ISSN:0003-2654    

DOI:10.1039/AN97601FX045

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

ANALAO 101 (1 209) 921 -1 008 (1 976)ISSN 0003-2654December 1976THE ANALYSTTHE ANALYTICAL JOURNAL OF 'THE CHEMICAL SOCIETYCONTENTS921 EDITORIAL: SI UnitsORIGINAL PAPERSFluorimetric Method for the Determination o f Low Concentrations o f DissolvedAluminium in Natural Waters-D. J,, Hydes and P. S. LissImproved Method f o r the Determination o f Acrylamide Monomer in Waterby Means o f Gas - Liquid Chromatography w i t h an Electron-captureDetector-A. HashimotoElimination o f Ionic Interferences in the Determination o f Sulphates in WaterUsing the Lead -sensitive lon-select ive Elect rode-Ada m H ula n icki, R yszardLewandowski and Andrzej Lewenstam943 Automatic Apparatus f o r the Determination of pH and Nitrate in Soils-D. Goodman949 Determination of Molybdenum in Geological Materials by Atomic-absorptionSpectrophotometry-P. Sutcliffe956 Interference Effects in the Determination of Barium in Silicates by FlameAtom ic-a bsorpt ion S pect r o p hot of n e t ry- R. Cioni, A.M azzucotell i andG. OttonelloDetermination o f Trace Elements in Titanium(lV) Oxide Pigments by Atomic-absorption Spectrometry Using an Aqueous Slurry Technique-C. W. FullerSemi-automated Determination o f Lead by Hydride Generation and Atomic-absorption Spectrophotometry-Prem N. Vijan and George R. WoodSpectrophotometric Determination o f Trace Amounts o f Iron in Pure ReagentChemicals by Solvent Extraction as the Ternary Complex o f Iron(ll),4-Chloro-2-nitrosophenol and Rhodamine B-Kyoji TBei, Shoji Motomizuand Takashi Korenaga982 Determination of Residues of Oxamyl in Crops and Soils by Gas - LiquidC h rom at og rap h y-R .H . B rom i lo w986 An Ultraviolet Spectral and Polarographic Study o f Nitrofurantoin, a Urinary-tract Antibiotic-J. S. Burmicz, W. Franklin Smyth and R. F. Palmer992 Rapid Assay of Proteinase Activity in Papain by Measuring the Increase inOpalescence of a Casein Solution During Hydrolysis-G. S. Skelton,J. Orszggh, J. Gregoire, G. Parent and Katombe Lukusa922932939961966974996 Rapid Determination of Primary Amides-J. Ellis and A. M. Holland1001 Book Reviews1005 Notice t o AuthorsSummaries o f Papers in this Issue-Pages iv, v, viii, ix, xiPrinted by Heffers Printers Ltd, Cambridge, EnglandEntered as Second Class at New York, USA, Post OfficDecember, 1976 SUMMARIES OF PAPERS I N THIS ISSUERapid Assay of Proteinase Activity in Papain by Measuring theIncrease in Opalescence of a Casein Solution During HydrolysisX rapid method of assay of the activity of proteinase in papain is describedin which the increase in opalescence of a solution of casein as i t undergoeshvdrolysis by the proteinase is measured.,4 number of practical applicationsof the method are discussed.G. S. SKELTON, J. ORSZAGH, J. GREGOIRE, G. PARENT and KATOMBELUKUSADepartment of Chemistry, Universit6 Nationale du Zaire, Lubumbashi, Republicof Zaire.Analyst, 1976, 101, 992-995.Rapid Determination of Primary AmidesA method for thc determination of primary amides is presented that involvestreatment with excess of nitrous acid - sulpliuric acid a t 0-5 "C. Unreactednitrite is measured spectrophotometrically by a diazo-coupling reaction afterpassage through a copper-coated cadmium reducing column in order toreduce any nitrate ions produced by disproportionation of nitrous acid andby aerial oxidation of nitrogen oxide liberated during diazotisation of theamide. The technique can be used to determine 1.5 and 0.015 mg ofamide nitrogen with relative standard deviations of 1.4 and 4.2y0, respectively.J. ELLIS and A. M. HOLLANDDepartment of Chemistry, University of Wollongong, P.O. Box 1144, Wollongong,NSW 2500, Australia.Analyst, 1976, 101, 996-1000.x

