman and J. GyllenspetzBook ReviewsSummaries of Papers in this lssue-Pages iv, v. viii, ixPrinted by Heffers Printers Ltd, Cambridge, EnglandEntered as Second Class at New York, USA, Post Offic

ISSN:0003-2654    

DOI:10.1039/AN97601BX007

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

i V SUMMARIES OF PAPERS I N THIS ISSUE February, 1976Summaries of Papers in this IssueAnalytical Optoacoustic SpectrometryA simple single-beam spectrometer suitable for the study of optoacousticspectra from small solid samples is described and the design of a suitable samplecell is reported. The performance characteristics of the spectrometer havebeen evaluated using different types of sample. A preliminary assessment ofthe predicted advantages of optoacoustic spectrometry over conventionaltechniques of ultraviolet - visible absorption and reflectance spectrometry forsolid samples has been made.M. J. ADAMS, A. A. KING and G. F. KIRKBRIGHTDepartment of Chemistry, Imperial Collcge, London, SW7 2AY.Part I. Instrument Assembly and Performance CharacteristicsAnalyst, 1976, 101, 73-85.Observations on the Limitation Imposed by Interferences in FlameAtomic- absorption Spectrometry at High Analyte ConcentrationsWhen burner rotation or an absorption line of poorer sensitivity is usedin flame atomic-absorption spectrometric analysis, care must be taken toestablish the absence of fresh or increased interferences a t higher concen-trations of the analyte element.At high concentrations, sulphate was foundto cause severe depressions in the determinations of magnesium, cobalt andnickel, although under normal conditions the interference is negligible. Therisk of substantial error can be reduced either by dilution of samples andstandards, or by taking measurements by using the upper part of a fuel-leanflame that is burning on a slot burner with a triangular cross-section, or byadding a suitable releasing agent.M.S. CRESSER and D. A. MACLEODSoil Science Department, University of Aberdeen, Aberdeen, AB9 2UE.Analyst, 1976, 101, 86-90.An Improved Digestion Method for the Extraction ofMercury from Environmental SamplesAn improved digestion procedure for the extraction of mercury from environ-mental material is reported. The method involves the digestion of the samplea t 60 "C with sulphuric acid - nitric acid (2 + l), containing a trace amountof hydrochloric acid, and subsequent oxidation with permanganate and per-sulphate solutions. With this procedure mercury is successfully recoveredfrom organic matter and resistant inorganic forms such as mercury(I1)sulphide. Unlike digestion with aqua regia, this procedure is simple andsafe, and is applicable to the digestion of a large number of samples simul-taneously. The method can be adapted to the automated cold-vapour andflame atomic-absorption techniques and is therefore ideal for routine monitor-ing.HAIG AGEMIAN and A.S. Y. CHAUCanada Centre for Inland Waters, Water Quality Branch, P.O. Box 5050, 867 Lake-shore Road, Burlington, Ontario, L7R 4A6, Canada.Analyst, 1976, 101, 91-95February, 1976 SUMMARIES OF PAPERS I N THIS ISSUEThe Application of a Wide-slot Nitrous Oxide - Nitrogen - AcetyleneBurner for the Atomic- absorption Spectrophotometric Determinationof Aluminium, Arsenic and Tin in Steels by the Single-pulseNebulisation TechniqueSingle-pulse nebulisation of 10 per cent.m/V iron or steel solutions into anitrogen-diluted nitrous oxide - acetylene flame maintained on a speciallydesigned wide-slot burner is a useful technique for the determination oftin, arsenic and soluble aluminium in iron and steels. Use of this methodavoids the need for prior separation of the analyte. A deuterium lampwas found to be unsatisfactory for measuring the background (non-specific)absorption when determining aluminium and tin, the explanation for which ispostulated.K. C. THOMPSON and R. G. GODDENVShandon Southern Instruments Limited, Frimley Road, Camberley, Surrey, GU165ET.Analyst, 1976, 101, 96-102.The Determination of Mobile Nitrogen in Steel Using anAmmonium Ion- selective ElectrodeAn absorption cell containing an ammonium ion-selective electrode has beenconstructed and used for the determination of mobile nitrogen in steel;this nitrogen is released as ammonia when the steel is heated a t 500 "C ina stream of hydrogen.The cell was used in conjunction with a digital volt-meter and a recorder in order to obtain a continuous record of the progressof the reaction between mobile nitrogen and hydrogen. Results are pre-sented for the determination of 0.0005-0.0108 per cent. of mobile nitrogenin 10 steels using the new equipment and are compared with those obtainedby using a spectrophotometric finish based on indophenol blue. The method,with relative standard deviations of 0.0001-0.0003 per cent., is more precisethan that with the spectrophotometric finish, with relative standard deviationsof 0.0002-0.0006 per cent.J.B. HEADRIDGE and G. D. LONGDepartment of Chemistry, The University, Sheffield, S3 7HF.Analyst, 1976, 101, 103-110.The Determination of Substituted Phenylurea Herbicides andTheir Impurities in Technical and Formulated Products byThe application of liquid chromatography to the identification and deter-mination of the active ingredient and the impurities in phenylurea herbicidescommonly employed in agriculture is described. Technical materials aredissolved in dichloromethane and chromatographed on microparticulatesilica with dichloromethane or dichloromethane - methanol as eluting agent,or on microparticulate silica bonded with octadecyltrichlorosilane withmethanol - water as eluting agent. An initial extraction procedure is requiredfor dispersible powders. Detection was by means of ultraviolet absorbance.J. A. SIDWELL and J. H. A. RUZICKADepartment of Industry, Laboratory of the Government Chemist, Cornwall House,Stamford Street, London, SE1 9NQ.Analyst, 1976, 101, 11 1-121.Use of Liquid Chromatograph

ISSN:0003-2654    

DOI:10.1039/AN97601FP009

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

...V l l l SUMMARIES OF PAPERS IN THIS ISSUEAn Improved Column-chromatographic Quantitative Isolation ofDiosgenin and Yamogenin from Plant Crude Extracts Priorto Their Determination by Infrared SpectrophotometryA previously described routine procedure inxrolving the use of a silica gel columnin determining diosgenin and yamogenin has been improved by using water-containing solvents. The advantages are that there is less variation betweenduplicate results, that each column can be used a t least five times and thatcomponent bands are eluted in predictable volumes of solvents. An apparatusand solvent sequence is described that allows twelve columns to be developedsimultaneously.The method has been successfully applied to crude extracts from Dioscoreadeltoidea tuber and to oily crude extracts from the seeds of Trigonella foenum-graecurn (fenugreek) and Balanites aegyptzaca.The over-all error of the pro-cedure, including sampling, extraction and infrared spectrophotometric deter-mination for duplicate analyses of 2-5-g samples of the fenugreek seed used,expressed as a 95 per cent. confidence interval of the mean sapogenin value,was 1.04 f 0.025 per cent. for diosgenin plus yamogenin, 0.64 f 0.019 percent. for diosgenin and 0.40 f 0.023 per cent. for yamogenin.As the method does not permit the separation of tigogenin from diosgeninnor that of neotigogenin from yamogenin, the results indicate maximumyields for diosgenin and yamogenin in fenugreek seed. The results excludesterols, steryl esters, dihydroxysapogenins and spirostadienes.T.M. JEFFERIES and ROLAND HARDMANFebraary , 19 76Pharmacognosy Group, School of Pharmacy arid Pharmacology, University of Bath,Bath, BA2 7AY.Analyst, 1976, 101, 122-124.Rapid Sample Dissolution and Determination of Total Ironin Iron Ore, Sinter, Concentrates and AgglomeratesA method is described for rapid and complete dissolution of iron ores with awide range of composition in order to determine total iron. The sample isfused with a mixed flux of sodium carbonate and sodium peroxide in avitreous carbon crucible and the melt is dissolved in hydrochloric acid. Ironis then determined by redox titration with dichromate. The wide range ofcomposition of the samples for which the method can be used, as exemplifiedby results obtained for ISO, BCS and NBS reference standards, demonstratesits universal applicability.The method is rapid, free from tedious and time-consuming manipulations, suitable for routine control and yields results com-parable with those obtained by the referee method.OM P. BHARGAVAMetallurgical and Chemical Laboratories, The Steel Company of Canada Limited,Wilcox Street, Hamilton, Ontario, L8N 3T1, Canada.Analyst, 1976, 101, 125-127.Spectrophotometric Determination of Free Chlorine in AirA spectrophotometric method that is suitable for the micro-determination offree chlorine in ambient air is described. When free chlorine is absorbed in analkaline solution of 4-nitroaniline, an orange - brown colour develops, with amolar absorptivity (E) of 19 000 a t 485 nm.The method is suitable for thedetermination of chlorine in the range from a few parts per hundred millionto about 20 p.p.m.Ozone, ammonia and sulphur reducing gases interfere with the reaction,but these interferences can be reduced. Features of the method include highsensitivity, stability of the reagents and coloured reaction product, highcollection efficiency, relative specificity to chlorine, reproducibility, simplicityand convenience.J. GABBAY, (the late) M. DAVIDSON and A. E. DONAGIResearch Institute for Environmental Health Nuisances, Tel-Aviv University,Sackler School of Medicine, Ramat-Aviv, Tel- Aviv, Israel.Analyst, 1976, 101, 128-135February, 1976 SUMMARIES OF PAPERS I N THIS ISSUEA Colorimetric Method for the Determination of IsoprenalheSulphate in Pharmaceutical PreparationsA rapid and convenient colorimetric method is described for the determinationof isoprenaline sulphate in pharmaceutical preparations.This method isbased on measuring the intensity of the orange colour developed when iso-prenaline sulphate is allowed to react with thiosemicarbazide in an alkalinemedium. Beer’s law is obeyed in the concentration range 2.0-16.0 pgml-l.Data on precision and accuracy are presented. As the catecholic functionwith unsubstituted adjacent positions is required for the development of thecolour, the method is highly specific.R. B. SALAMA and H. A. EL-OBEIDDepartment of Pharmaceutical Chemistry, Faculty of Pharmacy, University ofKhartoum, Khartoum, Sudan.Analyst, 1976, 101, 136-139.A Rapid Method for Detecting Erythrosine in Canned Red FruitsCommunicationJ. B. ADAMS and R. BUTLERCampden Food Preservation Research Association, Chipping Campden, Gloucester-shire, GL55 6LD.Analyst, 1976, 101, 140-141.X-ray Analysis of High- Alumina Cement ConcreteCommunicationC. PLOWMAN and J. GYLLENSPETZScientific Services Department, Central Electricity Generating Board, NorthEastern Region, Beckwith Knowle, Harrogate, HG3 1PR.Analyst, 1976, 101, 141-142.i