ISSN:0003-2654    

DOI:10.1039/AN97601BX047

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

i V SUMMARIES OF PAPERS I N THIS ISSUE December, 1976Summaries of Papers in thisFluorimetric Method for the Determination of Low Concentrationsof Dissolved Aluminium in Natural WatersDissolved aluminium occurs in natural waters down to concentrations of1 pg 1-1 or less. There are very few methods that are sufficiently sensitive tomeasure down t o these levels. A fluorimetric method using the reagentLumogallion has been investigated and Eound to have a detection limit of0.05 pg 1-1 of aluminium and a coefficient of variation of 5% a t the 1.0 pg 1-1level and 2.7% a t the 22pg1-1 level. The only anionic interference likelyto be important in most natural waters is that from fluoride, and this inter-ference can be dealt with by using an incremental calibration procedure.Foracidic waters iron is a potential interferent but has no significant effect atconcentrations of less than 100 pg 1-l. I n water abnormally rich in dissolvedorganic material there may be competition for the dissolved aluminiumbetween the natural organic ligands and the Lumogallion. This effect canbe overcome by ultraviolet irradiation prior to analysis for aluminium. Theanalysis detects all forms of aluminium in :filtered natural water samples exceptwhen the aluminium occurs in stable mineral structures, e.g., clay particlessmall enough to pass through the filter. Aluminium adsorbed on the surfaceof such particulate material appears to be determined.D. J. HYDES and P. S. LISSSchool of Environmental Sciences, University of East Anglia, Norwich, NR4 7T J.Analyst, 1976, 101, 922-931.Improved Method for the Determination of Acrylamide Monomer inWater by Means of Gas - Liquid Chromatography with anElectron-capture DetectorA gas - liquid chromatographic method involving electron-capture detectionhas been developed for the determination of an acrylamide monomer presentin trace amounts in water.The method is based on bromination of the doublebond by means of an ionic reaction and the 2,3-dibromopropionamide pro-duced is extracted twice with 10ml of ethyl acetate after salting out withsodium sulphate. The 2,3-dibromopropionamide from the reaction mixtureis then cleaned up by using a Florisil column.The calibration graph for the method is linear over the range 0-50 pg ofacrylamide monomer and the limit of detection is 0.032 pg 1-1 for an aqueousstandard solution.The yields of the brominated compound are 85.2 &- 3.3and 83.3 & 0.9% a t fortification levels {of 1.0 and 5.0 p g l-l, respectively.Recoveries of acrylamide monomer from river water, sewage effluent and seawater spiked a t a level of 0.20 pg per 50 ml are 99.4 3.0and 98.0 3.2%, respectively. No interferences were observed fromsea water or from 8.0% of ammonium ions present as ammonium bromide.A. HASHIMOTOKitakyushu Municipal Institute of Environmental Health Sciences, Tobata-ku,Kitakyushu, Japan 804.Analyst, 1976, 101, 932-938.2.5, 101.December, 1976 SUMMARIES OF PAPERS I N THIS ISSUEElimination of Ionic Interferences in the Determination of Sulphatesin Water Using the Lead- sensitive Ion- selective ElectrodeIn the titration of sulpliate ions with lead(I1) perchlorate solution usinga lead-sensitive electrode for end-point determination, the addition of75% V / V of methanol permits titration of M samples.Under theseconditions, a 200-fold excess of nitrate does not interfere, but chloride a tthis concentration introduces a small systematic error of 476. Significanterrors are caused by the presence of calcium, which co-precipitates as calciumsulphate with lead sulphate. In titrations carried out rapidly, the negativeerror increases with the amount of calcium and the true equilibrium maybe obtained after a relatively long time. Increasing the ionic strength ofthe solution by addition of sodium perchlorate eliminates the interferenceby calcium, but for concentrations of sulphate below M it is advantageousto remove calcium by using a cation-exchange resin in the hydrogen form.The method has been applied successfully to the determination of sulphatein water.ADAM HULANICKI, RYSZARD LEWANDOWSKI and ANDRZEJ LEWEN-STAMInstitute of Fundamental Problcms in Chemistry, University of Warsaw, 02-093Warsaw, Poland.Analyst, 1976, 101, 939-942.VAutomatic Apparatus for the Determination of pH and Nitratein SoilsAn apparatus is described that automatically extracts and analyses batchesof up to 60 soil samples.Analysis is performed electrocliemically by meansof either a pH or ion-selective and reference electrodes.The electrodeoutput is amplified and recorded on a chart recorder, thereby allowing theelectrode performance to be monitored. Results of tests with both pH andnitrate ion-selective electrodes are reported. Good agreement was obtainedbetween automatic and manual methods. Difficulty was experienced withthe operation of an Orion nitrate electrode but tests using a Corning electrodeon the apparatus gave excellent recoveries and reproducibility.D. GOODMANNational Vegetable Research Station, Wellesbournc, Warwick, CV35 9EF.Analyst, 1976, 101, 943-948.Determination of Molybdenum in GeologicalMaterials by Atomic-absorption SpectrophotometryA simple, direct method for the determination of molybdenum in geologicalsamples by atomic-absorption spectrophotometry has been developed. Theinterferences caused by some common metals and acids have been investi-gated. The procedure is free from interferences, and a t least as accurateas the thiocyanate colorimetric method at molybdenum levels below 25 mg kg-lwhile still suitable for use at levels of more than 1 000 mg kg-l.P. SUTCLIFFEMount Morgan Limited, Mount Morgan, Queensland 4714, Australia.Analyst, 1976, 101, 949-955