ISSN:0003-2654    

DOI:10.1039/AN97601BP013

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

FEBRUARY 1976 The Analyst Vol. 101 No. 11 99 Analytical Optoacoustic Spectrometry Part 1. Instrument Assembly and Performance Characteristics M. J. Adams, A. A. King and G. F. Kirkbright Department of Chemistry, Imperial College, London, SW7 2A Y A simple single-beam spectrometer suitable for the study of optoacoustic spectra from small solid samples is described and the design of a suitable sample cell is reported. The performance characteristics of the spectrometer have been evaluated using different types of sample. A preliminary assessment of the predicted advantages of optoacoustic spectrometry over conventional techniques of ultraviolet - visible absorption and reflectance spectrometry for solid samples has been made. In 1881, Alexander Graham Bell1 was able to demonstrate that the illumination of different solid and liquid substances with a rapidly interrupted beam of light resulted in the emission of acoustic energy at the same frequency as that at which the incident radiation was modu- lated. Tyndal12 repeated these observations and also studied gaseous samples; he was unable to observe a quantitative effect with gases, probably owing to the lack of detectors of sufficient sensitivity.With the advent of sensitive microphone detectors, a number of optoacoustic analysers were later reported for use in gas analysis using infrared sources of illumination3p4; for example, this type of equipment permitted the determination of carbon dioxide in air at the parts per million level. More recently, the optoacoustic effect has been used in molecular spectroscopy to measure vibrational relaxation times of gaseous molecule^^^^ and, utilising infrared sources, for the detection of trace levels of atmospheric pollutant gases.’ Although from the earliest work of Bell1 it was evident that some of the strongest opto- acoustic signals were observable from solid materials, it is only recently that the possibilities of developing a new analytical technique, based on optoacoustic spectrometry (OAS), for the direct examination of solid and semi-solid samples have become apparent This paper describes this novel technique, the development of an instrument assembly for optoacoustic spectrometry and our preliminary studies with the instrument utilising small solid samples.Principle of Operation The optoacoustic effect can be demonstrated with a very simple apparatus such as that shown in Fig. 1.When radiation in the visible and near infrared region of the spectrum from a 100-W quartz - iodine lamp is allowed to fall on a suitable absorbing material, e.g., carbon black, contained in a closed system, the energy is absorbed by the sample and, if the material does not luminesce or degrade photochemically, is converted into heat. This conversion of absorbed energy takes place rapidly, as the energy absorbed by electronic excitation of the molecules of the sample may be degraded through the cascade process through lower electronic and vibrational energy levels within 10-8s or less. If the incident radiation is interrupted periodically by use of a rotating sector, the absorption of energy is interrupted at the fre- quency of modulation by the sector and consequently the heat produced in the sample after energy conversion also appears at this frequency.In a closed system of constant volume, therefore, the heating produces a periodic increase in pressure which follows the modulation frequency of the incident radiation. At modulation frequencies between 30 Hz and 20 kHz, this varying pressure gives an acoustic signal whose amplitude can be measured with a simple microphone transducer. With gaseous samples, the periodic pressure change during irradia- tion is detected directly in the gas, whereas with solid samples the optoacoustic signal from the sample is detected “indirectly” by monitoring the periodic pressure change in the gaseous 7374 ADAMS, KING AND KIRKBRIGHT : ANALYTICAL Analyst, Vol.101 Source Q Microphone amp1 if ier Q Speaker I Chopper Fig. 1 . Apparatus for demonstration of the opto- acoustic effect. environment surrounding the sample ; the atmosphere around the gas becomes heated inter- mittently at the modulation frequency by heat transfer at the solid - gas interface. Despite the indirect nature of the detection for solid samples, as mentioned above some of the strongest optoacoustic signals observed have resulted from the observation of the effect in solid samples in this way. Indeed, in the original work of Bel1,l using the sun as the source, distinctly audible signals were detected for various samples using only the unaided ear as the detector. It is apparent, even with the simple apparatus for the demonstration of the optoacoustic effect shown in Fig.1, that the amplitude of the signal is directly proportional to the intensity of the source. In addition, the amplitude of the signal is inversely proportional to the modulation frequency, as at high operating frequencies the radiant energy supplied to the sample per pulse decreases and thus results in less heat energy per pulse being available after degradation to cause the pressure change. I t is also observed that the amplitude of the optoacoustic signal is greatest with samples of large surface area, c.g., fine powders, where the most efficient absorption of radiation by the solid and effective heat transfer to the surrounding gaseous atmosphere is possible. The optoacoustic effect can be observed only when the incident radiation is absorbed by the sample.Thus, if the wavelength of the ultraviolet, visible or near infrared radiation incident upon the sample is varied, the amplitude of the optoacoustic signal observed at a given wave- length will provide a measure of the ability of the sample to absorb the incident radiation, ie., the absorption spectrum of the sample will be obtained. The optoacoustic power spectrum obtained by measurement of the signal amplitude vcysus the wavelength of the incident radiation should therefore resemble the electronic absorption spectrum of the sample and be complementary to the reflectance spectrum. The use of ultraviolet - visible optoacoustic spectrometry in this way for the examination of solid samples should have a number of advantages over conventional optical absorption or diffuse reflectance spectroscopy. These advantages include the following.In contrast to optical absorption or reflectance spectroscopy, where the response of the photomultiplier or photocell detector employed is proportional to the photon flux, in the optoacoustic effect the microphone response is proportional to the optical power absorbed by the sample, i.e., proportional to both the number of photons per unit area per second and to their energy (frequency). A photon at 200 nm can therefore result in twice as much heat energy after absorption, and therefore a proportional intermittent increase in pressure monitored by the microphone, as a photon at 400 nm. I t is therefore a power spectrum that is obtained.This effect should be advantageous in the ultraviolet region, where difficulties can arise in absorption or reflectance spectroscopy if a rapid decrease in output intensity from the continuum source occurs at wavelengths of less than 300 nm. 1 .February, 1976 OPTOACOUSTIC SPECTROMETRY. PART I 75 2. Utilising optoacoustic spectrometry for gas analysis with laser sources, Kreuzer' has shown that, with a simple microphone and detector system, the absorption of W of optical power can give a signal to noise ratio of greater than unity in the optoacoustic signal; an improvement in sensitivity of about 100-fold can be expected upon optimising the instrument and detector parameters. With a suitable continuum source and a monochromator of large aperture used with a spectral band pass of 10 nm, a power of W should be attainable for sample illumination at all wavelengths in the visible and ultraviolet *region.The detection of materials with high absorptivities at low concentrations, or as very small samples, should therefore be possible. Alternatively, satisfactory spectra should be measurable for materials with very low absorptivities. A low absorptivity may arise as a result of the low oscillator strength of the transitions involved or as a result of the highly reflecting nature of the sample surface. I t can thus be expected that the technique should permit a higher detection sensi- tivity for fine powders and crystalline samples than is attained in reflectance or absorption spectroscopy and that spectra should be obtained when only very small amounts of samples are available.Although radiation must be absorbed by the sample in order to obtain optoacoustic signals, unlike conventional absorption spectroscopy there is no need to detect radiation transmitted by the sample. Unlike reflectance spectroscopy, as only absorbed energy is detected, no problems arise from scattered source radiation as an acoustic rather than an optical transducer is employed. The difficulties encountered in diffuse reflectance spectro- scopy owing to the variation in the relative contributions to the signal of specular reflectance (due to scattered radiation) and diffuse reflectance (due to radiation that has undergone some absorption) as the particle size varies should not be observed in optoacoustic spectrometry.The disadvantage encountered with fine powders of weakly absorbing species in diffuse reflectance spectroscopy, that only a small dffusely reflected component appears in the spectrum owing to failure of the incident radiation to penetrate the sample, should be offset in optoacoustic spectrometry by the ability to detect weak absorption effects and the increase in signal amplitude resulting from improved heat transfer efficiency between the sample and its gaseous environment as the particle size is reduced. For strongly absorbing powder samples, the latter effect should result in very strong optoacoustic signals for samples of very small particle size. Additionally, the freedom from problems with scattered light should make optoacoustic spectrometry suitable for the examination of other types of sample, such as hard and soft biological tissues, fibres and metallurgical samples. For small solid samples, high absolute sensitivity should result without resort to the microscope illumination and viewing optics required in the examination of such samples by optical absorption or reflectance spectroscopy.Although the optical arrangement used to illuminate the sample is still important, owing to the need to ensure maximum energy absorp- tion from the source, the requirement of wide solid-angle viewing of the sample by the detector should not be as important in optoacoustic spectrometry as in optical spectroscopy. In a sample cell of constant volume, it is necessary only to arrange to minimise the volume of gas in the cell in order to create the maximum pressure change for a given amount of energy transferred to it, and then to monitor this pressure change efficiently.The technique should provide valuable information in the study of materials that are fluorescent or phosphorescent. In these instances, part of the absorbed energy is re-emitted radiatively and is not degraded into thermal energy by radiationless transitions as for most absorbing species. Hence, in the optoacoustic spectrometric power spectra of such materials, the absorption bands of longest wavelength that normally give rise to fluorescence or phos- phorescence will be attenuated relative to those absorption bands at which radiationless transitions are responsible for de-excitation of the excited state.The technique should therefore prove complementary to spectrofluorimetry and phosphorimetry in the study of luminescence phenomena and for quantum efficiency measurements, 6, as it measures the complementary radiationless loss (1 -6). As the observation of optoacoustic signals for solid samples depends on the transfer to the surrounding gaseous environment of energy released after absorption of radiation, optoacoustic spectrometry may prove useful in studies of heat transfer efficiency at solid - gas interfaces and in thermal conductivity measurements. This paper reports the design and assembly of a simple single-beam spectrometer suitable for the study of the optoacoustic spectra of solid samples. The primary criteria for the 3. 4.5 . 6.76 ADAMS, KING AND KIRKBRIGHT ANALYTICAL Analyst, Vol. 101 successful design of the sample cell for use in optoacoustic spectrometry have been established and the performance characteristics of the spectrometer evaluated by using simple types of sample. A preliminary assessment of the above potential advantages of optoacoustic spectro- metry over conventional techniques of optical spectroscopy bas been made. Experimental Apparatus The single-beam optoacoustic spectrometer assembly constructed for this work is shown diagrammatically in Fig. 2. A 1 000-W mercury - xenon high-pressure continuum source (Oriel Corp., Stamford, Conn., USA, Model 6293) was employed. This source was operated in an air-cooled housing fitted with a UV-grade fused silica double-element condensing lens assembly and a spherical rear reflector.This optical arrangement provided forf/l .O collection efficiency for radiation from the arc lamp. Chopper Monochromator OAS Ref. Arnp./P.S.D. ‘-1 1-1 Recorder Fig. 2. Schematic diagram of the single-beam optoacoustic spectrometer. The source radiation was focused on to the blades of a rotating chopper (Brookdeal Elec- tronics Ltd., Bracknell, Berks., Model 9749) whose chopping frequency was variable between 1 and 1 000 Hz and which also generated a reference signal for the detector electronics at the chosen frequency by means of a simple source - photodiode assembly. The modulated radiation was collected by using a silica biconvex lens (40 mm diameter, 60 mm focal length) and focused on to the entrance slit of the grating monochromator employed.A smallf/4 monochromator (Farrand Optical Co., New York) fitted with a plane grating (50 x 50 mm, 600 lines mm-1) and interchangeable slits and a wavelength scanning motor were mounted on the optical rail of the system. The reciprocal linear dispersion obtained at the exit slit of the monochromator was 10 nm mm-l. The optoacoustic sample cell and microphone transducer assembly were mounted directly a t the exit slit of the monochromator. The sample cell employed is shown diagrammatically in Fig. 3. The cell was machined directly from a single piece of stainless steel and took the form internally of a cylinder (25 mm diameter, 25 mm long), normal to which a silica window of 1 mm thickness was mounted and held in position on the cell body by a knurled and threaded cap and a PTFE compression gasket.The other side of the cylinder opposite the window was employed to locate the sample holder. This took the form of a polished stainless-steel tray (10 mm diameter, 0.75 mm thick) mounted on a screw mechanism so as to provide for fine adjustment of the positioning of the sample. One end of the cylindrical cell body was employed to locate the condenser microphone transducer. The microphone diaphragm effectively formed part of the cell wall and was mounted in a pressure-tight fitting to the body of the cell. A pressure release valve was also located in the side-wall of the cell body; this served to allow insertion of the sample without an increase in internal pressure in the cell. The whole assembly formed a pressure- tight cell; the cell walls were polished in order to reduce energy losses and stainless steel was employed in order to provide low thermal conductivity.In the cell design employed in thisFebruary, 1976 OPTOACOUSTIC SPECTROMETRY. PART I 77 work, and with the optical geometry used, the sample tray receives maximum irradiation from the source only when it is in the vertical position. For the purposes of this work, therefore, milligram powder samples were mounted on the sample tray utilising small pre-cut discs (7 mm diameter) of transparent double-sided adhesive tape. The transducer used was a 1-in diameter condenser microphone (Bruel and Kjaer, Model 4144) of sensitivity 5 mV pbar-l, operated at a charging voltage of 200 V from a dry battery supply.\ Microphone 1 Window \ I Pressure release valve holder ;'amp'e Fig. 3. The optoacoustic cell. The optoacoustic signal from the microphone transducer was taken directly to a sensitive lock-in amplifier (Princeton Applied Research Corp., Model 186), which utilised the reference signal generated at the variabie-speed chopper to extract the signal waveform and present this as a d.c. potential to a potentiometric chart recorder. Optoacoustic spectra were then obtained by recording the output of the lock-in amplifier versus the wavelengths of the incident radiation at the sample cell during wavelength scanning of the monochromator. An Ideal Model for Prediction of the Strength of Optoacoustic Signals The optoacoustic effect can be observed only after absorption of radiation by the sample.For a simple model system, for example a single absorbing particle or thin absorbing layer of homogeneous composition, the power absorbed can readily be calculated from the classical relationships that govern the absorption of electromagnetic radiation by matter. In order to obtain an estimate of the optoacoustic signal power, it is then necessary to quantify the fraction of this absorbed radiation that is lost by the excited state in its return to the ground state by radiationless transitions. Thus, Efficiency of radiationless Optoacoustic signal - - Power absorbed power (W) from source (W conversion For a simple absorbing layer of thickness I of a species of molar absorptivity E and concen- tration c in the sample matrix, the power absorbed at wavelength A, Pabd, is given by Beer's law as Pabs = Pox - [POAexp(-2.3~&)] = POX [l-exp(-2.3~~cZ)] .. .. .. * . (1) where Pox is the power of the incident radiation at wavelength A. The optoacoustic signal power, Poasx (watts), is then obtained by multiplying Pabd by an78 ADAMS, KING AND KIRKBRIGHT: ANALYTICAL Analyst, Vol. 101 efficiency factor, /I, which is a measure of the conversion efficiency of absorbed power to heat by radiationless processes, i.e. PoAsh = Pox [l-exp(--2,3~cZ)]/3 . . For low absorption conditions, where the power absorbed is small, the higher terms in the expansion of l-exp(-2.3~cZ) can be neglected, and we can write From this expression, it is evident that the optoacoustic signal power under ideal conditions is directly proportional to the available power of the source at a given wavelength, the molar absorptivity of the absorbing species at this wavelength (E) and the concentration, path length and power conversion efficiency. It can be seen that when the optoacoustic signal is plotted against wavelength for a source whose output power does not vary with wavelength, the spectrum obtained will give a direct measure of the variation of E with wavelength, i e ., the electronic absorption spectrum will be obtained. I t is also apparent from equation (3) that for weakly absorbing species, the amplitude of the optoacoustic signal will be directly proportional to the concentration of the absorbing species, so that linear calibration graphs can be expected. In addition, the direct proportionality between PoAsh and the absorption path length ( I ) will affect the manner in which the amplitude of the signal varies for strong and weakly absorbing species with sample thickness and/or particle size (for powdered samples).For absorbing species that luminesce by fluorescence or phosphorescence, where a fraction of the absorbed energy is lost by radiative transitions rather than radiationless internal conversion and collisions, the optoacoustic signal power expected might be written as . . .. where 4 is the luminescence quantum efficiency, which is complementary to p. In a real experimental system, a number of factors relating to the instrument arrangement employed and to the characteristics of the sample will lead to modification of the ideal expres- sion of equation (3).Thus, for the continuum source used in optoacoustic spectrometry, the power incident upon the sample, POL, is given by . . . . . . where Ioh W cm-2 nm-l, is the spectral irradiance of the source, o cm is the monochromator slit width, H is the slit height, s nm is the spectral band width employed, TA is the transmittance of the monochromator at wavelength h and CI sr is the solid angle, for sample illumination available from the particular source and monochromator assembly. Thus the optoacoustic signal power becomes In contrast to fluorescence or reflectance spectroscopy, in which the re-emitted or reflected optical power received at the radiation detector is further dependent on the optical viewing geometry (solid angle), in optoacoustic spectrometry in the ideal case, all of the absorbed power that appears as kinetic energy after radiationless de-excitation is available to cause an increase in pressure in the cell and to generate the optoacoustic signal at the microphone transducer.In practice, however, a power transfer efficiency from sample to transducer of unity is difficult to achieve owing to loss of energy at the cell walls. For a real experimental system, therefore, it is necessary to introduce an additional power transfer efficiency factor, a, into equation (6) in order to account for this loss and for any loss in efficiency at the transducer. In addition, for solid samples, the optoacoustic signal power may be expected to vary with particle size, for two reasons: (a) when the particle size (or surface area) varies, the power absorbed will vary owing to changes in the average path length for absorption and the reflec- tivity of the sample, and (b) as the particle size (surface area) varies, the efficiency of power transfer to the surrounding gas may vary.The power transfer at the solid - gas interface may also be influenced by the thermal conductivity of the sample and the filler gas; this wouldFebruary , 1976 OPTOACOUSTIC SPECTROMETRY. PART I 79 then also be expected to affect the manner in which the observed optoacoustic signal varies with sample characteristics. The amplitude of the signal would also be expected to be inversely proportional to the heat capacity of the filler gas. As part of our preliminary studies on optoacoustic spectrometry, we have undertaken a number of experiments with the instrument assembly described here in order to investigate the predicted variation in optoacoustic signal strength with the sample and instrument characteristics outlined above.Results and Discussion In order to evaluate the performance characteristics of the optoacoustic spectrometer, and the fundamental instrument and sample parameters that govern the analytical application of the optoacoustic effect to solid samples, it was necessary initially to choose a “standard” sample of well defined absorption characteristics whose optoacoustic spectrum could be studied. Carbon black was chosen for this purpose, as it is readily obtainable, can be em- ployed as a black-body reference absorber of known absorptivity at different wavelengths and is a strong absorber for which optoacoustic signals are easily attainable.This material was therefore used to demonstrate the relationships between the optoacoustic signals obtained and the instrument and sample parameters. Optoacoustic Spectrum of Carbon Black With the instrument assembly described above, the amplitude of the optoacoustic signal versus the excitation wavelength from the mercury - xenon source was obtained for particulate carbon black (particle size less than 76 pm) utilising a source modulation frequency of 86 Hz and a broad spectral band pass at the monochromator exit slit (40 nm). The optoacoustic spectrum obtained is shown in Fig. 4, A. An optoacoustic signal, caused by absorption of the source radiation by the sample, was observed a t all wavelengths between 200 and 650nm selected by the monochromator.The optoacoustic spectrum exhibits broad peaks at those wavelengths which correspond to the intense mercury line emission from the source super- imposed on the continuum background of xenon. Fig. 4, B, shows the spectrum obtained when a thermocouple detector was positioned at the monochromator exit slit in place of the optoacoustic sample cell and microphone transducer. In this way, the variation with 600 500 400 X/nm 300 Fig. 4. A, the optoacoustic spectrum of particulate ( < 7 6 pm) carbon black; and B, the power spectrum from the 1000-W mercury - xenon lamp employing a thermopile detector. wavelength of the integrated incident power per unit time (in millijoules) from the source was measured. The source power spectrum obtained is seen to be virtually identical with the optoacoustic spectrum of the carbon black and clearly indicates that power spectra are80 ADAMS, KING AND KIRKBRIGHT : ANALYTICAL Analyst, Vol.101 obtained by optoacoustic spectrometry. These results also confirm the suitability of carbon black as a reference absorption material whose optoacoustic spectrum can be used to correct the observed spectra of other samples for the variation of source power with wavelength. This correction procedure for the optoacoustic spectra can be effected either manually, as with the spectra obtained in .the work reported here with the single-beam spectrometer, or automatically in a double-beam spectrometer in which the carbon black absorber is incor- porated into the reference beam.Effect of Source Power and Modulation Frequency The predicted proportionality between the amplitude of the optoacoustic signal and the incident source power was readily confirmed using a carbon black sample by varying the operating current of the mercury - xenon lamp source. The results illustrated in Fig. 5 were obtained utilising a modulation frequency of 85 Hz and a peak excitation wavelength of 570 nm selected at the monochromator. 35 ?? O 3 0 2i J E 25 Fig. Acoustic signal 5. Variation in the amditude of the optoacoustic signal for carboh black a t 570 nm with source lamp current. The variation of the amplitude of the optoacoustic signal with the source modulation frequency was studied for carbon black (particle size less than 76 pm) at several different excitation wavelengths over the frequency range 10460 Hz. Essentially similar results were ‘obtained at each excitation wavelength employed. The variation of the optoacoustic signal amplitude at 570 nm with modulation frequency is shown in Fig.6, (a). As expected, the signal is inversely proportional to the modulation frequency and the graph of signal amplitude versus the reciprocal of the frequency is linear [see Fig. 6, ( b ) ] . This relationship follows from the direct proportionality obtained between signal amplitude and source power. 0 200 400 600 800 0 100 200 300 400 500 Modulation frequency/Hz i i f ( x 104) Fig. 6. (u), Variation in the amplitude of the optoacoustic signal for carbon black with modulation frequency; and ( b ) , variation in the amplitude of the optoacoustic signal for carbon black with the reciprocal of the modulation frequency.February , 1976 OPTOACOUSTIC SPECTROMETRY. PART I 81 Thus, as the modulation frequency increases, the power per pulse available for absorption by the sample decreases.All further studies reported here were conducted at modulation frequencies of less than 200 H z , in order to maintain an adequate signal amplitude and to optimise the signal to noise ratio. Effect of Nature of Filler Gas Used in Sample Cell As optoacoustic signals for solid samples are obtained “indirectly” by monitoring the periodic pressure change in the gaseous environment surrounding the sample in the constant-volume cell, it might be expected that the amplitude of the optoacoustic signals observed would depend on the physical properties of the gas in the sample cell.The particular properties of interest are the heat capacity of the filler gas at constant volume (C,) and its thermal con- ductivity. The effect of using different filler gases on the spectrum observed for carbon black in the region 500-650 nm was therefore investigated. The spectra shown in Fig. 7 were obtained for a carbon black sample in a closed sample cell that contained helium, argon, nitrogen or carbon dioxide at atmospheric pressure. The corresponding heat capacity and thermal conductivity values for these gases are shown in Table I. It might be expected that He Ar Xlnm Fig. 7. Variation in the optoacoustic signal for carbon black with the nature of the cell filler gas.the use of a filler gas of high thermal conductivity would lead to efficient heat transfer at the sample - gas interface and that this would be beneficial to the amplitude of the optoacoustic signal but that correspondingly greater energy losses to the cell walls might then offset this effect. At the low source modulation frequencies employed in this work, the results obtained indicate that the rate of heat transfer at the sample - gas interface does not limit the amplitude of the optoacoustic signal observed ; with the stainless-steel cell employed, no relationship is observed between the thermal conductivity of the filler gas and the amplitude of the signals obtained. A comparison of the results shown in Fig.7 with the heat capacities of thegases employed, however, reveals an apparent correlation: as the heat capacity of the gas increases, the amplitude of the optoacoustic signal decreases. This results from the fact that, under otherwise constant conditions, the energy absorbed by the sample can produce only a small change in temperature and pressure (and therefore a small signal amplitude) in the constant volume of gas in the cell when a gas of large heat capacity is employed. TABLE I HEAT CAPACITIES (C,) AND THERMAL CONDUCTIVITIES (K) OF GASES Gas C,/J K-l mol-1 klW m-1 K-I Helium . . .. .. . . 12.48 Argon .. . . .. . . 12.58 Nitrogen . . .. .. .. 20.70 Carbon dioxide . . .. . . 28.27 0.150 0.017 0.025 0.01682 ADAMS, KING AND KIRKBRIGHT ANALYTICAL Analyst, VoL.101 It is apparent that no significant practical advantage or significantly higher sensitivity would accrue from the use of filler gases other than air with the sample cell and detector assembly used in this work. Effect of Particle Size of Sample The effect of the particle size of the sample for a strongly absorbing material was investi- gated, utilising samples of carbon black and a source modulation frequency of 85 Hz and examining the optoacoustic spectrum in the region. 500-650nm. Fig. 8 shows the spectra obtained using a set of standard test sieves for three particle size ranges: A, less than 76 pm, B, 76-150 pm and C, 150-250 pm. It is clear that the amplitude of the optoacoustic signal increases at all wavelengths in the range examined as the particle size of the sample decreases.This effect is not a mass effect, as the sample mass increases as the ?article size increases. - m m an 0 3 .- .- + 8 a 570 Wavelengthhm Fig. 8. Effect of particle size of the sample on the amplitude of the optoacoustic signal for carbon: A, <76pm; B, 76-150pm; and C, 150-250 pm. 570 Wavelengthhm Fig. 9. Effect of particle size of the sample on the amplityde of the hydrated copper(I1) sulphate optoacoustic signal : A, <76 pm ; B, 76-150 pm; and C, 150-250 pm. As the surface area increases, the power absorbed (which alone gives rise to the optoacoustic signal) thus also increases. The implication is that as the particle size decreases the absorbed power increases, owing to an increase in the mean effective absorbing path length [I in equation (3)].I t is also possible that the power transfer efficiency at the solid - gas interface increases as the particle size decreases. The variation of the amplitude of the optoacoustic signal with particle size was also investi- gated for a weakly absorbing species. Crystalline hydrated copper( 11) sulphate was employedFebruary, 19 76 OPTOACOUSTIC SPECTROMETRY. PART I 83 in these experiments using the same particle size ranges as those of the carbon black samples. Fig. 9 shows the uncorrected optoacoustic spectra obtained in the wavelength range 550- 650 nm for these samples. The signal amplitude is again seen to increase as the particle size decreases, although the effect observed is not as pronounced as with carbon black.It is apparent that for powdered samples, control of particle size will be important in quanti- tative applications to optoacoustic spectrometry. The effect on the analytical signal of variation in particle size, however, is less complex than that encountered in diffuse reflectance spectroscopy. In the latter technique it has commonly been observed that either an increase or decrease in the magnitude of the observed reflectar-e at different wavelengths may be obtained, depending on whether the sample is a weak or strong absorber. The complicating factor is that both the diffuse reflectance component of the measured intensity (which has experienced absorption by the sample) and the specular (non-absorbed) reflectance component may vary independently with particle size and the molecular absorptivity for the species studied at different wavelengths ; the measured reflectance, however, indicates only the observed net effect for both of these components, Thus the effect of particle size and sample absorptivity in optoacoustic spectrometry is fundamentally simpler to interpret owing to the fact that it is only the portion of the incident radiation that is absorbed that gives rise to the optoacoustic signal ; no complications arise from the specular (non-absorbed) reflected radiation, which is not measured.Effect of Amount of Sample and Concentration of the Absorbing Species The variation of signal amplitude with the sample size was investigated for a carbon black sample of particle size less than 76pm. The carbon powder was spread uniformly on to a 7 mm diameter disc of double-sided clear adhesive tape.The mass of the sample was deter- mined by weighing the adhesive tape with and without the powder sample. The amplitude of the optoacoustic signal was measured for this sample in the wavelength range 500-650 nm. The mass of the sample was then progressively decreased by removing sections of the 7-mm disc of tape and re-weighing. The optoacoustic signal was recorded in each instance. Fig. 10 shows the variation in signal amplitude at 570 nm with sample mass. A linear relationship is obtained; as the sample mass and surface area decrease, the power absorbed from the incident beam of radiation becomes less so that the optoacoustic amplitude is decreased. The minimum detectable mass of sample for carbon black, based on an estimation of the signal to noise ratio observed for the small background signal produced by the stainless-steel sample holder, was about 10 pg.The effect of the concentration of the absorbing species was investigated for a strong absorber using milligram samples of mixtures of carbon black in magnesium oxide of particle size less than 76 pm. Fig. 11 shows the variation in the amplitude of the optoacoustic signal at 570 nm with the concentration of carbon black in magnesium oxide covering the range O-lOO%. The graph obtained is linear at low concentrations, i.e., under conditions of low power absorp- tion, but deviates towards the concentration axis at high concentrations when the simple form of equation (3) is not valid.0 2 4 6 8 Mass of carbon ( X 104)/g 1 1 I I I 0 20 40 60 8 0 1 Carbon in MgO, % 0 Fig. 10. Variation in the amplitude of the Fig. 11. Calibration graph for mixtures optoacoustic signal with the mass of sample, for carbon ( < 7 6 pm) a t 570 nm. of carbon black and magnesium oxide.84 ADAMS, KING AND KIRKBRIGHT : ANALYTICAL Analyst, Vol. 101 Mnm Xlnm Fig. 12. (a) Optoacoustic spectra and (b) Fig. 13. Optoacoustic spectrum reflectance spectra of glossy red and blue of a 1 + 9 cadmium sulphide- papers. magnesium oxide mixture. Preliminary Optoacoustic Spectra of Different Types of Sample Optoacoustic spectrometry shows considerable promise for the characterisation md quanti- tative determination of a wide range of molecular species present in different types of solid samples.Investigation of the analytical application of optoacoustic spectrometry to materials of inorganic and biological origin is at present in process; the results of these studies will be reported in later papers. In order to demonstrate that “real” spectra are obtained, and that these spectra give information relating to the molecular absorption characteristics of samples, however, some representative spectra of different samples are shown in Figs. 12-14. The I 1 I 3 550 450 350 Wnm Fig. 14. (a) Optoacoustic spectrum and (b) reflectance spectrum of a 1 + 9 chromium(II1) oxide - magnesium oxide mixture,February, 1976 OPTOACOUSTIC SPECTROMETRY. PART I 85 spectra shown have been manually corrected for variation in output power of the source with wavelength using the optoacoustic spectrum of carbon black as reference.Fig. 12, (a) and ( b ) , shows the spectra of samples of highly reflecting artists’ gummed paper. These spectra are compared with the corresponding reflectance spectra and are clearly virtually identical. Although the highly reflecting nature of the papers gave rise to relatively weak absorption, and some degradation in the signal to noise ratio was observed in the reflectance spectra, the optoacoustic signal for these samples was readily measured. Fig. 13 shows the optoacoustic spectrum of a 1 + 9 (m/m) mixture of cadmium sulphide and magnesium oxide. Cadmium sulphide is a direct-band semiconductor whose band-edge occurs at 2.4 eV.* The band-edge measured from the optoacoustic spectrum occurs at 500 nm, which corresponds to 2.48 eV; this value is thus in fair agreement with the literature value and with a previous value obtained by optoacoustic spectrometry by Rosencwaig.lo Fig. 14 shows the opto- acoustic and reflectance spectra of chromium( 111) oxide powder. The resolution attainable is similar to that obtained in earlier measurements by Rosencwaig and the Cr3+ ion crystal field bands are readily observed. The, sample size taken for the optoacoustic measurements was 100 pg of chromium(II1) oxide, &luted to 1 mg with magnesium oxide, whereas the attachment for the spectrophotometer employed to obtain the reflectance spectra required a sample size of about 750 mg. Modifications to the simple optoacoustic spectrometer described here so as to permit the automatic correction of the optoacoustic spectra for variation of the source power with wave- length are at present in progress; this improved system, and the results of its application to different types of sample, will be described in a later paper. We are grateful to the Analytical Chemistry Trust of the former Society for Analytical Chemistry (now the Analytical Division of the Chemical Society) for the award of a Student- ship to one of us (A.A.K.) and to the Royal Society for an equipment grant for the assembly of the spectrometer . 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Bell, A. G., Phil. Mag., 1881, 11, 510. Tyndall, J., Proc. R. SOC., 1881, 31, 307. Pfund, A. H., Science, N.Y., 1939, 90, 326. Luft, K . F., 2. Tech. Phys., 1943, 24, 97. Read, A. W., Adv. Molec. Relaxation Processes, 1967, 1 , 257. Cottrell, T. L., Macfarlane, I. M., Read, A. W., and Young, A. H., Trans. Faraday SOC., 1966, 62, Kreuzer, L. B., J . .4ppl. Phys., 1971, 42, 2934. Harshbarger, W. R., and Robin, M. R., Accts Chewz. Res., 1973, 6, 329. Pankove, J. I., “Optical Process in Semiconductors,” Prentice Hall, Englewood Cliffs, N.J., 1971. Rosencwaig, A., Analyt. Chem., 1975, 47, 592A. 2655. Received October 24th, 1975 Accepted December l s t , 1975