ISSN:0003-2654    

DOI:10.1039/AN97601FP093

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

viii SUMMARIES OF PAPERS I N THIS ISSUEInterference Effects in the Determination of Barium in Silicatesby Flame Atomic-absorption SpectrophotometryA study has been made of single- and multi-element interactions of the elementspresent in a silicate matrix with barium during its determination by meansof atomic-absorption spectrophotometry. A general review of the problemis made and an experimental study is reported. It is necessary to separateoff the silicate matrix in the determination of trace amounts of barium.Such a separation is obtained by use of an ion-exchange technique and hasbeen described in previous papers. It is coincluded that in all geochemical workdetailed information on the technique employed in obtaining the analyticalresults is needed.R. CIONIIstituto di Mineralogia e Petrogrsfis, UniversitA di Pisa, Pisa, Italy.A.MAZZUCOTELLIIstituto di Petrografia, UniversitA di Genova, Genoa, Italy.and G. OTTONELLOScuola Normale Superiore di Pisa, Pjsa, Italy.December, 1976Analyst, 1976, 101, 956-960.Determination of Trace Elements in Titanium(1V) Oxide PigmentsAqueous Slurry TechniqueA simple procedure involving a slurry has been used for the determinationof trace elements in titanium(1V) oxide pigments by means of atomic-absorp-tion spectrometry. Both electrothermal and flame atomisation techniqueshave been employed. The method has been used to determine copper, iron,manganese and lead by dispersing the pigments in water that contained0.005y0 wz/ V of sodium hexametaphosphate. The limits for determinationare as low as 0.1 pg g-l using electrothermal atomisation and 2 pg g-l usingflame atomisation and a discrete sampling procedure.The advantages of this technique are its simplicity, speed of analysis andreduction of blank levels.C.W. FULLERTioxide International Limited, Stockton-on-'Tees, Cleveland, TS18 2NQ.by Atomic-absorption Spectrometry Using anAnalyst, 1976, 101, 961-965.Semi-automated Determination of Lead by HydrideGeneration and Atomic-absorption SpectrophotometryThe technique of gas-phase atomisation of the volatile hydrides in a heatedsilica tube and measurement of the resulting atomic-absorption signal hasbeen extended to the determination of lead in air, water and vegetation.The sample solutions, containing 0.7% V/V of nitric acid or 1.0% V / V ofperchloric acid, are made to react with l!2yo V / V hydrogen peroxide solutionfollowed by 4% wz/V sodium borohydride solution by means of a peristalticproportioning pump.The gaseous hydricles generated are swept by a stream ofnitrogen into a heated silica tube and the absorbance of lead a t 217 nm isrecorded. The effects of interferences and their elimination are discussed.Results obtained on a few typical samples by the proposed method and analternative method are presented.A sensitivity of 0.6 ng nil-l and a detection limit of 0.1 ng ml-1 have beenachieved. The relative standard deviation of the method is 2.5% at a leadconcentration of 10 ng ml-l.PREM N. VIJAN and GEORGE R. WOODAir Quality Laboratory, Ministry of the Environment, 880 Bay Street, Toronto,Ontario, Canada.Aaalyst, 1976, 101, 966-973.Up to 50 samples can be analysed in 1 dDecember, 1976 SUMMARIES O F PAPERS I N THIS ISSUESpectrophotometric Determination of Trace Amounts of Iron inPure Reagent Chemicals by Solvent Extraction as the TernaryComplex of Iron(II), 4-Chloro-2-nitrosophenol and Rhodamine BTrace amounts of iron in pure reagent chemicals have been determined bysolvent extraction - spectrophotometry with 4-chloro-2-nitrosophenol andRhodamine B. The ternary complex of iron(I1) , 4-chloro-2-nitrosophenoland Rhodamine B was extracted quantitatively into toluene a t about pH 4.8.The absorbance of the organic phase was measured in a glass cell of 10-mmpath length a t 558nm. The apparent molar absorptivity of the ternarycomplex in the organic phase was 9.0 x 1041mol-1cm-1 a t 558 nm.Theternary complex was very stable and not decomposed by addition of EDTA.By using the above procedure, trace amounts of iron (10-7-10-4~0) in alkalimetal salts, alkaline earth metal salts, ammonium salts, acids and bases, etc.,were determined. The standard deviations of the determinations were 1-3%.KYOJI TaEI, SHOJI MOTOMIZU and TAKASHI KORENAGADepartment of Chemistry, Faculty of Science, Okayama University, 3-1-1, Tsushi-manaka, Olrayama-shi, 700, Japan.Annlysf, 1976, 101, 974-981.ixDetermination of Residues of Oxamyl in Crops andSoils by Gas - Liquid ChromatographyA method is described for the determination of oxamyl residues in cropsand soils by gas - liquid chromatography.Oxamyl is extracted with di-chloromethane or acetone - dichloromethane and is separated from inter-fering co-extractives by chromatography on a Florisil column. I t is thendetermined by gas - liquid chromatography 1- ring on-column reaction withtrimethylphenylammonium hydroxide, the dei-ivative so formed being deter-mined by a flame-photometric detector operated in the sulphur mode. Approxi-mately 0.5 pg of oxamyl can be detected in a 50-g sample.R. H. BROMILOWChemical Liaison Unit, Rothamsted Experimental Station, Harpenden, Hert-fordshire, AL5 2JQ.Analyst, 1976, 101, 982-985.An Ultraviolet Spectral and Polarographic Study of Nitrofurantoin, aUrinary- tract AntibioticNitrofurantoin, a urinary-tract antibiotic, exists in four different forms overthe pH range 0-14, corresponding to the divalent cation H3A2+, the mono-valent cation H,A+, the neutral species HA and the monovalent anion A-,as investigated by means of ultraviolet spectrophotometry.The compoundis polarographically reducible over the whole pH range, the nitro group beingreduced to the hydroxylamine group in a 4e process and subsequently theamine being formed in a 2e process a t a pH value below 5. The C=N--Nlinkage is reduced by a mechanism involving reductive fission of the N-Nbond. The best defined differential pulse polarographic waves for thereduction of the nitro group to hydroxylamine are obtained in the pH rangeExperiments involving spiking of urine have been carried out and a methodis suggested for the determination of nitrofurantoin in urine over the concen-tration range 2-24 pg ml-l.J. S. BURMICZ, W. FRANKLIN SMYTHChemistry Department, Chelsea College, University of London, Manresa Road,London, SW3 6LX.and R. F. PALMERG. D. Searle and Co. Ltd., Lane End Road, High Wycombe, Buckinghamshire.6-12.Analyst, 1976, 101, 986-991

ISSN:0003-2654    

DOI:10.1039/AN97601BP097

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

DECEMBER 1976 Vol. 101 No. 1209 The Analyst Ed itor ial SI Units There are obvious advantages that would accrue from the adoption of a universal system of units for use in scientific publications. Clearly, comparison of data given by different authors in different journals would be much facilitated if the numerical information provided were expressed in the same system of units. Currently, the nearest approach to a universally accepted system is that of SI units. In common with many other journals, The Analyst has for a number of years encouraged authors to use the units recommended and allowed in the SI system. This is reflected in the infor- mation on units and symbols given in the “Notice to Authors,” which is published periodically in The Analyst and which appears in full in this issue (see p.1005). The obvious and most effective stage at which to ensure conformity to SI units is in pre- paration of the author’s manuscript. Unfortunately, old habits die hard, and a proportion of authors submitting papers to The Analyst persist in using units that have been superseded in the SI system. This practice poses a problem to the editorial staff. When conformity to the SI system merely involves a change in the nomenclature of the units involved, with no arithmetical conversion, e.g., the change of cycles per second to hertz, it is simple though time consuming to make such changes editorially and it has been the practice to do so. How- ever, when conversion would require juggling with numbers, e.g., a change of the unit of pressure from pounds per square inch to pascals, it is impracticable for the editorial staff to undertake all of the calculations necessary for conversion.For this reason, an author’s use of non-SI units is likely to result in their appearance in the finally published text. Clearly, then, we need to enlist the help of authors. This Editorial is directed to all authors intending to submit papers to The Analyst, and is a plea to them to ensure that their manuscripts conform as far as possible to the usage of units recommended in our “Notice to Authors.” Only with the effective co-operation of authors can the goal of a uniform system of units in the journal be achieved. Basically, the problem is that of requiring people brought up on older units to make the effort to use the newer units. Perhaps one might even suggest that there could be a measure of self-interest in their making such an effort, in addition to the obvious one of the need to comprehend publications that use the newer units. With SI units being taught in schools, there is a danger of a schism between those who are reluctant to adopt these units and the younger fraternity who will have known no other system. It would be a sad day for the scientific community if such a generation gap were allowed to develop. H. J. Cr,urmr Chairnzaii, The A pzalyst Publicatiom Coinnzittcc 92 1