ISSN:0003-2654    

DOI:10.1039/AN9760100073

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

86 Analyst, February, 1976, Vol. 101, pfi. 86-90 Observations on the Limitation Imposed by Interferences in Flame Atomic-absorption Spectrometry at High Analyte Concentrations M. S. Cresser and D. A. MacLeod Soil Science Department, University of Aberdeen, Aberdeen, AB9 2 UE When burner rotation or an absorption line of poorer sensitivity is used in flame atomic-absorption spectrometric analysis, care must be taken to establish the absence of fresh or increased interferences at higher concen- trations of the analyte element. At high concentrations, sulphate was found to cause severe depressions in the determinations of magnesium, cobalt and nickel, although under normal conditions the interference is negligible. The risk of substantial error can be reduced either by dilution of samples and standards, or by taking measurements by using the upper part of a fuel-lean flame that is burning on a slot burner with a triangular cross-section, or by adding a suitable releasing agent.An inherent limitation of atomic-absorption spectrometry is the incidence of curvature of calibration graphs a t high concentrations of analyte elements. In practice, the useful working range of the technique is often extended by the use of an alternative absorption wavelength, or, when flame atomisation is employed, by rotating the burner to give a shorter absorption path length. We have recently found in our laboratory, however, that when magnesium is determined over the range 0-15 pg ml-l, using burner rotation to attain higher (poorer) sensitivity, very substantial interference was encountered from sulphate, although under normal working conditions, over the concentration range 0-2 pg ml-l, the interference from this anion was negligible in the air - acetylene flame.The depression of calcium absorbance by phosphorus is known to decrease with decreasing calcium concentration, a fact which leads to increased curvature of calibration graphs in the presence of phosphate.' Very little work has been carried out on the effect of the concentration of the analyte element on the incidence and extent of interferences for other elements, however. According to Aldous and Reynolds,2 the sensitivity of the determination of magnesium is slightly diminished when sulphate, rather than chloride or nitrate, is used to prepare mag- nesium standards for atomic-absorption analysis ; most books on atomic absorption do not mention this possible interference.As we observed a very substantial depression at higher magnesium concentrations, it was decided to investigate this interference further, and to search for further instances of enhanced or fresh interferences occurring at high analyte concentrations. Experimental Apparatus The instruments used were : an EEL, Model 240, equipped with either a conventional 100-mm air - acetylene burner or a laboratory-built 50-mm brass slot burner with a triangular cross-section, constructed as shown in Fig. 1, and a Shandon Southern Instruments A3400 with a standard air - acetylene burner. A tomic-absorption spectrophotometers. Standard Solutions Solutions containing lo00 pg ml-l of magnesium, nickel and cobalt were prepared from the analytical-reagent grade chlorides and sulphates.The magnesium solutions were stan- dardised by complexometric analysis before further dilution. The magnitude of the effects observed was so substantial that standardisation of the other stock solutions was unnecessary. Results and Discussion For the EEL, Model 240, instrument, the effects of burner type, fuel flow-rate (at a constant air flow-rate for both burners) and height of measurement on the change in mag- nesium absorbance at 285.2 nm caused by nebulising 10 pg ml-' of magnesium, as the sulphateCRESSER AND MACLEOD 87 , Side , A , End 1 , 1 1 1 0 1 2 3 4 cm Fig. 1. Design of burner with triangular cross-section. rather than the chloride, are shown in Fig.2. Both burners were rotated through an angle of 90" in order to increase the sensitivity for 1 per cent. absorption. The interference became greater for both burners as the ratio of fuel to oxidant was increased, or as measurements were made at lower heights in the flame. The interference was significantly reduced when the burner with a triangular cross-section was used, and with this burner could be eradicated completely over a wide range of burner heights and fuel to oxidant ratios. -60 -80 1 I I 1 I 0.9 1 .o 1.1 1.2 1.3 Fuel flow-rate/l min" Fig. 2. Effect of fuel flow-rate on the change in magnesium absorbance caused by using sulphate in place of chloride at various heights above a flat-topped, 100-mm burner (broken lines) and a triangular cross-section, 50-mm burner (solid lines), with burners rotated through an angle of 90".Air flow-rate, 6.6 1 min-1 for both burners.88 CRESSER AND MACLEOD: INTERFERENCES IN FLAME Analyst, VoZ. 101 The extent of the interference varied significantly with magnesium concentration : it was negligible at 2 pg ml-1 of magnesium, but rose to a 30 per cent. depression for 15 pg ml-l of magnesium under the routine operating conditions used in our laboratory (i.e., wavelength, 285.2 nm; air flow-rate, 6.5 1 min-1; acetylene flow-rate, about 1 1 min-l; and burner height about 4 mm). The atomisation of magnesium sulphate proceeds, at least in part, via the formation of magnesium oxide. Magnesium sulphate decomposes at 1397 K, whereas the oxide melts and boils at 3073 and 3873 K, respectively. The chloride, on the other hand, boils at 1685 K, so that while the chloride is readily volatilised, and hence atomised, the sulphate tends to form stable oxide particles, the size of which depends primarily upon the droplet size distribu- tion produced by the nebuliser, and the magnesium concentration in solution. Larger particles, which Secure the magnesium atoms more efficiently, are formed as the magnesium concen- tration is increased.It should be emphasised that as the droplet size distribution varies between instruments and between nebulisers, the precise analyte concentration at which such an interference becomes significant will also vary between instruments. The effect and trends described above were still observed when an A3400 atomic-absorption spectrophotometer was used in place of the EEL, Model 240, instrument, for example, but the extent of the effect was much smaller.Even when measurements were made at a low height in a fuel-rich flame, the use of magnesium sulphate instead of the chloride reduced the absorbance by only about 20 per cent. If, however, the impact bead was displaced from its normal, optimised position, the effect became more pronounced, which indicates that a finer mist is normally produced by the A3400 nebuliser, although the higher oxidant and fuel flow-rates used with this instrument may also contribute to the improvement in selectivity. Magnesium calibration graphs for the concentration ranges 0-15 and 0-25 pg ml-l exhibited far greater curvature when sulphate rather than chloride solutions were employed, as would be expected under these conditions.Typical curves for the EEL, Model 240, instrument, are shown in Fig. 3. 0.6 0.4 0.1 /+’ / I+’ / 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5 10 15 Magnesium/pg ml-’ Fig. 3. Magnesium calibration graphs at 202.5 nm: broken line, MgC1,; and solid line, MgSO,. The advantages of the burner with a triangular cross-section must arise from the fact that burners of this design give a stiffer, narrower flame, because of the different pattern of air entrainment, The burner with a triangular cross-section required a slightly higher acetyleneFebruary, 1976 ATOMIC-ABSORPTION SPECTROMETRY AT HIGH ANALYTE CONCENTRATIONS 89 flow-rate in order to give a luminous flame, which indicates that more air is entrained by the stiffer flame.This effect would make the flame leaner and hotter, and could result in the observed decrease in the extent of the interference when this burner was used. The increase in the extent of the inter!erence with burner rotation is probably attributable to the effects of lateral diffusion interferen~e,~-S which might be expected to occur under these conditions, although they are not normally observed in air - acetylene flames. A brief in- vestigation of the absorption profiles showed a small but significant effect in this instance, but it would be difficult to relate it quantitatively to the increase in interference when burner rotation is used. It was found that the interference effect of sulphate on the absorbance of magnesium could readily be overcome by the addition of a suitable excess of one of the commonly used releasing agents, such as lanthanum chloride or strontium chloride.The main risk of un- suspected error arising at higher concentrations of the analyte element will therefore be in analyses in which the addition of a releasing agent is not normally regarded as essential. However, in view of the extent of the effect in the determination of magnesium, it is surprising that a more widespread occurrence of sulphate interference has not been reported. Nickel chloride, for example, sublimes6 at 1246 K, whereas the sulphate decomposes at 1121 K, and the oxide melts at 2263 K, so that the effect should be observed, particularly at a low height in the flame, for nickel. Cobalt chloride melts and boils at 997 and 1322 K, respectively, while the sulphate decomposes at 1008 K, and the oxide only melts at 2208 K.It has been stated in the literature,2 however, that nickel chloride and sulphate and cobalt chloride and sulphate will give the same responses. I t was found in practice, however, that considerable depressions could be observed when sulphates were used to prepare standard solutions at higher concentrations than those normally employed under the most sensitive conditions for the determinations of these two elements. The extents of the effects for 50 pg ml-l solutions of cobalt and nickel are shown in Fig. 4 for various observation heights as functions of the fuel flow-rate. The same trends were observed as for magnesium: the degree of interference became greater at lower heights in the flame, and as the fuel flow-rate + 2c 0 -20 +- 0, a L n -40 cn m .c m - -60 .- v, - 80 - 100 0.9 1 .o 1.1 1.2 1.3 Fuel flow-rate/l min-' Fig.4. Effect of fuel flow-rate on the change in cobalt absorbance a t 341.3 nm (broken lines) and nickel absorb- ance a t 234.8 nm (solid lines) caused by using sulphate in place of chloride, a t various heights above a lOO-mm, flat-topped burner in line with the optical axis.90 CRESSER AND MACLEOD was increased. The effects were slightly less pronounced when the burner with a triangular cross-section was used, but became more pronounced when either burner was rotated, and as the concentrations of the analyte element increased. With nickel, it is perhaps worth mentioning here that although under normal working conditions the effect of using sulphate rather than chloride to prepare standard solutions was negligible for 1 and 5 pg ml-l concentrations, and only an 8 per cent.depression occurred for 50 pg ml-l concentration, if measurements were made at a low height in a fuel-rich flame depressions of 48, 72 and 90 per cent., respectively, could be observed for these three con- centrations. There can be little doubt that many other instances will be found when the incidence and extent of interferences increases at higher concentrations of the analyte element. The effect is not confined to sulphate, and may be observed for other oxy-anions. Thus, for example, the depression of the absorbance of magnesium by the addition of different excess amounts of phosphate was found to vary with magnesium concentration (see Table I).TABLE I EFFECT OF PHOSPHATE ON MAGNESIUM ABSORBANCE AT A LOW HEIGHT IN A FUEL-RICH FLAME Change in absorbance, per cent., with phosphate (POIa-) concentrations of- A Magnesium 7 1 0.6’’ 0 0 0 6*0t 4 9 10 60*0$ 12 29 46 * 286.2 nm, burner aligned. t 286.2 nm, burner rotated through an angle of 90”. 202.5 nm, burner rotated through a small angle. (as MgC12)/mg ml-l 2 p g ml-l 20 p g ml-l 200 pg ml-1 When possible cationic interferences are being investigated, care must be taken to ensure that the associated anion does not cause a variation in the incidence or extent of apparently simple cationic interferences. Thus, for example, 50 pg ml-l of magnesium (as sulphate) interfered considerably in the determination of 2 pg ml-l of cobalt or nickel, whereas the same concentration of magnesium as the chloride caused no interference.It can be concluded that care should be taken to establish the absence of additional or increased interferences from concomitant elements and ions when employing burner rotation or an absorption line of poorer sensitivity for analysis at high concentrations of the analyte element. The risk of increased interference can be reduced by making measurements by using the upper part of a fuel-lean flame, particularly if a burner with a triangular cross- section is used. Although the results obtained with one particular instrument provide an indication of interference trends, significant variation can be expected to occur between different instru- ments and even between different nebulisers, so that it is essential to check for interferences on the instrument to be used for the analysis. The authors are indebted to Messrs. G. Wilson and D. Strath, for the construction of the burner with a triangular cross-section, and to Mrs. S. Reid, for assistance with some of the experiment a1 work. References 1. 2. 3. 4. 5. 6. Kirkbright. G. F., and Sargent, M., “Atomic Absorption and Fluorescence Spectroscopy,” Academic Aldous, K., and Reynolds, R. J., “Atomic Absorption Spectroscopy,” Charles Griffin & Co. Ltd., Koirtyohann, S. R., and Pickett, E. E., Analyt. Chem., 1968, 40, 2068. West, A. C., Fassel, V. A., and Kniseley. R. N., Analyt. Chem., 1973, 45, 1686. West, A. C., FasseI, V. A., and Kniseley, R. N., Analyt. Chem., 1973, 45, 2420. “Handbook of Chemistry and Physics,” 53rd Edition, The Chemical Rubber Publishing Co., Cleveland, Received August 8th, 1975 Accepted Octobev 6th, 1975 Press, London, New York and San Francisco, 1974. London, 1970. Ohio, 1972-1973.