ISSN:0003-2654    

DOI:10.1039/AN9760100921

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

922 Analyst, December, 1976, Vol. 101, p p . 922-931 Fluorimetric Method for the Determination of Low Concentrations of Dissolved Aluminium in Natural Waters D. J. Hydes" and P. S. Liss School of Environmental Sciences, University of East Anglia, Norwich, NR4 7T J Dissolved aluminium occurs in natural waters down to concentrations of 1 pg 1-1 or less. There are very few methods that are sufficiently sensitive to measure down to these levels. A fluorimetric method using the reagent Lumogallion has been investigated and found to have a detection limit of 0.05 pg 1-1 of aluminium and a coefficient of variation of 5% at the 1.0 pg 1-1 level and 2.7% a t the 22 p g 1-1 level. The only anionic interference likely to be important in most natural waters is that from. fluoride, and this inter- ference can be dealt with by using an incremental callibration procedure.For acidic waters iron is a potential interferent but has no significant effect a t concentrations of less than 100 p g 1-l. In water abnormally rich in dissolved organic material there may be competition for the dissolved aluminium between the natural organic ligands and the Lumogallion. This effect can be overcome by ultraviolet irradiation prior to analysis for aluminium. The analysis detects all forms of aluminium in filtered natural water samples except when the aluminium occurs in stable mineral structures, e.g., clay particles small enough to pass through the filter. Aluminium adsorbed on the surface of such particulate material appears to be determined. The object of this work was to find a method for the determination of aluminium that would be sufficiently sensitive for it to be used to investigate changes in dissolved aluminium concentrations during the mixing of river and sea waters in estuaries, where the over-all change may be from the order of 10 pg l-l, or more, of aluminium in river water to only 1 pgl-l, or less, in sea water.Reported spectrophotometric methods are not sufficiently sensitive for use at these 1evels.l Recent methods, involving chelation of the aluminium with an organic ligand and subsequent determination of the resulting complex by gas chromatography2 or atomic-absorption spectr~pfiotometry,~ have detection limits of about 2-3 pg 1-1 of aluminium and are therefore not sufficiently sensitive for application to estuarine samples of moderate to high salinity.Measurements of the required sensitivity can be made by atomic-absorption spectroscopy using a heated graphite tube for atomisation4s5 in fresh or waters of low salinity, but in sea water the analysis cannot be carried out directly on the water because of interferences caused by the high salt ~ o n t e n t . ~ Fluorimetry using the diazodihydroxy dye Solochrome Dark Blue has been applied to sea-water determinations6 but is not suitable for use with estuarine samples because of the wide range of interferences to which it is subject. More recently Lumogallion [3-(2,4-dihydroxyphenylazo) -2-hydroxy-5-chlorobenzenesulphonic acid] has been reported as being a sensitive and relatively interference-free reagent for the deter- mination of aluminium in natural In this paper, we describe a re-investigation of the conditions of the Lumogallion method and describe its use in natural waters. Iron was considered to be the most likely serious interferent, but the results presented here show this interference to be less than previously reported.8 Considerable interference can occur in waters rich in organic matter, which complexes the aluminium preferentially to Lumo- gallion.Ultraviolet radiation, which has been used previously to oxidise organic matter in natural waters to carbon dioxide in the determination of rnicron~trients,~ is used here as a method of making complexed aluminium available for determination. The reproducibility of the method is generally better than 5% when applied to natural waters. However, what remains uncertain is the nature of the aluminium species determined. The analytical conditions are not sufficiently rigorous to detect aluminium present in very * Present address : Sedimentpetrographisches Institut, Universitiit Gottingen, D-3400 Gottingen, West Germany.HYDES AND LISS 923 fine clay particles (which are able to pass through the filter membrane used) because of their very low solubilities at the pH used in the analysis.Atomic-absorption spectroscopy using a heated graphite tube for atomisation should provide an analysis that detects all aluminium species, the temperature of the tube reached during atomisation being sufficient to break down clay structures. The results of analyses by fluorimetry and atomic-absorption spectroscopy with graphite tube atomisation for clay suspensions and natural water samples are compared.Experimental The method involves buffering the sample to a pH of 5.0, adding the Lumogallion reagent and allowing it to react with the dissolved aluminium to form a complex whose fluorescence is measured using an excitation wavelength of about 465nm and an emission wavelength of about 555 nm. Optimum pH, Analytical Wavelengths and Buffering System The basic analytical conditions reported previouslya were re-investigated, using standard solutions in distilled water and natural river and sea waters. No significant differences were found. There was a slight discrepancy between the position of the optimum analytical wavelengths, probably owing to differences in the characteristics of the detection systems and wavelength calibration of the machines used.The graph of solution pH venzts fluorescence intensity was found to have a maximum at pH 5.0, which decreased approximately 2% within kO.1 pH unit so that precise buffering at pH 5.0 is not critical. A sodium acetate - acetic acid buffer mixture was found to provide an adequately stable pH during the analytical procedure. Rate of Formation of the Lumogallion - Aluminium Complex For solutions of aluminium in distilled water, the formation of the complex with Lumo- gallion exhibits second-order kinetics. For low aluminium concentrations with the Lumo- gallion present in excess, the kinetics approximate to first order. Reaction half-times at laboratory temperatures were compared for a number of different waters (Table I).Measure- ments were made by rapidly mixing the sample plus reagents and then transferring the mixture directly into the fluorimeter. The rate of formation was then followed as the increase in fluorescence. In natural waters the rate is slower than in distilled water, probably as a result of the aluminium being present as polymeric species and in complexes t h a t reduce the concentration of free aluminium ions. TABLE I HALF-TIMES FOR DEVELOPMENT OF LUMOGALLION - ALUMINIUM FLUORESCENCE I N DIFFERENT TYPES OF WATER Half-timeslmin r > Water Natural aluminium Added aluminium 5 Distilled . . . . River Yare . . . . 16 16 River Conway . . . . 70 16 Sea . . . . . . 22 22 - By heating the mixtures under analysis on a water-bath for 14 h, complete development The complex when formed of the fluorescence was achieved with all of the waters tested.was stable for at least 48 h. Effect of Temperature Increasing the temperature of the solutions under analysis produces a marked decrease in the fluorescence yield equivalent to approximately 1% "C-l. Samples must therefore be cooled after heating in order to improve the fluorescence yield, and must be allowed to equilibrate with the laboratory temperature in order to reduce the amount of inter-sample replicate variation.924 HYDES AND LISS : FLUORIMETRIC 1)ETERMINATION OF LOW Analyst, VOZ. 101 Reagents and Contamination Deterioration of the 0.02y0 Lumogallion solution, noticeable as an increased reddening of the solution after about 1 month, produces erratic results.A buffer solution of sodium acetate and acetic acid is used that is approximately 4 M with respect to acetate. When it is prepared from. analytical-reagent grade sodium acetate, the reagent blanks are found to be very variable, ranging from undetectable to equivalent t o 1 pg 1-1 of aluminium. When the solution is prepared from sodium acetate recrystallised from a filtered solution of the analytical-reage:nt grade chemical, contamination is not detectable (equivalent to less than 0.05pg1-1 of aluminium). All reagents should be kept as free from dust as possible. Throughout the work, determinations were carried out in wide-necked polypropylene bottles of nominal capacity 100 ml. New bottles give high and variable blanks but are easily cleaned by soaking with 10% hydrochloric acid. Bottles can be used several times with only thorough rinsing with distilled water between analyses.However, for optimum repro- ducibility, a rinse with acid should be included. When determinations were carried out using glass beakers, the blanks measured were high and variable, equivalent to 1-3 pg 1-1 of aluminium. Contamination of the Lumogallion reagent is nol; detectable. Range of Applicability The standard deviation for the analysis of ten