ISSN:0003-2654    

DOI:10.1039/AN9760100086

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, Febmary, 1976, Vol. 101, pp. 91-95 91 An Improved Digestion Method for the Extraction of Mercury from Environmental Sam pies Haig Agemian and A. S. Y. Chau Canada Centre for Inland Waters, Watev Quality Branch, P.O. Box 5050, 867 Lakeshore Road, Burlington, Ontario, L7R 4A6, Canada An improved digestion procedure for the extraction of mercury from environ- mental material is reported. The method involves the digestion of the sample at 60 "C with sulphuric acid - nitric acid (2 + l), containing a trace amount of hydrochloric acid, and subsequent oxidation with permanganate and per- sulphate solutions. With this procedure mercury is successfully recovered from organic matter and resistant inorganic forms such as mercury(I1) sulphide. Unlike digestion with aqua regia, this procedure is simple and safe, and is applicable to the digestion of a large number of samples simul- taneously.The method can be adapted to the automated cold-vapour and flame atomic-absorption techniques and is therefore ideal for routine monitoring. Agemian and Chaul reported a method for the determination of mercury in sediments, in which a comprehensive manual digestion method developed by Iskandar et aL2 was adapted to the automated cold-vapour atomic-absorption technique. The extraction technique involved the digestion of the sediment with sulphuric acid - nitric acid (2 + 1) at 60 "C and subsequent oxidation of the organomercury compounds with permanganate and persulphate solutions. Complete recovery2 of several organomercury compounds was obtained by using this technique.Jacobs and Keeneys showed that this procedure does not adequately dissolve mercury( 11) sulphide (cinnabar), which may be formed under reducing conditions. They suggested treating the sediment with boiling aqua regia. However, we found that this treatment is hazardous, impractical and unnecessary for routine analysis of a large number of samples. Moreover, Jonasson et aL4 have reported that the presence of excessive amounts of hydro- chloric acid should be avoided as the consequent generation of chlorine5 may cause the loss of volatile mercury chlorides. They proposed the use of a solution of nitric acid plus a trace amount of hydrochloric acid, which technique enables mercury sulphides to be satisfactorily dissolved, and is suitable for application to geological samples.Agemian et aL6 showed the adaptability of both sulphuric acid - nitric acid (2 + 1) and hydrochloric acid - nitric acid (1 + 9) digestion mixtures to the automated cold-vapour atomic-absorption technique for mercury determination. They compared these extraction methods for a number of rock, soil and sediment samples and obtained identical results, presumably owing to the absence of cinnabar in the samples analysed. Samples that contain substantial amounts of this mercury compound are not commonly analysed in this laboratory. However, in this study, the intention was to obtain a compre- hensive method that is applicable to all types of samples, including those containing mercury in forms such as mercury(I1) sulphide. Cinnabar is resistant to attack by sulphuric and nitric acids.* However, the inclusion of the minimum4 amount of hydrochloric acid pro- motes its rapid decomposition.Thus, a slight modification of the digestion method of Iskandar et aZ.,2 by the addition of a trace amount of hydrochloric acid, enables large amounts of mercury(I1) sulphide, at levels much higher than are found in natural samples, to be recovered. The proposed method is simple, rapid and adaptable to the determination of all organic and inorganic forms of mercury in a large number of samples by the automated cold-vapour atomic-absorption technique. The method can be suited to the digestion of biological (fish)' or geological (rocks, soils and sediments) samples by using the procedure with or without the inclusion of hydrochloric acid. The extraction medium is especially suitable for sediments and soils as its oxidising nature enables the large amounts of organic matter frequently found in these samples to be oxidised;92 AGEMIAN AND CHAU : AN IMPROVED DIGESTION METHOD FOR THE Analyst, VoZ.101 also, the strongly acidic medium, in the presence of a trace amount of hydrochloric acid, extracts inorganic forms of mercury including that found in cinnabar. Experimental Apparatus Samples were digested in 100-ml calibrated flasks in a temperature-controlled shaker bath (Precision Scientific Co. , Model 75). The equipment used for the automated cold-vapour analysis consisted of ( a ) an automatic sampler (Technicon AutoAnalyzer I1 sampler with 20-1/5 cam) ; ( b ) a proportionating pump (Carlo Erba, Model 08-59-10202) ; (c) Technicon AutoAnalyzer tubing of specified dimensions ; ( d ) a gas separator, as used by Agemian and Chaul; (e) a mercury monitor (Pharmacia Fine Chemicals) ; and (f) a strip-chart recorder (Hewlett-Packard, Model 7101B).The system is similar to that used by Agemian and co-workers.l,s For high levels of mercury, a Perkin-Elmer 503 atomic-absorption spectrophotometer, equipped with an Intensitron lamp, was used with an air - acetylene flame. Reagents High-purity certified reagents were used for all analyses. Sulphuric acid, 36 N. Nitric acid, 16 N. Hydrochloric acid, 12 N. Potassium permanganate solution, 6 per cent. m/V. Potassium persulphate solution, 5 per cent. m/V. Tin(II) sulphate solution, 10 per cent.m/V in 2 N sulphuric acid. Hydroxylammonium sulphate (6 per cent. m/V) - sodium chloride (6 per cefzt. m/V) solution. Procedure Determine the water content of the wet sediment by drying it overnight to constant mass at 110 "C. Weigh a representative sample of wet sediment, equivalent to 0.1-2 g of the dry mass, into a 100-ml calibrated flask. Wash the sediment down to the bottom of the flask with mercury-free distilled water, place the flask in an ice - water bath and slowly add 15 ml of sulphuric acid - nitric acid (2 + 1). After cooling, add 2 ml of hydrochloric acid. Shake the mixture and let it stand for 5 min. After expelling the acid fumes from the flask, place it in a shaking water-bath at a tempera- ture of 50-60 "C and digest for 2 h. Then allow the flask to cool for 30 min and carefully add 10 ml of potassium permanganate solution while cooling in an ice - water bath.If the colour does not persist for 15 min, add a further amount of potassium permanganate solution. After 30 min, add 5 ml of potassium persulphate solution, with gentle stirring, and allow the mixture to stand overnight. If all of the permanganate is reduced, as witnessed by the absence of the purple colour, add potassium permanganate solution until the colour persists. Add 10 ml of hydroxylammonium sulphate - sodium chloride solution and stir the mixture gently until the solution becomes clear and all of the precipitated manganese(1V) oxide has dissolved. Make the solution up to volume and centrifuge an aliquot a t 2500 rev min-l for 5 min.Transfer an aliquot of the clear supernatant liquid into a glass sample cup and place it in the automatic sampler for analysis. Use a cam designed for 20 samples per hour and a sample to wash ratio of 1 : 5 , corresponding to a sampling time of 30 s and a wash time of 2.5 min. The linear concentration range in solution is 0.0002-0.006 mg 1-l for the non-flame detection system1 and 10-300 mg 1-1 for the flame detection system.8 For sample concentrations of 0-1-2 g per 100 ml, the non-flame detection system has a range of 0.01-6 mg kg-l and the flame method 5-300 mg k g l of mercury in the sediment. Further dilution of the extracts could extend the upper limit of both detection systems. Treat all standards exactly as for the above samples. For concentrated solutions of mercury use the flame technique.Results and Discussion Sulphuric acid is used extensively for the extraction of mercury from biological materials, It has also been used separatelyll-l3 or together with either alone7,Bs10 or with nitric acid.llFebruary, 1976 EXTRACTION OF MERCURY FROM ENVIRONMENTAL SAMPLES 93 nitric acid2J4J5 for sediments and rocks. Iskandar et aL2 also showed that a mixture of sulphuric and nitric acids is necessary for the adequate digestion of organic matter in sedi- ments and soils and for the subsequent liberation of mercury. As already indicated, the presence of hydrochloric acid is necessary for the decomposition of cinnabar. Initially, an attempt was made to use large amounts of hydrochloric acid, as suggested by Jacobs and K e e n e ~ , ~ but the procedure proved to be very impractical and difficult and gave rise to low recoveries of mercury from sediments.The use of a fume hood in which to perform the digestion is essential when hydrochloric acid is heated, even to 60 "C: Further, on adding potassium permanganate in order to oxidise the organomercurials, violent frothing of the solution occurs, with evolution of chlorine, which makes the analysis very difficult. In addition, the decompositions of the permanganate to manganese(1V) oxide caused by the hydrochloric acid, renders it ineffective as an oxidant for organomercurials, and a low recovery of mercury may therefore result. We found that as little as 5ml of hydrochloric acid caused the rapid decomposition of 25 ml of a 5 per cent.m/V potassium permanganate solution. Table I shows the effect of excess of hydrochloric acid on the recovery of mercury from sediments. The low results obtained (third column) are attributed to the loss of mercury due to violent frothing of the solution and possible volatilisation, and also to the reduced efficiency of oxidation of the organic matter by permanganate owing to its decomposition by hydrochloric acid. The analyses were performed by using the automated method reported by Agemian and co- workers.lt6 The samples examined were ordinary samples as analysed in our laboratory and therefore did not give any indication of the presence of cinnabar. Comequently, the method of Iskandar et aL2 gave results for recovery that are similar to those given by the proposed met hod.TABLE I EXTRACTION OF MERCURY FROM SEDIMENT SAMPLES BY USING DIFFERENT EXTRACTION MEDIA Results are expressed in pg kg-1. Extraction medium 7 - HISO, - HNO, €I,S04 - HNO, - HC1 H,SO, - HNO, - HCf Sample (2 + I)* (10 + 5 + 2)t (2 + 1 + 1): Silt .. .. 310 320 250 Silty clay . . .. 600 600 450 Clay . . .. 970 960 600 Clay . . .. 1000 1000 800 Clay . . .. 1600 1600 1400 * Method of Iskandar et al., t Present method. : Contains an excessive amount of hydrochloric acid. In order to check the efficiency of the method, sediment samples were spiked with mercury(I1) sulphide powder and the recovery of the mercury was determined. As only milligram amounts could reliably be weighed, the analysis had to be performed by the flame atomic-absorption method; the results obtained are given in Table I1 and show a satisfactory recovery of 100 & 5 per cent.of mercury added as mercury(I1) sulphide. The small amount of hydrochloric acid added in the digestion procedure causes the decomposition of the cinnabar to become visually apparent, the disappearance in a few seconds of the bright red colour indicating its complete dissolution. TABLE I1 RECOVERY OF MERCURY FROM MERCURY(II) SULPHIDE BY THE PROPOSED METHOD Mercury added (as HgS)/g . . . . 0.0018 0.007 1 0.0103 0.0222 0.029 1 Recovery, per cent. . . . . .. 102 98 105 95 101 The levels of the spikes used for the recovery tests are much higher than would be found in any real samples, the mercury contents16 of most uncontaminated solid earth materials being in the range 10-500 ng k g l .Based on l-g samples, the results shown in Table I1 correspond94 AGEMIAN AND CHAU: AN IMPROVED DIGESTION METHOD FOR THE Analyst, Vol. 101 to levels of lo00 mg k g 1 and above, which are much higher than those reported for most contaminated samples; Konradl' reported levels of 700-800 mg k g l for samples from chlor- alkali plants in Wisconsin. Therefore, the satisfactory recovery of amounts of mercury that are much larger than those encountered in highly contaminated samples indicates that the proposed method is adequate. As confirmed by Jacobs and Keeney,a use of the method of Iskandar et aL2 did not permit the recovery of any of the metcury introduced as mercury(I1) sulphide, which was completely resistant to the sulphuric acid - nitric acid mixture.Table I11 shows the effect of hydrochloric acid on the recovery of mercury from geological samples that contain large amounts of sulphide. The results given in this table show the superiority of the proposed method, using hydrochloric acid, for determining low levels of mercury. The precisions of the two methods are similar and the coefficient of variation6 TABLE I11 EXTRACTION OF MERCURY FROM GEOLOGICAL SAMPLES THAT HAVE HIGH SULPHIDE CONTENTS Results are expressed in mg kg-l; all samples were analysed in triplicate. Extraction medium , Sample Sphalerite . . .. Copper concentrate . . Pyrite concentrate Ore head .. .. Tailing . . .. .. Tailing . . .. .. Lead concentrate . . Ore head .. .. Copper concentrate . . Oxidised tailing . . HSSO, - HNO, 11.1 (2 + 1)s 0.30 0.19 29.7 16.8 8.6 32.6 6.2 6.2 0.9 HSSO, - HNO, - HCl (10 + 6 + 2)t 12.6 0.31 0.33 29.3 16.6 8.7 36.7 6.7 6.2 0.9 I Difference, per cent.+ 11.2 + 3.2 + 42-4 - 1.4 - 1.8 + 1.1 + 9.0 + 4-6 0 0 * Method of Iskandar et al.' t Present method. varies with the concentration of mercury as follows: 14, 2 and 2 per cent. for 0.1, 0.6 and 2 mg kg-1 of mercury, respectively, levelling off at values of about 2 per cent. above this concentration range. Thus, from Table I11 it can be seen that the first, third, seventh and eighth samples statistically show higher levels of mercury by the proposed method. This evidence is consistent with the hypothesis that in these samples some of the mercury is in the form of mercury(I1) sulphide. In the remaining samples there is again, apparently, no mercury(I1) sulphide in spite of the high sulphide content.Therefore, the above results show that use of the proposed method is advantageous and convenient for the safe digestion of a large number of samples. In addition, the proposed method is directly applicable, without modification, to the digestion of organic and biological tissues, We have not shown recoveries for organomercury compounds because both Iskandar et aL2 and Jacobs and KeeneyS have shown that with their methods the mercury in these compounds can be successfully recovered. Their methods represent the two extremes of the proposed method, that is, our method is intermediate between the two digestion techniques and was found to give similar recoveries. Conclusion A digestion method applicable to biological and geological samples for the extraction of mercury has been shown to enable mercury to be satisfactorily recovered from mercury(I1) sulphide.The method permits the decomposition of all forms of mercury and is simple, rapid, safe and adaptable to the automated cold-vapour and flame atomic-absorption tech- niques for the determination of mercury. The authors thank Y. K. Chau for his comments on the original manuscript, I. R. Jonasson and K. I. Aspila for supplying the geological samples and S. Scott for her secretarial help.February, 1976 EXTRACTION OF MERCURY FROM ENVIRONMENTAL SAMPLES 95 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. References Agemian, H., and Chau, A. S. Y., Analytica Chim. Acta, 1975, 75, 297. Iskandar, I. K., Syers, J. K., Jacobs, L. W., Keeney, D. R., and Gilmour, J. T., Analyst, 1972, 97, Jacobs, L. W., and Keeney,‘D. R., Envir. Sci. Technol., 1974, 8, 267. Jonasson, I. R., Lynch, J. J., and Trip, L. J., Geol. Surv. Pap. Can., 1973, Paper 73-21. Lingane, J. J., “Analytical Chemistry of Selected Metallic Elements,” Reinhold Publishing Corpora- Agemian, H., Aspila, K. I., and Chau, A. S. Y . , Analyst, 1975, 100, 253. Uthe, J. F., Armstrong, F. A. J., and Stainton, M. P., J. Fish. Res. Bd Can., 1970,27, 806. “Analytical Methods for Atomic Absorption Spectrophotometry,” Perkin-Elmer Corp., Norwalk, Polley, D., and Miller, V. L., Analyt. Chem., 1966, 27, 1162. Thorpe, V. A., J . Ass. Off. Analyt. Chem.. 1971, 54, 206. Jeffus, M. T., and Elkins, T. S., J . Ass. 08. Analyt. Chem., 1970, 53, 1172. Hatch, W. R., and Ott, W. L., Analyt. Chem., 1968, 40, 2085. Head, P. C., and Nicholson, R. A., Analyst, 1973, 98, 63. Cranston, R. E., and Buckley, D. E., Envir. Sci. Tcchnol., 1972, 6, 274. Melton, J. R., Hoover, W. L., and Howard, P. A., Proc. Soil Sci. SOC. Am., 1971, 35, 850. “Mercury in the Environment,” Prof. Pap. U.S. Geol. Surv., 1970, Paper 713. Konrad, J . G., Research Report 74, Wisconsin Department of Natural Resources, Madison, Wis., Received March 24th, 1976 Accepted August 27th, 1976 388. tion, New York, 1966. Connecticut, 1973. 1971.