replicates of a sample containing 1.0 pg 1-1 of aluminium was equivalent to 0.05 pg l-l, which is the probable practical detection limit of the method as described here. Using a :Luniogallion concentration of 2 mg 1-1 in the solution under analysis, the calibration graph is linear up to 120 pg 1-1 of aluminium.In the presence of interfering species that reduce the effective Lumogallion concentration, the length of the linear portion of the calibration graph will be correspondingly reduced. I t is possible to move the linear portion of the calibration graph to aluminium concen- trations higher than 120 pg 1-1 by increasing the Lumogallion concentration. However, at the emission wavelength, absorption by the uncoinplexed Lumogallion is stronger than that of the Lumogallion - aluminium complex (Fig. 1). In solutions with high Lumogallion concentrations (above 2 mg 1-l) and low levels of dssolved aluminium, absorption by the uncomplexed Lumogallion can be significant and lead to non-linearity of the calibration graph. The magnitude of this effect will depend on the path length of the emitted radiation in the analytical cell of the spectrometer.600 500 400 Wavelength/nm Fig. 1. Absorption spectra, measured against distilled water, for: A, 2 mg 1-1 of Lumogallion + 100 pg 1-1 of Al; B, 2 mg 1-1 of Lumogallion; and C, 2 mg 1-1 of Lumogallion + 2150 pg 1-1 of Fe + 20 pg 1-1 of Al.December, 1976 CONCENTRATIONS OF DISSOLVED ALUMINIUM IN NATURAL WATERS 926 Storage of Samples The following conditions were investigated for the storage of filtered river- and sea-water samples in polyethylene bottles prior to analysis: unfrozen; unfrozen and acidified to pH 2.0; and frozen a t -20 "C; all samples were stored in the dark in order to inhibit biological activity. The aluminium concentration in the bulk sample of the test water was measured.Aliquots of the sample were then stored in separate 500-ml polyethylene bottles that had been cleaned with 50% hydrochloric acid and distilled water and thoroughly rinsed with the sample water before being filled. For the unfrozen samples, a gradual decrease in the aluminium concentration occurred over the 14-day observation period; both the observed river- and sea-water aluminium concentrations decreased by approximately 10%. Storage of acidified samples was found to be impracticable owing to large and variable contamination (1-5 pg 1-1 of aluminium) from the additional reagents that had to be added to these samples. No change in the concentration in the frozen sea water was observed. The analysis of the frozen river water was complicated by the need to allow 2 d for the samples to equilibrate after thawing, as the reproducibility of replicate analyses is poor after shorter times.This effect probably occurs because aluminium is co-precipitated as a white precipitate that is sometimes observable on thawing river-water samples. A similar need to allow an equilibration period for silicate analyses has been reported.1° On the basis of these results, which will be reported in greater detail elsewhere, it is recommended that samples which cannot be analysed within a few days of collection should be stored frozen rather than acidified. Method Apparatus Knick 233 pH meters were used. bottles and Eppendorf microlitre automatic pipettes were employed. Reagents Perkin-Elmer 204 and Zeiss ZFM4 fluorescence spectrophotometers and EIL 7030 and Polyethylene and polypropylene reagent and reaction Lumogallion solution, 0.02yo in distilled water.B u f e r solution. Recrystallised sodium acetate plus acetic acid (4 M with respect to acetate), Standard aluminizwz solution. Dissolve 1.758 g of analytical-reagent grade aluminium Tests showed that this solution is stable A 0.5-ml volume of this solution diluted 1 + 999 and added to 50 ml of adjusted to give a pH of 5.0 in the water type being analysed. potassium sulphate in 100ml of distilled water. for at least 3 weeks. sample increases the aluminium concentration of the sample by 10 pg 1-l. Procedure Dispense 50-ml portions of a well shaken sample into the reaction bottles using a 50-ml measuring cylinder or, when optimum precision is required, an automatic burette.Add the reagents, 0.5 ml of the buffer solution and 0.25 or 0.5 ml of the Lumogallion solution (use 0.5 ml when concentrations higher than 30 ,ug 1-1 of aluminium are expected). Shake the bottles, transfer them to a water-bath and heat at approximately 80 "C for 14- h. The bottles should then be cooled and allowed to equilibrate with laboratory temperature. Finally, the