ISSN:0003-2654    

DOI:10.1039/AN9760100091

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

96 Analyst, February, 1976, Vo,?. 101, pp. 96-102 The Application of a Wide-slot Nitrous Oxide = Nitrogen -Acetylene Burner for the Atomic-absorption Spectrophotometric Determination of Aluminium, Arsenic and Tin in Steels by the Single-pulse Nebulisation Technique* I<. C. Thompson7 and R. G. Godden Shandon Southern Instruments Limited, Frimley Road, Camberley, Surrey, G U16 5E T Single-pulse nebulisation of 10 per cent. m/V iron or steel solutions into a nitrogen-diluted nitrous oxide - acetylene flame maintained on a specially designed wide-slot burner is a useful technique for the determination of tin, arsenic and soluble aluminium in iron and steels. Use of this method avoids the need for prior separation of the analyte. A deuterium lamp was found to be unsatisfactory for measuring the background (non-specific) absorption when determining aluminium and tin, the explanation for which is postulated.Atomic-absorption flame spectrophotometry has traditionally depended on the continuous introduction of the sample into the flame during the measurement period. The nebulisation of steel solutions containing more than 1-2 per cent. m/V of steel into a conventional nitrous oxide - acetylene flame may result in a rapid partial blockage of the burner slot. This effect can be overcome by nebulising discrete sample aliquots, typically 25-200 p1, and recording the resulting pulse absorption ~ignal.l-~ Single-pulse nebulisation studies in flames have also been reported for the analysis of microlitre samples by flame-emission and atomic- fluorescence spectroscopy4 and in conjunction with an ultrasonic nebuli~er.~ Other workers6J have shown that the addition of air or argon to the nitrous oxide - acetylene flame increases the operating safety margin and minimises inter-element effects in the determination of magnesiums and ~trontium.~ Conventional atomic-absorption methods for the determination of low levels of aluminium and some other metals in steels usually require the prior extraction of the iron.*PB Shaw and OttawaylO have proposed the use of electrothermal atomisation using a graphite tube to overcome this problem.This paper reports the application of the nitrogen-diluted nitrous oxide - acetylene flame , in conjunction with a wide-slot burner and pulse nebulisation of the sample, to the deter- mination of arsenic, tin and soluble aluminium in steels.This study also shows that the use of a deuterium hollow-cathode lamp for background correction can lead to erroneous results for aluminium and tin. The reason for this behaviour is postulated. Experimental Results were obtained by using a Shandon Southern Instruments A3400 atomic-absorption spectrophotometer with an A3429 air - nitrous oxide change-over valve. The output was monitored on a Shandon Southern Autograph potentiometric recorder. A nebuliser with a platinum-iridium capillary was used. The nitrogen supply to the burner was fed, via a non-return valve, into a 2.5-1 metal reservoir. The exit of this reservoir was connected to the inlet of the supplementary air flow meter. The needle valve of this flow meter was permanently set to give a nitrogen flow-rate of 2 1 min-l at a nitrogen pressure of 20 p.s.i.g.(138 kN m-2). A pressure-sensing device was incorporated into the reservoir such that if the nitrogen pressure in the reservoir fell below 16 p.s.i.g. (110 kN m-2) a cut-off valve (Dew- range Controls Ltd.) in the acetylene line to the A3400 was activated. Tests showed that if the nitrogen supply to the unit failed during operation the flame was always extinguished without explosive flashback. The system is depicted in Fig. 1. The design of the wide-slot burner grid is shown in Fig. 2. * Presented a t 18th Colloquium Spectroscopicum Internationale, Grenoble, 15th-19th Septembe, 1975. t Present address : Severn - Trent Water Authority, Malvern Regional Laboratory, Milbourne Lodge, 141 Church Street, Malvern, Worcestershire.THOMPSON AND GODDEN B 97 50 mm il h A I Fig.1. Gas arrangement. A, A3400 spectrophotometer; B, flow meter + needle valve; C, 2.5-1 reservoir; D, non- return valve; E, pressure sensor; F, cut-off valve; and G, A3429 change-over valve. An air - acetylene mixture, with nitrogen flowing at the rate of 2 1 min-l, was ignited and the acetylene flow-rate set to give a fuel-rich flame. The change-over valve was then operated so as to substitute nitrous oxide for the air. A nitrous oxide flow-rate of 11.5 1 min-l and a nitrogen flow-rate of 2 1 min-l were used for all studies; the acetylene flow-rate was optimised for each element. The burner was removed and washed out with dilute hydrochloric acid after operating for about 2 h.Occasionally the burner slot was polished with a very fine grade of emery paper, which was found to minimise carbon build-up on the jaws of the burner. Tests showed that even if a severe reduction in the nitrous oxide flow-rate occurred, the acetylene supply could be turned off without risk of explosive flashback. This is not true for the standard burner (without nitrogen) with which, if the nitrous oxide flow-rate is reduced by 30 per cent. or more, the nitrous oxide must be turned off prior to the acetylene in order to prevent an explosive flashback. With the wide-slot burner it was possible to nebulise a 5 per cent. m/V steel solution continuously for 15 min before signs of burner clogging became evident. If 5 per cent.m/V steel solutions were nebulised for 15 s, with an equal nebulisation period for distilled water between samples, partial blockage of the A98 THOMPSON AND GODDEN: AAS DETERMINATION OF Al, As AND Sn Analyst, Vol. 101 burner slot occurred after 4045 min of operation. The burner was allowed to warm up for 5 min before any steel solutions were nebulised, otherwise partial clogging of the burner slot occurred more rapidly. Optimisation of Operating Conditions Nitrogen $0 w -rate For most elements tested the sensitivity decreased with increasing nitrogen flow-rate. A nitrogen flow-rate of 2 1 min-l gave a satisfactory safety margin with respect to flashback, and adequate sensitivity. In Table I the sensitivity of the wide-slot burner (slot width 0-60 mm) is compared with that of the standard burlier (slot width 0.43 mm).It can be seen that for most elements tested the use of the wide-slot 1)urner results in an improvement in sensitivity even though the flame temperature must be lower. This effect was attributed to a larger flame volume where the breakdown of refractory oxides can occur. The burner height setting and acetylene flow-rate were not so critical with the wide-slot burner, and the optimum burner height setting tended to be 2-3 mm lower than that of the standard burner. The optimum acetylene flow to the wide-slot burner was not dependent on the nitrogen flow-rate. Although it was possible to use air6 instead of nitrogen, this change reduced the safety margin and the optimum acetylene flow was then very dependent on the air flow-rate.TABLE I COMPARISON OF SENSITIVITY OF STANDARD AND WIDE-SLOT BURNER Characteristic concentration (continuous nebulisation) / pg ml-1 Standard burner Wave- (norm a1 operating Standard burner Element length/nm conditions) (+ 2 1 min-l of nitrogen) A1 309.3 1.0 1.2 A1 396.2 As 193.7 1.7 1.6 Ba* 653.6 0-37 0.4 1 Si 251.6 2.0 3.6 Sn 236.6 286.3 2.6 2-4 364.3 1.9 3.6 Ti V 318-4 1.6 1-8 * + 1000 pg ml-l of potassium. t Better signal to noise ratio a t 286.3 nm than a t 236.6 nm. (doublet) - - - - Snt Wide-slot burner (+ 2 1 min-1 of nitrogen) 0.76 0.84 1.1 0.30 2.1 1.2 1.9 2.1 1.3 Optimum sample volume for pulse nebulisation A sample volume of 2 0 0 ~ 1 was found to give the best compromise with respect to the signal to noise ratio and absence of blockage of the burner slot when nebulising 10 per cent.m/V steel solutions, which is in agreement with other ~tudies.l-~ Damping operated a t a time constant of 2s. Spectral bandpass and wavelengths this study are listed in Table 11. Sample dissolution The steel sample (5 g) was carefully dissolved by slowly adding to it 50 ml of aqua regia. The solution obtained was then boiled for 15 min, allowed to cool and filtered into a 50-ml (for a 10 per cent. m/V solution) or 100-ml (for a 5 per cent. m/V solution) calibrated flask. The filter was washed with a small volume of 10 per cent. aqua regia solution and the com- bined filtrate and washings were diluted to volume with water. Calibration was carried out by the method of standard additions. The aluminium, arsenic and tin contents of the acid blank were below the detection limits of the technique.In order to attain the optimum signal to noise ratio the A3400 spectrophotometer was The resonance lines, background correction wavelengths and spectral bandpasses used inFebwary, 1976 IN STEELS BY THE SINGLE-PULSE NEBULISATION TECHNIQUE Analyte +--7 Element nm A1 309.3 Wave- length/ (doublet) A1 396.2 As 193.7 Sn 286.3 TABLE I1 RESULTS FOR BACKGROUND CORRECTION Background correction - Wave- length/ Element nm 309.3 311.4 309.3 Pd 396.9 193.7 286.3 286.3 3 286.6 3 Mg c u 309.4 D2 D2 Spectral band- passlnm 0.3 0.18 0.18 0.18 0.18 0.6 0.3 0.6 0.3 Lower energy level of background linelev 0.96 2.72 1.39 1.46 - - 4.00: Background absorption* fi-om 200 p1 of los pg ml-1 steel? soh tion 0.0012 0.000 26 0.000 26 0.0003 0.000 26 0.002 0.006 0.004 0.001g 99 Signal" from 200 p1 of a loa pg ml-1 solution of background correction element 0,017 0.0016 0.0012 0.003 - - 0.01 1 * Expressed as per cent.of analyte in steel. t BCS 466. 1 Ionic line. Probably caused by stray light from very intense 276.3-nm palladium resonance line. Measurement Procedure The burner was always allowed to run for 5 min in order to attain the normal operating temperature prior to nebulising the steel solutions; 200 pl of the sample were pipetted into a 20-ml disposable polystyrene beaker. The plastic capillary tube was removed from the blank solution and placed in the 20O-pl sample until the latter had completely disappeared from the beaker; the capillary tube was then replaced in the blank solution.A blank solution of 0.1 M hydrochloric acid was continuously nebulised between samples at all times. Some typical traces are shown in Fig. 3. 8 i! 8 ?I B 1 c 1 2 A 1 2 3 30 s 2 - Time --e Fig. 3. Typical traces (10 per cent. m/V mild steel solutions, 200-4 pulse nebulisation). A, Aluminium, 396.2 nm: 1, sample (0.0008 per cent. of Al); 2, as for (1) + 2.6 pg ml-l of Al; and 3, background (Pd 395.9 nm). B, Tin 286.3 nm: 1, sample (0.0075 per cent. of Sn); and 2, background (Pd 285.6 nm). C, Arsenic, 193.7 nm: 1, sample (0.032 per cent. of As); and 2, background (D2 193.7 nm).100 Background Absorption Measurements When concentrated solutions (5-10 per cent. m/V) are nebulised it is essential to check for non-specific background absorption caused by the saniple matrix.Initially, background absorption measurements were made by using a deuterium hollow-cathode lamp. However, for steel samples that contained very low levels of aluminium, the background signals from both 5 and 10 per cent. m/V steel solutions were larger than those observed by using the aluminium and tin lamps at 309.3 and 286.3 nm, respectively. This behaviour was repro- ducible and was also observed, to a slightly lesser degree, if the spectral bandpass was increased from 0.3 to 0.6 nm. The background absorption was also monitored by using the non-resonance lines of various elements. These lines are listed in Table 11, which also gives the lower energy level of the transition,I1J2 the absorption signal for a 200-p1 sample of a lo00 pg ml-l solution of the non-resonance line element and the background absorption signal expressed as a concentration of the analyte.The deuterium lamp had too low an intensity to be satisfactorily used at the aluminium wavelength of 396.2 nm. This curious behaviour by which the background absorption (deuterium lamp) was apparently higher than the sum of atomic plus background absorption (element lamp) was attributed to weak atomic absorption by the large concentration (about 106 pg ml-1) of iron. It was not considered to be justified to ignore the absorption by the iron non-resonance lines over the monochromator spectral bandpass. A typical absorption line half-width in the nitrous oxide - acetylene flame is about 0.005 nm.13 Assuming a triangular line profile, a spectral bandpass of 0.3 nm and complete absorption over the absorption line profile, then (0.005/0.3) x 100 = 1.7 per cent.of the radiation of the deuterium lamp will be absorbed for a single absorption line that falls within the monochromator spectral bandpass (ignoring any additional absorption from the wings of the line). In Table I11 the iron lines12 that are within 0-15 nm of the aluminiuni and tin lines used in this work are listed. It can be seen that this explanation would account for the anomalously high absorbance values observed when using the deuterium lamp at 286.3 and 309.3nm, and was further substantiated by the fact that very large background absorption signals at 248.3 and 279.6 nm were observed when using a deuterium lamp and nebulising a 5 per cent.m/V solution of a manganese steel sample. TABLE I11 THOMPSON AND GODDEN : AAS DETERMINATION OF Al, As AND Sn AnaZyst, VoZ. 101 Results IRON (I) LINES WITHIN & 0-15 rlm OF THE ALUMINIUM, ARSENIC AND TIN RESONANCE LINES Analyte -7 Element Wavelength/nm A1 doublet A1 396.153 As 193.696 Sn 286.333 Iron linesle/nm 309.168 309.278 309.336 309.381 309.388 396.029 396.116 396.235 None 286,250 286-344 280.386 Lower energy level of ironla line/eV 1.02 2.96 1.61 2.67 3.64 2.85 3.27 1.01 1-48 0.09 - - The deuterium lamp had a molybdenum cathode. However, a 1000 pg ml-l molybdenum solution (equivalent to 1 per cent. of molybdenum in steel for 10 per cent. m/V solutions) gave no response at any of the deuterium background wavelengths used. A hydrogen hollow-cathode lamp with a nickel cathode gave background absorption results that were similar to those given by the deuterium lamp..The outputs from both lamps were scanned in the emission mode using a spectral bandpass of 0.18nm. The 248.3-nm iron line and the 279.6-nm manganese line could not be detected above the continuum emission. Wagenaar and de Galan13 have shown that the profile of the 396.2-nm aluminium absorptionFebrzlary , 1976 IN STEELS BY THE SINGLE-PULSE NEBULISATION TECHNIQUE 101 line in the nitrous oxide - acetylene flame has a half-width of 0.0048 nm. The line half-width of the 396-2-nm resonance line from the aluminium hollow-cathode lamp was quoted as being 0.0013 nm. If it is assumed that the iron lines given in Table I11 have a similar absorption line half-width, spectral overlap of the profiles of the iron absorption lines with the aluminium 396.2-nm resonance line or the palladium non-resonance lines used in this work should not occur.There could be slight overlap of the aluminium 309471/309&4 and iron 309.278-nm lines, but as similar analytical results were obtained by using the 309.3 and 396.2-nm alu- minium resonance lines, this was not thought to be of any practical significance. (Also the lower energy level of the iron 309.278-nm line is 2.96 eV above the ground state.12) There could be weak overlap of the tin 286-333-nm and iron 286-344-nm lines, but the steel sample (BCS 456) gave an identical (very small) signal at the tin resonance and palladium background wavelengths, which would indicate that iron atoms are not absorbing radiation at the wavelength of the tin resonance line.Also tin measurements made at the 235.485-nm tin resonance line, with background correction measurements made using the palladium 235.134-nm line, gave similar results. For arsenic measurements the deuterium lamp gave satisfactory results. This is to be expected as no iron lines are listed1l9l2 in the wavelength region 191-198 nm. It can be seen that great care must be exercised in the choice of non-resonance lines for background correction. The response (0406 absorbance unit) for a 1000 pg ml-l magnesium solution at the 309.3-nm magnesium non-resonance line was found to be caused by stray light from the magnesium 285.2-nm resonance line (the intensity of the 309.3-nm line was only 1.5 per cent.of that of the 285.2-nm line). Thus the use of this line for background correction would be invalid if the sample contained minor amounts of magnesium. If an OX9 filter (Barr & Stroud Ltd.), which is effectively opaque at 285.2 nm but transmits at 309.3 nm, is placed between the flame and the monochromator, no response is observed from the 1000 pg ml-l magnesium solution. Thus, when selecting a non-resonance line for back- ground correction in complex matrices it is essential to check for a response using a solution containing the maximum concentration of the non-resonance lamp element that is likely to be encountered. If the cathode is an alloy (e.g., brass) all the major elements of the alloy should similarly be checked.It is also advisable to check that the selected non-resonance line does not overlap any atomic line profiles of the main components of the analyte. A palladium lamp was used for all background correction measurements for aluminium and tin, while for arsenic a deuterium lamp was used. I t was found that the noise level and drift of the output from the deuterium lamp were greater than those from the aluminium, tin and palladium lamps. Analysis of Steel Samples Preliminary studies were performed using 2OO-pl aliquots of 5 per cent. m/V solutions. However, subsequent work showed that there were no burner clogging problems with 200-4 aliquots of 10 per cent. m/V solutions and the latter were utilised in all further work. The response for a given amount of aluminium from the 10 per cent.m/V steel solution was 80 per cent. of that from a 5 per cent. m/V steel solution. The calibration graphs were linear up to concentrations 50 times greater than the detection limits (see Table IV). Higher concentrations were not studied. TABLE IV DETECTION LIMITS FOR ARSENIC, TIN AND SOLUBLE ALUMINIUM IN STEEL Element Linelnm 20 detection limit, per cent. A1 309-3 0.000 25 A1 396.2 0.000 25 As 193.7 0.002 Sn 286.3 0*0008 Table IV shows the 20 detection limits obtained, and Table V some results obtained for arsenic, tin and soluble aluminium in standard steel samples. The BCS 494 manganese steel contained 13 per cent. of manganese, but a 15 000 pg ml-l manganese solution gave negligible background absorption at all of the analytical wavelengths used.The relative standard deviation for a solution containing 2.5 pg ml-l of aluminium (396-2 nm) in 10 per cent. m/V steel solution (20 measurements) was 3.1 per cent.102 THOMPSON AND GODDEN Conclusions The single-pulse nebulisation technique in conjunction with a nitrogen-diluted nitrous oxide - acetylene flame maintained on a wide-slot burner allows the direct nebulisation of 10 per cent. m/V steel solutions, thus avoiding the need to carry out a prior separation of the analyte from the matrix. The method is also very useful if the amount of sample available is limited. TABLE V DETERMINATION OF ARSENIC, TIN AND SOLUBLE ALUMINIUM IN IRON AND STEEL Concentration, per cent. r \ A Sample Element Pulse nebulisation Other methods BCS 466 mild steel . . . . A1 0.0008 0*0007* BCS 494 manganese steel .. A1 0.000 85 0-0007* BCS 260/3 high-purity iron . . A1 o*ooo 55 0.000 42: BCS 451 mild steel . . * . Sn 0.0075 O-OOS$ BCS 451 mild steel . . .. As 0.032 0.0315 BCS 453 mild steel . . .. As 0.056 0.0525 iron with isobutyl acetate. 0.001 t BCS 453 mild steel . . .. Sn 0.0185 0.0 193 * Results obtained by Mr. R. C. Rooney using a method based on the extraction of t Single BCS value. : Results obtained by Shaw and Ottaway.lo § BCS standard values. A deuterium hollow-cathode lamp was found not to be satisfactory for making background correction measurements at wavelengths 286.3 and 309.2 nm when nebulising 10 per cent. m/V steel solutions, which was attributed to atomic iron lines within the monochromator spectral bandpass, which absorb radiation from the deuterium lamp.Care should be exercised when using a deuterium lamp for background correction measurements with this type of analysis. A better technique for background correction is to utilise a non-resonance line from an element that is unlikely to be encountered in the steel samples. Ideally, the chosen back- ground correction line and analyte resonance line should not overlap any atomic line profiles of the major elements in the sample. The authors thank the Directors of Shandon Southern Instruments Limited for permission to publish this paper and Mr. R. C. Rooney of Kooney and Ward Limited, Blackwater, Camberley, Surrey, for the results for soluble aluminium in steel using the isobutyl acetate extraction method. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. References Sebastiani, E., Ohls, K., and Riemer. G., 2. Analyt. Chem., 1973, 264, 105. Berndt, H., and Jackwerth, E., Spectrochim. Acta, 1975, 30B, 169. Manning, D. C., Atom. Absorption Newsl.. 1975, 14, 99. Sarbeck, J. R., St. John, P. A., and Winefordner, J . D., Mikrochim. Acta, 1972, 65. Korte, N. E., Moyers, J . L., and Denton, M. B., Analyt. Chern., 1973, 45, 530. Fleming, H. D., Spectrochim. Acta, 1967, 23B, 207. Fulton, A., and Butler, L. R. P., Spectrosc. Lett., 1968, 1, 317. Headridge, J . B., and Sowerbutts, A., Analyst, 1973, 98, 57. Jenkins, R. H., and Jones, C. P., British Steel Corporation Research Report No. SM/464/A, 1972. Shaw, F., and Ottaway, J . M., Analyst, 1975, 100, 217. Meggers, W. F., Corliss, C. H., and Scribner, B. F., “Table of Spectral Line Intensities,” NBS Zaidel, A. N., Prokof’ev, V. K., and Raiskii, S. M., “Tables of Spectrum Lines,” Pergamon Press, Wagenaar, H. C., and de Galan, L., Spectrochim. Acta, 1973, 28B, 157. Monograph 32, Part 1, 1961. London, 1961. Received August 8th, 1975 Accepted October 24th. 1975

ISSN:0003-2654    

DOI:10.1039/AN9760100096

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, February, 1976, Vol. 101, pp. 103-110 103 The Determination of Mobile Nitrogen in Steel Using an Ammonium .Ion-selective Electrode J. B. Headridge and G. D. Long Department of Chemistry, The University, Shefield, S3 7HF An absorption cell containing an ammonium ion-selective electrode has been constructed and used for the determination of mobile nitrogen in steel; this nitrogen is released as ammonia when the steel is heated a t 500 "C in a stream of hydrogen. The cell was used in conjunction with a digital volt- meter and a recorder in order to obtain a continuous record of the progress of the reaction between mobile nitrogen and hydrogen. Results are pre- sented for the determination of 0.0005-0.0108 per cent. of mobile nitrogen in 10 steels using the new equipment and are compared with those obtained by using a spectrophotometric finish based on indophenol blue.The method, with relative standard deviations of 0.0001-0.0003 per cent., is more precise than that with the spectrophotometric finish, with relative standard deviations of 0.0002-0.0006 per cent. The mechanical properties of steel are greatly affected by the content of nitrogen, which is usually present within the range 0.001 to 0.05 per cent. Its presence can be harmful, causing age-hardening and flaws in pressings, and the presence of aluminium nitride can result in inter-granular fracture. The presence of nitrogen can also be beneficial by improving the strength and creep properties of stee1s.l Usually nitrogen occurs in steels both as mobile nitrogen and as stable nitrides of elements such as aluminium, silicon, titanium, zirconium, vanadium, niobium and chromium; the nitrogen in stable nitrides is referred to as combined nitrogen.The mobile nitrogen is less strongly bound in steel, occurring as atomic nitrogen or as less stable nitrides of iron and manganese. The ratio of mobile to combined nitrogen, and thus the properties of a steel, can be changed by heat treatment. Hence, for the control of heat treatments, a knowledge of the mobile and combined nitrogen contents is most helpful. The total nitrogen content of a steel can be determined by use of a Kjeldahl method, in which the nitrogen is converted into ammonia and determined by a suitable titrimetric or colorimetric procedure. The combined nitrogen content of a steel can be determined by using a similar procedure after separation of stable nitrides and other compounds from the steel, following dissolution of the metals, by using a methyl acetate - bromine mixture.The difference between the total and combined nitrogen values is the mobile nitrogen. However, a more convenient way to determine mobile nitrogen is to pass hydrogen over steel millings at 500 "C. The mobile nitrogen is thus converted into ammonia, which is absorbed in a suitable solution and determined spectrophotometrically as indophenol The success of this thermal method depends on two factors. Firstly, there must be no reaction at 500 "C between the mobile nitrogen and a metal in the steel, which will result in the precipitation of more stable nitride.Such a reaction would lead to a low result for mobile nitrogen. It has been established that such a reaction does not occur between mobile nitrogen and aluminium, silicon or titanium at 500 "C, hence, such an interfering reaction will not occur with most commercial steels. The interfering reaction occurs only with certain special steels, which contain appreciable concentrations of elemental vanadium and niobium.2 Secondly, it is important that no stable nitrides should dissociate at 500 "C as this would give rise to high results for mobile nitrogen. It appears that this is an interfering effect only for steels that contain appreciable concentrations of chromium nitride,2 and carbon and low-alloy steels of this type are rarely produced. Therefore, it is generally agreed that the thermal method, using hydrogen at 500 O C , is reliable for the determination of mobile nitrogen in most commercial carbon and low-alloy steels.Actually, the optimum temperature for this extraction method for mobile nitrogen may not be 500" C for all types of steels, but this information can be acquired only by more extensive use of the technique. The thermal method using hydrogen, coupled with an indophenol blue spectrophotometric finish, has been used successfully in our laboratory for some time. However, a spectro-104 HEADRIDGE AND LONG: THE DETERMINATION OF MOBILE NITROGEN Analyst, VoZ. 101 photometric determination of the ammonia that is evolved has the disadvantage that one cannot be certain that all of the mobile nitrogen from a particular alloy sample has been converted into ammonia at 500 "C within a particular period of time, unless multiple runs with increasing collection times are undertaken.The collection time for ammonia depends on the flow-rate of the hydrogen and on the dimensions of the steel millings. Millings that have at least one dimension less than 0.5 mm are usually used; also, collection times of the order of 1 h are often used. Clearly, it would be advantageous to have available a simple method for continuously recording the amount of mobile nitrogen that has been extracted as ammonia during a run and to know beyond doubt when all of the mobile nitrogen has been extracted. Such a method is described in this paper. To this end a special absorption cell for ammonia, incor- porating an ammonium ion-selective electrode in a triethanolamine - triethanolammonium ion buffer solution, has been constructed.The potential difference between this electrode and a mercury - mercury(1) sulphate reference electrode is passed to a digital voltmeter and continuously recorded on a potentiometric recorder. The point at which all of the mobile nitrogen has been collected can be seen at a glance and the mobile nitrogen content of the steel can then be determined immediately by reference to a suitable calibration graph. Because it was also our intention to compare the quality of the ion-selective electrode results for mobile nitrogen with those obtained by use of colorimetry, the 10 steel samples analysed by the proposed method were first analysed by using the spectrophotometric indophenol blue In order to obtain useful values for the precision of the methods, six samples of each steel were analysed by use of each method.Apparatus The apparatus for the release from a 1-g sample of steel of mobile nitrogen as ammonia and its collection and determination was straightforward in design and is shown in Fig. 1. Experimental Hydrogen I Flow meter Tap Tap cupbo: P Tap 3 \ To fume 3 rd . nryvi I iriayi I ~ J I U I I I perch lorate Digital mV meter Potentiometric recorder I / Silica tube Silica boat containing steet sample ~~ Fig. 1. Apparatus for the determination of mobile nitrogen in steel. The various components were as follows. Flow meter. This was a Meterate, Type RS2, fitted with a hydrogen tube (Glass Precision Engineering Ltd., Heme1 Hempstead).Resistance heated tube furnace (maximum temperature 1000 " C ) . This was made by the Amalgams Co. Ltd., Sheffield, and controlled by a Pye Ether Mini temperature controller. The furnace contained a silica tube of 40 mm internal diameter with a hot zone approximately 150 mm in length. Silica boats. These boats had internal dimensions of 100 x 17 x 7 mm and were formed from silica sheet 2.5 mm thick.February, 1976 IN STEEL USING AN AMMONIUM ION-SELECTIVE ELECTRODE 105 This contained a mercury - mercury(1) sulphate reference electrode (see Fig. 2). A bsorftion cell. Ammonium - potassi.um ion-selective electrode. EIL, Model 1057 200. Digital millivoltmeter. EIL, Model 7060. Potentiometric recorder. Oxford, 3000 series, single pen.Constant-temperature bath, thermostatically controlled at 25 "C. For immersion of the absorp- tion cell. Joint to accommodate Copper wire \ Rubber bung, Saturated magnesium sulphate \ Mercury ( I ) sulphate \ Mercury ' Platirium wire Fig. 2. Absorption cell for the ammonia produced by reaction of hydrogen with mobile nitrogen from a steel sample. Reagents Acetone. Redistilled analytical-reagent grade. Hydrogen. Air Products Ltd., high-purity grade, 99.99 per cent. Argon. Air Products Ltd., ultra-high-purity grade, 99.999 per cent. Triethanolamine. Fisons SLR grade. Hydrochloric acid, 1.000 M. This was prepared from ConvoL solution, Hopkin and Williams Ammonium ion free water. This was prepared by passing distilled water through a column Magnesium 9erchlorate.Fisons LR grade. Ammonium chloride solution A. Dissolve 3.821 g of analytical reagent grade ammonium chloride (dried at 140 "C) in ammonium ion free water and dilute to 1 1 in a calibrated flask with the same solvent. 1 ml of solution = 1 mg of nitrogen. Ltd. of Amberlite IR-120 ion-exchange resin (H+ form) and was used throughout this work. Ammonium chloride so2ution B. Dilute 10ml of solution A to 100ml with ammonium ion free water. 1 ml of solution ZE 100 pg of nitrogen. Ammonium chloride solution C. Dilute 10 ml of solution B to 100 ml with ammonium ion free water. Solution for the Absorption of Ammonia To respond to the ammonium ion-selective electrode the ammonia must be fixed as ammonium ion. Triethanolamine and its conjugate acid form a suitable buffer solution for the absorption of ammonia because the pK, value for triethanolammonium ion at 25 "C is 7.76.In theory, a buffer solution of triethanolamine and triethanolammonium ion in a concentration ratio of 1 : 10 should have a pH of 6.76 and in such a solution the ratio of 1 ml of solution = 10 pg of nitrogen.106 HEADRIDGE AND LONG: THE DETERMINATION OF MOBILE NITROGEN Analyst, Vd. 101 ammonium ion to ammonia resulting from the absorption of ammonia should be 309: 1, indicating that the fixing of ammonia as ammonium ion is virtually complete. Commercial triethanolamine contains some diethanolamine and it' was therefore necessary to determine the mass of triethanolamine that was equivalent to 1 mol of hydrogen ion. This was determined by titrating potentiometrically a suitable mass of the triethanolamine with 0.1000 M hydrochloric acid.One mole of hydrogen ion was found to be equivalent to 147.7 g of the commercial base. Preparation of the bufer solzltion triethanolamine (1-1 mol) dissolved in approximately 800 ml of water. then diluted to 2 1. at 25 "C. One litre of 1 M hydrochloric acid was added, with agitation, to 162-5 g of commercial The solution was The pH of this BH+ - B buffer solution (10 + 1) was measured as 6.88 Construction of a Suitable Calibration Graph A graph of cell potential vcyszls the logarithm of the concentration of ammonium ion took the form of a straight line down to M with a slope of 54 mV per unit of log[NH,+]. At lower concentrations the line curved increasingly towards the log concentration axis, but the electrode responded satisfactorily to changes in ammonium-ion concentration down to 7 x 10-6~, which is the concentration of ammonium ion produced when the ammonia equivalent to 0.0001 per cent.of mobile nitrogen in 1 g of steel is absorbed in 10 ml of buffer solution. A calibration graph of cell potential veisus concentration of nitrogen as ammonium ion (0-5-8 pg ml-l of nitrogen) was constructed after measuring the potential of the ion-selective electrode versus the reference electrode for seven solutions of ammonium ion in the buffer solution; these solutions were prepared as follows. To 50 ml of buffer solution in each of seven 100-ml calibrated flasks add, in turn, 5 ml of ammonium chloride solution C, and 1,2, 3, 4, 6 and 8 ml of ammonium chloride solution B arid dilute each solution with ammonium ion free water to 100ml. -570 -560 > E -550 L c -540 8 ' -520 .- c1 - 8.-530 - -510 -500 I 1 1 1 1 1 1 0 1 2 3 4 5 6 7 8 Concentration of nitrogen as ammonium ion/pg mi" A typical calibration graph for the deter- mination of nitrogen as ammonium ion collected from steel samples. Fig. 3. The ammonium ion-selective electrode was conditioned before use by standing it in the most dilute calibration solution (0.5 pg ml-l of nitrogen) for 2 d. The solutions were placed in the absorption cell in turn, in order from the most dilute to the most concentrated, and the cell potential was measured with each solution. Several minutes were allowed for the electrode to attain a steady potential, the electrode being agitated during this time.The electrode was washed only with the next solution between readings and not with distilled water. A typical calibration graph is shown in Fig. 3; calibration was carried out every 72 h. The changes in potential of the ammonium ion-selective electrode, when used with the cali-Febrzcary , 1976 IN STEEL USING AN AMMONIUM ION-SELECTIVE ELECTRODE 107 bration solutions, from one calibration to the next were never in excess of and usually less than 2 mV. When not in use the electrode was stored in the most dilute calibration solution. Method for the Determination of Mobile Nitrogen in Steel Using the Ion-selective Electrode With argon flowing through the furnace and the by-pass tube at a rate of 300 ml min-l, turn on the furnace and allow the temperature to become steady at 500 "C.Wash'the absorption compartment of the cell with water, followed by acetone, these solvents being removed by means of a hypodermic needle connected to a suction pump. Close tap 3 and open tap 4 so that hot argon passes through the sinter and completely dries the absorption compartment (Note 1). Then close tap 4 and open tap 3 so that the argon by-passes the cell. Increase the flow-rate of argon to approximately 1.5 1 min-l and remove the rubber bung at the end of the silica tube. Place the silica boat containing exactly 1 g of steel into the cool end of the furnace tube nearest the flow meter and replace the bung. Next, purge the furnace tube with argon for approximately 3 min (Notes 2 and 3). Reduce the flow-rate of argon to 300mlmin-l and switch the flow of argon through the absorption cell.Pipette 10 ml of the buffer solution containing ammonium ion equivalent to 5 pg ml-l of nitrogen into the absorption compartment of the cell and insert the ion-selective electrode. Allow the argon to flow through the solution for 5 rnin and switch the flow of gas to the by-pass tube. Then increase the flow-rate of argon to 1.5 1 min-l and push the boat to the centre of the furnace tube. Continue the passage of argon for 2 rnin and then switch to a stream of hydrogen at a flow-rate of 140 mlmin-l. Immediately switch the flow of gas to the absorption cell a,nd continue to pass hydrogen through the cell until the electrode potential has stabilised, showing that no more ammonia is being released from the steel.For most steels this is a period of 120 min. Note the cell potential on the digital voltmeter and read off the concentration of nitrogen in the absorption cell from the calibration graph; then switch back to an argon stream. Determine the blank value for an empty silica boat after passing hydrogen through the apparatus for 120 min. NOTES- It is important that the inlet arm t o the absorption cell should be completely dry while the ammonia is being collected, otherwise some ammonia is absorbed in solution on the sides of the inlet arm before the sinter and low results are obtained. 1. 2. 3. Use steel millings with a t least one dimension less than 0.5 mm. For samples containing more than 0-0070 per cent. of mobile nitrogen use 0.5 g of steel.Calculation of the Concentration of Mobile Nitrogen in the Steel The passage of hydrogen through the absorption cell for 120 min caused the volume of the solution to diminish from 10.0 to 9.7 ml. The volume change is most easily determined by adding the appropriate buffer solution to the absorption cell from a microburette until the volume is restored to its original value, as shown by a mark on the outside of the cell. During this operation the cell is removed from the bath, tap 4 being open. Initial cell potential = 0.5 pg ml-l of nitrogen = 5 pg of nitrogen for 10 ml of solution Final cell potential = x pg ml-1 of nitrogen = x x 9-7 pg of nitrogen for 9.7 ml of solution Hence, nitrogen as ammonia from the hydrogen (the blank) = [(x x 9.7) - 51 pg Initial cell potential = 5 pg of nitrogen for 10 ml of solution Final cell potential = y pg ml-1 of nitrogen = y x 9.7 pg of nitrogen for 9.7 ml of solution Hence mobile nitrogen from the steel + blank = [(y x 9.7) - 51 pg Thus mobile nitrogen from the steel = [ ( y - x ) x 9.71 pg cent.For an empty tube: For the steel sample: For a 1-g sample the concentration of nitrogen in the steel = [(y - x) x 9-71 x low4 per Method for the Spectrophotometric Determination of Mobile Nitrogen2-* The equipment was identical with that for the ion-selective electrode method except that the ammonia in the stream of hydrogen from the furnace tube was absorbed in 50 ml of 0.001 M hydrochloric acid after passing through a larger No. 0 sinter. A collection time of 70 rnin was generally employed, with a flow-rate of 220 ml min-l.108 HEADRIDGE AND LONG: THE DETERMINATION OF MOBILE NITROGEN AutdYSt, VOl.101 Results Compositions of the Steels electrode methods are shown in Table I. The compositions of the steels analysed by both the spectrophotometric and ion-selective TABLE I COMPOSITIONS OF THE STEELS Concentrations of elements in the alloy, per cent. Alloy c Si Mn Ni A 0.31 0.67 0.87 0.04 C 0.30 0.67 0.87 0.04 D 0.29 0.65 0.86 0.04 12N 0.16 0.12 1.17 - 13N 0.15 0.17 0.97 - 14N 0.15 0.25 1-00 - 7 0.33 0.40 1.48 - B3275 (A) y7: (B) 0.28 1-23 0.61 0.04 0.07 1.56 1.11 21.4 Cr 0.03 0.03 0.03 - 0.03 15.0 A1 Mo 0.054 - 0.1 5 - 0.066 - - - - 0.32 - 0-05 Ti Cu Total N - 0.07 0-0108 - 0.07 0.0092 - 0.07 0.0122 0.024 - 0.0087 0.058 - 0.0220 0.0228 0.099 - - - 0.012 - - 0.012 - - 0.0169 Analysis of the Steels by Using the Ion-selective Electrode Method Results for the analysis of the steels by this method are shown in Table I1 and are also compared with the results obtained by the spectrophotometric procedure.The blank corre- sponded to 0.5 pg or less of nitrogen in 9.7 ml of solution. A typical recording of cell potential versus time for a steel is shown in Fig. 4. By using the calibration graph, these cell potentials were converted to concentrations of nitrogen as ammonium ion and a graph of concentration of nitrogen veysus time is also shown in Fig. 4. TABLE I1 RESULTS FOR THE DETERMINATION OF MOBILE NITROGEN I N STEEL SAMPLES USING THE OBTAINED BY THE SPECTROPHOTOMETRIC PROCEDURE ION-SELECTIVE ELECTRODE AND A COMPARISON OF THE RESULTS WITH THOSE Results for the determination of mobile nitrogen using the ion-selective electrode, Steel per cent.x lo4* 5.5, 3.5, 6.5, 5.5, 5.5, 6.5 6-5, 9.5, 6.5, 6, 8.6, 8.5 4.5, 7.5, 5, 5.5, 6, 6 61, 63.5, 63.5, 65.5, 61.5, 65.5 106, 109, 106, 109, 109, 109 46.5, 44.5, 43, 49, 41, 45 32.5, 31.5, 31, 30.5, 31, 32.5 31, 31.5, 35.5, 35.5, 36, 37 4, 4.5, 6, 5.5, 5, 6 A C D 12N 13N 14N 7 B3275 (A) :7*5, 5.5, 6, 9.5, 9, 8.5 B3275 (B) P74 Average result using the ion- selective electrode, per cent. x lo4 5.5 7.5 6 63.5 108 45 3 1-5 7.5 34.5 5 Standard deviation using the ion- selective electrode, per cent. x lo4 1 1.5 1 2 1.5 3 1 1.5 2.5 1 Average result using the spectrophoto- metric method, per cent. x lo4 6 9 6 64 1061. 45 28 8 36 7 Standard deviation using the spectrophoto- metric method, per cent.x lo4 2 2 3 7 ? 3 2 6 3 * Collection time 120 min for all samples. t Ammonia collected for 120 min. : Heat treated and quenched form of B3275 (B). For most steels the rate of release of ammonia was greatest after the passage of hydrogen for 15 min in the ion-selective electrode method, but ammonia continued to be released at a diminishing rate up to 2 h after switching from argon to hydrogen. A typical graph of the rate of accumulation of nitrogen as ammonium ion versus time is shown in Fig. 5. Discussion The results shown in Table I1 are considered to be very satisfactory, excellent agreement between the two methods being achieved. With the spectrophotometric procedure, steel 13NFebruary, 1976 IN STEEL USING AN AMMONIUM ION-SELECTIVE ELECTRODE 109 -570 8 17 0 0 10 20 30 40 50 60 70 80 90 100 110 120 Ti me/m in Fig.4. A, A recording of cell potential versus time for steel 12N. B, A graph of nitrogen concentration as ammonium ion in the collecting solution versus time for steel 12N. appeared to contain only 0.0098 per cent. of mobile nitrogen, but this was raised to 0.0106 per cent. when the collection period was increased from 70 to 120 min. This result illustrates the drawbacks of the spectrophotometric procedure with a fixed collection time. Different collection periods arose when using the two procedures because of the different sizes of the No. 0 sinters used in the absorbing solutions. A lower flow-rate inevitably leads to a longer collection period.It is evident from Table I1 that the precision of the ion-selective electrode method is superior to that of the spectrophotometric method, but its main advantage lies in the fact that the completion of the collection of ammonia for any steel sample is at once obvious to the operator. The use of a potentiometric recorder coupled to a digital voltmeter is not essential but it is a distinct advantage because the amount of ammonia collected is automatically recorded and, while this is occurring, the operator can be engaged in other work. ‘0 Tim e/m i n Fig. 5 . A graph of the rate of accumulation of nitrogen as ammonium ion in the collecting solution versus time for steel 12N. (C is the concentration of nitrogen and T the time). The response of the EIL, Model 1057 200, ion-selective electrode to changes in ammonium- ion concentration from more to less concentrated solutions was rather sluggish.The potential of the ion-selective electrode versus the reference electrode changed rapidly over a period of a few minutes, and then tended to drift to more negative values very slowly when the electrode was placed in the most dilute standard solution (05pgml-l of nitrogen) after being in a more concentrated ammonium-ion solution. For the first run on a particular110 HEADRIDGE AND LONG day the cell potential stabilised quickly at the start of the run because the ion-selective electrode had been standing overnight in the most dilute calibration solution, but for the second and later runs it was inconvenient to wait until the starting potential had stabilised completely before commencing the collection of ammonia.The starting potentials were often 1-2 mV more positive than the corresponding potential on the calibration graph. This was of no consequence as the cell potential that was required for the calculation of mobile nitrogen in a steel was that measured after the passage of hydrogen for 2 h, and a steady potential coiresponding to the concentration of ammonium ion in the absorption cell was always achieved under these conditions. It has been suggested that a suitable poly(viny1 chloride) membrane electrode for the ammonium ion would respond more rapidly than the glass membrane electrode to changes in the ammonium-ion concentration, but the EIL 1057 200 glass membrane electrode can be used almost indefinitely and its rate of response is sufficiently rapid for the work described in this study. All of the steels except P74 were low-alloy steels containing only trace amounts, if any, of vanadium and niobium. Therefore, the results obtained for mobile nitrogen from these nine steels should be reliable. Alloy P74 is a high-alloy steel and might possibly contain some chromium nitride, although silicon forms a more stable nitride than does chromium.6 However, if the alloy does contain chromium nitride there is little evidence for its breakdown at 600 "C because it can be seen from Table I1 that its mobile nitrogen content, 0.0005 per cent., is very low, although the total nitrogen content is 0.0169 per cent. The authors thank the Steel Castings Research and Trade Association for a maintenance grant (to G.D.L.) and for the loan of the furnace. They are grateful to Mr. D. G. Swinburn for his helpful discussions in connection with this project. References 1. 2. 3. 4. 5 Scholes, P. H., and White, G., Steel Times A . Rev. Steel Ind., 1970, 172. Fisher, R., and White, G. , Advanced Process Laboratory, British Steel Corporation, Research Report Jenkins, R. H., Open Rep. Br. Iron and Steel Res. Ass., MG/CC/520/72. Hill, R., and Swinburn, D. G., Steel Founders' Society of America Project No. 88, Restricted Report No. 4, Steel Founders' Society of America, New York, 1972. Pearson, J., and Ende, U. J. C., J . Iron Steel Inst., 1953, 175, 53. CAPL/SM/G/51/73. Received September 21st, 1976 Accepted October 16th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9760100103

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, February, 1976, Vol. 101, $9. 111-121 111 The Determination of Substituted Phenylurea Herbicides and Their Impurities in Technical and Formulated Products by Use of Liquid Chromatography J. A. Sidwell and J. H. A. Ruzicka Department of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE1 9NQ The application of liquid chromatography to the identification and deter- mination of the active ingredient and the impurities in phenylurea herbicides commonly employed in agriculture is described. Technical materials are dissolved in dichloromethane and chromatographed on microparticulate silica with dichloromethane or dichloromethane - methanol as eluting agent, or on microparticulate silica bonded with octadecyltrichlorosilane with methanol - water as eluting agent.An initial extraction procedure is required for dispersible powders. Detection was by means of ultraviolet absorbance. Substituted phenylurea compounds (urons) are widely used in agriculture as selective herbi- cides for the pre- or post-emergence control of various weeds. Technical materials and dispersible powders are usually analysed by means of acid or alkali hydrolysis with titration of the liberated aliphatic amine1-3 or a colorimetric determination of the aromatic amine f ~ r m e d . ~ Methods based on hydrolysis, however, lack specificity for individual urons, as impurities which may be present will be included in the results for the assay. Gas - liquid chr~matography~,~ can be used but suffers from the disadvantage that carefully controlled conditions are required in order to prevent thermal decomposition of the phenylurea, either on the column or during injection.A need therefore exists for a rapid and specific procedure for the determination of these compounds and their impurities in technical and formulated products. Methods have been developed that involve liquid chromatography, a technique first applied to analysis for some urons by Kirkland,' and these methods do not have the shortcomings of the procedures mentioned above. The urons examined are shown in Table I. Experimental Apparatus and Reagents All reagents were of analytical-reagent grade unless otherwise specified. The equipment used for liquid chromatography was of modular construction. Solvent delivery. A Waters Associates, Model 6000, constant-volume solvent delivery system was used.Sample injection. A Varian Associates stop-flow injector was used throughout this work. For maximum resolution samples were injected with a standard 10-p1 syringe on to a stainless- steel fine-mesh gauze fitted on top of the column packing. In order to prevent tailing of peaks, a needle guide was incorporated in the injector so that samples were introduced on to the centre of the column. On-column injection was not employed as it was found that the top part of the column gradually became disturbed, which resulted in a dramatic reduction in efficiency. Silica column. A slurry of LiChrosorb SI 60 5-pm silica packing (E. Merck, Darmstadt) in 2,2,4-trimethylpentane saturated with tetrachloroethylene was prepared. By using pressures of 5000-6000 lb in-2 the slurry was forced with 2,2,4-trimethylpentane into a 150 x 4-6 mm i.d.stainless-steel column fitted with a low dead-volume connector containing a frit of pore size 2 pm. The packed column required a pressure of 600 lb in-2 for a flow-rate of dichloromethane of 1 ml min-l. The column was maintained at 30 "C by means of a water-jacket. C,, bonded silica column. This column packing material was prepared in the laboratory by using the following procedure. Crown Copyright.112 SIDWELL AND RUZICKA: DETERMINATION OF PHENYLUREA HERBICIDES Analyst, Vol. 101 LiChrosorb SI 60 (5 g), dried at 120-140 "C for 2 h, was placed in a 100-ml centrifuge tube together with 50 ml of light petroleum (boiling range 60-80 "C, sodium-dried) and 20 ml of octadecyltrichlorosilane.The mixture was then shaken and placed in an ultrasonic bath for 30 min with shaking at intervals. The mixture was centrifuged, the light petroleum discarded and the packing treated with a further 50 ml of light petroleum and 20ml of trimethylchlorosilane to react with any remaining non-bonded surface-active sites on the silica. The packing, after removal of excess of reaction mixture by centrifuging, was then washed with three 50-ml portions of light petroleum, one 50-ml portion of propan-2-01 and two 50-ml portions of methanol. The bonded material was then packed into a 250 x 4.6 mm i.d. stainless-steel column fitted with a low dead volume connector containing a frit of pore size 2 pm, using a balanced- density slurry in chloroform - bromoform and pressures of 5000-6000 lb in-2.The packed column required a pressure of 2000 lb in-2 for a flow-rate of 1 ml min-l for water containing 60 per cent. of methanol. The column was maintained at 30" C by means of a water-jacket. Solvents. For chromatography on the silica column dichloromethane (containing 0.1 per cent. of methanol as stabiliser) was used as the eluting agent. For the determination of certain urons small amounts of methanol were added to the dichloromethane in order to increase the polarity. Mixtures of methanol (Spectro Grade, Eastman) and water, which were warmed and stirred continuously in order to remove air bubbles, were used with the cl8 bonded packing material. Detector. A variable wavelength ultraviolet detector (Cecil CE 212) fitted with a 10-p1 cell (1-cm light path) was employed.With methanol - water and dichloromethane as eluting agents, determination of the urons plus their impurities was carried out at 240 and 245 nm, respectively. A wavelength of 254 nm can also be used but measurement at this wavelength is not as sensitive. The detector output was connected to a 10-mV input chart recorder. Peak measurement. A digital integrator (Autolab 6300) was used to measure peak areas and a transparent bevelled rule to measure peak heights. It is important that sample solutions that are kept for more than 2 d before examination should be stored in the dark in order to prevent photodecomposition. All washings were removed after centrifuging. Methods Identity Test for Phenylurea Herbicides For positive identification, comparison of retention of the sample with that of a reference standard on at least two types of liquid-chromatography packing is recommended. Use of a 5-pm silica column and a c18 bonded 5-pm silica column meets this requirement.Xormal phase separation on a silica column Prepare solutions of the technical uron sample and the reference standard (about 1-5 mg of each) in 10 ml of dichloromethane. For dispersible powders extract into 10 ml of dichloro- methane about 1-5 mg of the uron active ingredient from a 5-ml slurry of the sample in water. By reference to Table I select a dichloromethane - methanol mixture of such a composition that elution of the uron from the column will occur in a reasonable time.Using a flow-rate of 06-1-0 ml min-l inject 2 pl of both sample and standard solutions, and set the detector sensitivity so that peak heights of 60-80 per cent. full scale are obtained. Check that the sample and standard are eluted with the same reten.tion time. Alternatively, for unknown urons, check the retention time of the sample against those obtained for a mixture of standard urons with dichloromethane as the eluting agent (Fig. 1). Reverse-phase separation on a cl8 bonded silica cohmn Prepare solutions of the technical uron sample and the reference standard (about 1-5 mg each) in 10 ml of methanol. For dispersible powders proceed initially as described for the silica column but evaporate the dichloromethane extract to dryness and dissolve the uron residue in 10 ml of methanol.By reference to Table I select a methanol - water mixture of such a composition that elution of the uron from the column will occur in a reasonable time. Using a flow-rate of 0.5-1.0 ml min-l, inject 2 p1 of both sample and standard solutions, and set the detector sensitivity so that peak heights of 60-80 per cent. full scale are obtained. Check that the sample and standard are eluted with the same retention time. Alternatively,February, 1976 AND THEIR IMPURITIES BY LIQUID CHROMATOGRAPHY TABLE I SUITABLE ELUTING AGENTS FOR THE LIQUID-CHROMATOGRAPHIC 113 DETERMINATION OF URONS Column f A \ Uron Chlorbromuron ci Diuron . . Linuron Structure Silica Chloroxuron . . Chlortoluron . . ci Methabenzthiazuron Metobromuron Metoxuron . . Monolinuron .. Monuron . , Dichloromethane ci ci 0 . . B r e N H ! N l c H 3 OCH3 CI C,, bonded silica 66% methanol in water 70% methanol in water 1% methanol in 66% methanol dichloromethane in water 0.5% methanol in 65% methanol dichloromethane in water Dichloromethane 65% methanol in water 0.5% methanol in 65% methanol dichloromethane in water Dichloromethane 60% methanol in water 1% methanol in SOY0 methanol dichloromethane in water Dichloromethane 60% methanol in water 1% methanol in 60% methanol dichloromethane in water114 SIDWELL AND RUZICKA: DETERMINATION OF PHENYLUREA HERBICIDES Analyst, Vol, 101 for unknown urons, check the retention time of the sample against those obtained for a mixture of standard urons with methanol - water as the eluting agent (Fig.2). Determination of the Content of Active Ingredient The following procedure, which involves the use of an internal standard, is suitable for the determination, using a silica column, of monuron, diuron and chlortoluron in technical materials and formulated products. The general principles can be applied to analysis for other urons. Prepare a sufficient amount of the eluting agent (dichloromethane containing 0.6 per cent. of methanol) for the analysis. Technical materials Accurately weigh 0.25g of uron sample and 0.30g of internal standard (acetanilide for monuron and chlortoluron and 4'-chloroacetanilide for diuron) into a 100-ml calibrated flask. Dissolve the contents of the flask in dichloromethane and dilute to 100 ml. Repeat with a reference standard uron of known content of active ingredient.Using a flow-rate of 1 ml min-l for the eluting agent and a column temperature of 30 "C inject separately several 1-pl portions of the uron sample and standard solutions on to the silica column, adjusting the sensitivity of the detector so that peak heights of approximately 80 per cent. full scale are obtained on the recorder. Calculate the content of active ingredient of the sample by com- parison of the peak heights or peak areas of the uron and internal standard, using the following equation : A , x I , x w, x 100 A , x I , x W , x P Active ingredient, per cent. = where A , and A , are the mean peak areas (or peak heights) for the uron sample and the standard uron, respectively; I , and I , are the mean peak areas (or peak heights) for the internal standard in the sample and standard solution, respectively; W , is the mass of uron sample taken; W , is the mass of standard uron taken; and P is the percentage content of active ingredient' of the standard uron.Dispersible powders Accurately weigh enough of the dispersible powder to contain 0.25 g of active ingredient into a 500-ml separating funnel. Add 50 ml of distilled water and shake to disperse the powder. Extract with three 100-ml portions of dichloromethane, collecting the dichloromethane layer in a round-bottomed flask. Evaporate to dryness by using a rotary evaporator (see Note). Transfer the residue quantitatively into a 100-ml calibrated flask by rinsing with three 10-ml portions of dichloromethane. Accurately add 0.30 g of internal standard and proceed as for technical materials.NOTE- adsorption characteristics of the column. Determination of Impurities Accurately weigh 0.25 g of uron into a 100-ml calibrated flask and dilute to volume with dichloromethane. Inject 2 p1 of the solution on to the silica column and examine the im- purities present by eluting the sample with the eluting agent suggested in Table I. Set the detector sensitivity so that the impurities are of sufficient peak height for accurate measure- ment. Identify the impurities present by comparison of their retention times with those of known compounds eluted under similar conditions. Unknown impurities can be identified by means of mass spectrometry after collection of the peak fraction eluted from the column following injection of 100 p1 of the solution.Determine the impurities by constructing a calibration graph with standard material. Should difficulty be experienced in dissolving certain impurities, e.g., substituted diphenylureas, in dichloromethane, use 1,4-dioxan to effect dissolution, warming if necessary, and dilute with dichloromethane to give a ratio of 1,4-dioxan to dichloromethane of 1 : 9 prior to injection. Results and Discussion Identity Test Figs. 1 and 2 show the separation of urons on 5-pm silica and C,, bonded 5-pm silica. A difference in elution order is given by these two columns. The separation mode of urons Complete removal of any remaining traces of water is advised in order t o prevent changes in theFebmary, 1976 AND THEIR IMPURITIES BY LIQUID CHROMATOGRAPHY 116 1 0.03 0-02 Q, c e El 2 0.01 3 I 5 9 6 7 8 10 Time/m in Fig.1. Separation of uron herbicides on a microparticulate (5 pm) silica column: 1, chlorbromuron; 2, linuron ; 3, monolinuron; 4, metobromuron ; 5 , methabenzthiazuron ; 6, diuron; 7, chlortoluron; 8, monuron; 9, chloroxuron; and 10, metoxuron. Mobile phase, dichloromethane at a flow-rate of 1-2 ml min- -1. 0.03 0-02 0.01 0 0 10 20 30 40 50 60 Time/m in Fig. 2. Separation of uron herbicides on a C,, bonded microparticulate (6-pm) silica column : 1, metoxuron; 2, monuron; 3, monolinuron; 4, metobromuron; 6 , diuron; 6, linuron; 7, chlor- bromuron ; and 8, chloroxuron. Mobile phase, methanol- water (3 + 2) at a flow-rate of 0.6 ml min-'.116 SIDWELL AND RUZICKA: DETERMINATION OF PHENYLUREA HERBICIDES Analyst, VoZ. 101 on the bonded packing is primarily partition whereas on the silica, separation depends on the extent of adsorption of the individual urons on the active sites of the packing.As a result, certain urons, e.g., chlorbromuron and chloroxuron, have similar retention times on the bonded packing but widely differing times on silica. These two columns can therefore be used for confirmation of the identity of urons. In Fig. 1, better resolution of the earlier peaks can be achieved by reducing the flow-rate or by decreasing the polarity of the eluting agent. The retention time of the later peaks will, however, be increased. Alternatively, a gradient elution separation technique could be used. Urons were also separated on a column packed with 5-pm silica bonded with 3-(trifluoro- methy1tetrafluoroethoxy)propyltrichlorosilane.Using water containing 40 per cent. of methanol as eluting agent, the elution order of the urons on this polar column was the same as on the less polar C,, bonded silica. However, the resolution of the urons was not as good. Quantitative Analysis For the determination of the content of active ingredient a procedure involving the use of an internal standard was adopted in order to overcome any dilution or injection errors. The analysis of technical monuron and diuron materials and dispersible powders has been studied in depth. The silica column was used rather than the C,, bonded silica column so that any substituted diphenylureas present as impurities could be determined under the same con- ditions.For monuron and diuron, acetanilide and 4’-chloroacetanilide, respectively, were selected as internal standards, being structurally related to and eluted closely after the uron but not having the same retention time as any impurity (Fig. 3). Masses taken were such that the peak heights of the internal standard and the uron were approximately the same. Relative retention times of certain urons and acetanilides are shown in Table 11. For repeat injections of solutions of the uron and internal standard, the coefficients of variation of the 1 *o 0.8 0 0.6 -2 51 n a 0.4 0.2 0 Fig. 3. A 0 5 10 15 Time/m in Separation of uron from internal standard in the quantitative determina- tion of diuron using a 5-pm silica column: A, diuron; B, 4’-chloroacetanilide.Mobile phase, dichloromethane con- taining 0.6 per cent. of meth- anol, at a flow-rate of 1 ml min-1.February, 1976 AND THEIR IMPURITIES BY LIQUID CHROMATOGRAPHY 117 peak-height and peak-area ratios were 0.38 and 0.39 per cent., respectively, for monuron and acetanilide and 0.40 and 0.54 per cent., respectively, for diuron and 4‘-chloroacetanilide. A small gradual increase in retention time was observed over a large number of injections. However, the relative retention times of the uron and the internal standard remained constant and thus the over-all precision of the method was unaffected. TABLE I1 RELATIVE RETENTION OF SOME URONS AND ACETANILIDES ON SILICA Eluting agent: 0.6% methanol in dichloromethane a t a flow-rate of 1 ml min-1.Relative Relative Uron retention Acetanilide reten tion Linuron . . .. . . 1.0 3’,4’-Dichloroacetanilide . . 2.7 Diuron . . .. . . 2.3 4’-Chloroacetanilide . . 3.2 Chlortoluron . . . . 2.8 Acetanilide . . . . 3.6 Monuron . . .. . . 3.1 Metoxuron . . .. . . 4.1 Various technical monuron and diuron materials and dispersible powders were analysed, using certified samples of urons (monuron, 99-92 per cent. ; diuron, 99.90 per cent.) as reference primary standards. For formulated products containing less than 90 per cent. of active in- gredient it was found necessary to take a mass of sample containing an amount of active ingredient comparable with that in the uron standard solution. Otherwise correction factors have to be applied in order to compensate for an observed non-linearity of the uron to internal standard ratio at low concentrations. The content of active ingredient of the samples was also determined by the CIPAC acid hydrolysis procedure in order to check the accuracy of the analysis by use of liquid chromato- graphy (Table 111).Uron samples were partitioned between 4 N hydrochloric acid and chloroform in order to remove any basic impurities or wetting agents, etc., and the residue, after evaporation of the chloroform, was hydrolysed by refluxing with 24 N sulphuric acid. After making the solution alkaline, the liberated dimethylamine was steam distilled, absorbed in excess of acid and back-titrated with standard sodium hydroxide solution. The results obtained by liquid chromatography were slightly lower than those obtained by hydrolysis.In the instance of monuron, sample 1, this difference can be accounted for by the presence of 0.34 per cent. of diuron. There would appear to be little significant difference between contents of active ingredient calculated from peak-height ratios or from integrated peak-area ratios. The determination of other urons by a similar procedure should also be possible provided that suitable internal standards are selected. When the determination of impurities is of secondary importance a C,, bonded packing can be substituted for the TABLE I11 COMPARISON OF RESULTS OBTAINED WITH LIQUID CHROMATOGRAPHY AND ACID HYDROLYSIS FOR THE ACTIVE INGREDIENT CONTENT OF MONURON AND DIURON Active ingredient, per cent., obtained by- p u i d chromatograp; y and calculated from- Sample Technical monuron .. (sample 1) (sample 2) Technical monuron . . Technical diuron .. Diuron dispersible powder acid hydrolysis .. 97.9 98.2 . . 98.0 97.9 .. 98-5 .. 79.0 79.3 integrated r g h t peak-area ratio ratio 97.8 97.4 97.7 97.2 97.8 974 97.8 97.9 98.3 98.0 98.0 78.9 79.1 98.2 98.2 98.3 78.6 78.6118 SIDWELL AND RUZICKA: DETERMINATION OF PHENYLUREA HERBICIDES Analyst, VoZ. 101 silica column. Monuron, for example, has been determined on a Corasil C,, column using diallyl phthalate as internal standard. Determination of Impurities The nature and extent of impurities present in monuron, diuron, linuron, metoxuron and chlortoluron have been investigated. The relative retentions of some possible impurities together with the percentage amounts of those identified in technical samples are given in Table IV.Typical chromatograms are shown in Figs. 4 and 5. Major impurities found were substituted diphenylureas formed from the reaction of excess of isocyanate (used in the manufacture of urons) with water. In general, these compounds have poor solubility in methanol and other polar solvents, which makes their determination on reverse-phase bonded packings difficult or impossible. For this reason the 5-pm silica column only was used for the determination of impurities. TABLE IV IMPURITIES IN URONS Some possible impurities Linuron . . .. .. .. .. .. 3,4-Dichloroaniline . . . . . . . . Methyl 3,4-dichlorophenylcarbamate . . .. 1,3-Bis( 3.4-dichlorophenyl) urea .. .. Monolinuron . . .. . . .. .. Diuron .. .. .. .. .. . . 3.4-Dichloroaniline . . .. .. .. 1,3-Bis( 3,4-dichlorophenyl)urea .. . . Monuron . . . . .. .. .. . . 4-Chloroaniline . . .. .. . . . . 1,3-Bis(4-chlorophenyl)urea . . .. . . Diuron . . . . .. . . .. .. Chlortoluron . . .. . . .. .. .. 3-Chloro-4-methylaniline . . . . . . 1,3-Bis (3-chloro-3-methylphenyl) urea . . 3-(4-Methylphenyl)-l, l-dimethylurea . . .. 3-Chloro-4-methoxyaniline . . . . . . . . 3-( 3-Chloro-4-methylphenyl)-l-methylurea . . Metoxuron . . .. .. .. .. .. 1,3-Bis(3-chloro-4-methoxyphenyl)urea . . Diuron . . .. .. .. .. .. 3-(4-Methoxyphenyl)-l, l-dimethylurea . . 3-( 3-Chloro-4-methoxyphenyl)-l-methylurea . . Relative retention 1*00* 0.4 1 0.48 1.3 1 1.66 1-00* 0.06 0.21 0.20 0.38 0.75 0.20 0.29 1-54 3.57 0.24 0.48 0.57 1.66 3 4 6 1*oot 1.oot 1.oot Impurities in technical samples, per cent.-- A B C - - - 4.8 0.05 3.4 1.5 2.7 1.5 0.1 0.1 0.4 - - 0.35 0.40 - 0.78 0.34 - 0.10 0.40 - 0.47 0.36 0.77 - * Silica column ( 5 pm) with dichloromethane as eluting agent at a flow-rate of 0.5 ml min-l. t Silica column (6 pm) with 1 per cent. methanol in dichloromethane as eluting agent a t a flow-rate of-0.5 ml min-l. All of the uron samples examined dissolved completely in dichloromethane. In preparing calibration graphs with which to determine impurities, it was noted that the rate of dissolution of certain compounds in dichloromethane was very slow. Impurity standards were therefore dissolved in 1,4-dioxan, by warming if required, and then diluted with dichloromethane to give a 1:9 ratio of 1,4-dioxan to dichloromethane.No differences were observed in the retention times for impurities dissolved in this way and for impurities partly dissolved in dichloromethane. No difference in the level of impurity was found whether the uron was dissolved in dichloromethane ?r in 1,4-dioxan - dichloromethane (1 + 9). The calibration graphs for the impurities were linear. The presence of methyl 3,4-dichlorophenylcarbamate in samples of linuron was verified by collecting the peak fraction following elution from the detector and analysing it by means of mass spectrometry (Fig. 6). A parent ion, m/e 218-9854, corresponding to C,H,NO%Cl, confirmed the identity of the impurity. The 1,3-bis(3,4-dichlorophenyl)urea present in diuron pyrolysed during mass spectrometry to give parent ions corresponding to 3,4-dichloro- aniline and 3,4-dichlorophenyl isocyanate, and was thus confirmed.Checks were made forFebruary, 1976 AND THEIR IMPURITIES BY LIQUID CHROMATOGRAPHY 119 A - 0.06 0.04 8 5 f! 5) 3 0.02 0 (a) A B D E 0 10 20 30 40 50 Time/m in 0 10 20 30 40 50 Time/min Fig. 4. Impurities in technical metoxuron. (a), Some possible impurities : A, unreacted 3-chloro-4- methoxyphenyl isocyanate ; B, 3-chloro-4-methoxyaniline ; C, 1,3-bis (3-chloro-4-methoxypheny1)urea ; D, diuron; E, metoxuron; F, 3-(4-methoxyphenyl)-l,l-dimethylurea; G, 3-(3-chloro-4-methoxyphenyl)-l- methylurea. (b), Technical sample : A, 1,3-bis(3-chloro-4-methoxyphenyl)urea, 0-47 per cent. ; B, diuron, 0.35 per cent. ; C, 3-(4-methoxyphenyl)-I, 1-dimethylurea, 0.77 per cent.Column, 5-pm silica. Mobile phase, dichloromethane containing 1 per cent. of methanol, a t a flow-rate of 0.5 ml min-I. 0.16 0.12 @ ii 9 2 0.08 0.04 3 0 5 Time/m in Fig. 5. Technical sample of linuron : A, methyl 3,4-dichlorophenylcarba- mate, 4.8 per cent. ; B, 1,3-bis(3,4-dichlorophenyl)urea, 1-5 per cent.; C, monolinuron, 0.1 per cent. Column, 5-pm silica. Mobile phase, dichloro- methane a t a flow-rate of 0.5 ml min-I.120 SIDWELL AND RUZICKA : DETERMINATION OF PHENYLUREA HERBICIDES Analyst, VoZ. 101 the presence of any free isocyanate in the urons by comparing chromatograms of urons dissolved in methanol-free dichloromethane (ie., methanol stabiliser removed) with chro- matograms of urons dissolved in dichloromethane - propan-2-01 (20 + 1).No changes in the chromatograms owing to reaction between isocyanates and propan-2-01 were observed and hence the absence of free isocyanate was inferred. 100 90 80 70 2 60 E c, 50 > c, .- - 2 40 30 20 10 0 h (M-OC 0 II (C-OCH3)' (M - OCH - 0)' \ M' 13 - H)' 0 20 40 60 80 100 120 140 160 180 200 2.20 m/e Fig. 6. in linuron). Mass spectrum of methyl 3,4-dichlorophenylcarbamate (an impurity During work on the determination of impurities it was noted that peaks were appearing in the chromatograms of uron solutions that had been standing in sunlight, indicating that photochemical breakdown had occurred. For example, with diuron, the methyl 3,4-di- chlorophenylcarbamate slowly appeared in the chromatogram. This is explained by : ci CI CH30H I Ci the methanol being present as stabiliser in the solvent. I t is therefore important that im- purities should be determined in freshly prepared solutions. The large amounts of methyl 3,4-dichlorophenylcarbamate found in linuron did not arise from photochemical breakdownFebrzlary , 1976 AND THEIR IMPURITIES BY LIQUID CHROMATOGRAPHY 121 of linuron on standing in solution. Its formation may arise either during manufacture, from side reactions between 3,$-dichlorophenyl isocyanate and NO-dimethylhydroxylamine in- volving the methoxy radical, or by photochemical breakdown of the solid. The authors thank the Government Chemist for permission to publish this paper, Dr. G. Cox for his advice on aspects of the liquid chromatography and Dr. K. S. Webb for carrying out the mass spectrometry. Ciba-Geigy Ltd., DuPont Ltd., Fisons Ltd., Hickson and Welch Ltd., Sandoz Ltd. and Staveley Chemicals Ltd. are-thanked for the supply of phenylurea herbicides and their impurities and also the National Physical Laboratory, Teddington, Middlesex, for the supply of pure samples of diuron, linuron and monuron. References 1. 2. 3. Yuen, S. H., and Milosevic, B., Analyst, 1969, 94, 820. Yuen, S. H., and Palmer, J. M. C., Analyst, 1972, 97, 921. Lowen, W. K., Bleidner, W. E., Kirkland, J . J., and Pease, H. L., in Zweig, G., Editor, “Analytical Methods for Pesticides, Plant Growth Regulators and Food Additives, Volume IV, Herbicides,” Academic Press, New York, 1964, p. 157. Yuen, S. H., Analyst, 1970, 95, 408. Katz, S. E., and Strusz, R. F., J . Agric. Fd Chem., 1969, 17, 1409. Buser, H., and Grolimund, K., J. Ass. 08. Analyt. Chem., 1974, 57, 1294. Kirkland, J . J., J . Chromat. Sci., 1969, 7, 7. 4. 5. 6. 7. Received August 21st, 1975 Accepted September 26th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100111

出版商:RSC    

年代:1976

数据来源: RSC