fluorescence of the samples is measured using an excitation wavelength of 465 nm and an emission wavelength of 555 nm. Calibration Because the concentration of interfering species may vary from sample to sample and because of the temperature dependence of the fluorescence yield, the samples need to be calibrated by an incremental addition of aluminium. Normally, for estuarine work five 50-ml sub-samples were taken, to two of which an aluminium increment corresponding t o approximately one third of the expected aluminium concentration in the sample was added.Blanks Fluorescence not produced from aluminium in the sample can arise from either impurities in the reagents or when the natural fluorescence spectrum of the sample overlaps with that926 HYDES AND LISS : FLUORIMETRIC DETERMINATION OF LOW Analyst, VoZ. 101 of the aluminium - Lumogallion complex. The portion of the blank due to impurities in the reagents can be dealt with by replacing the sample with an equal volume of aluminium- free water. Doubly de-ionised water can contain a n undetectable amount of aluminium, but singly distilled water may occasionally contain over 0.5 pg 1-1 and often up to 0.3 pg 1-1 of aluminium.For water containing aluminium, the reagent blank can be determined by measuring the difference between aliquots of the same sample containing the normal and twice the normal amounts of reagents. This fluorescence difference is then subtracted from the fluorescence produced by the sample plus the normal amounts of reagents. For a reagent blank determined in distilled water, because the ratio of the fluorescence intensity to alu- minium may be different from that in the sample owing to interference, the amount of aluminium in the reagents must first be calculated and then subtracted from that found in the sample. Any blank due to natural fluorescence can be quantified by analysing a second aliquot of each sample, but with the omission of the Lumogallion reagent.The fluorescence of this solution is then subtracted from that of the solutions that included the Lumogallion reagent. Interferences and Application to Natural Waters Anionic For an anion, the magnitude of any interference is governed by the stability of its complex or complexes with aluminium, compared with the stability of the Lumogallion - aluminium complex. Of the ions that are likely to occur in significant amounts in natural waters, fluoride forms the strongest complex with aluminium. The next most stable complexes are those with phosphate and sulphate, but they have equilibrium constants which are 103 and 106 times smaller, respectively, than that for the aluminium fluoride complex.11 For an aluminium solution in distilled water containing 5 pg 1-1 of aluminium and 1 mg 1-1 of Lumogallion, a 5% reduction in the fluorescence intensity is produced by 0.3 mg 1-1 of fluoride added a s sodium fluoride.Under the same conditions, phosphate added as potassium dih ydrogen orthophosphate produces an equal reduction at a concentration of 3.0 mg 1-1 of phosphate. No interference was detectable with concentrations of sulphate added as sodium sulphate up to 5.0 g 1-1 of sulphate. These phosphate and sulphate levels are unlikely to be exceeded in natural waters. However, fluoride attains a concentration of 1.5 mg 1-1 in sea water.12 The relationships between the amount of the aluminium - Lumogallion complex formed and the amounts of fluoride and alumininm in the solution are sufficiently linear for it to be possible to measure the actual amount of aluminium present in natural waters when samples are calibrated with an aluminium increment (Figs.2 and 3). For saline waters, the extent of fluoride interference is less than expected, as the effective fluoride concentration 0 0.5 1.0 1.5 2.0 Fluoride/mg I-' 90 I 1 Fig. 2. Effect of increasing concentra- tions of fluoride on the fluorescence intensity produced in samples containing 6 pgl-l of A1 and 1 mgl-1 of Lumo- galhon. 0 2.5 5.0 7.5 10.0 Added aluminium/pg I-' Fig. 3. Calibration graphs for: A, fresh water containing 4.1 pg 1-1 of Al; and B. sea water of salinity 3.2% containing 2.2 pg I-' of Al.December, 1976 CONCENTRATIONS OF DISSOLVED ALUMINIUM IN NATURAL WATERS 927 is reduced owing to its complexing with other cations, principally magnesium.For sea water, the reduction in fluorescence intensity is approximately 10%. Cationic In natural waters, the only cation likely to be present in sufficient concentration to produce interference is iron. The results presented here disagree with the magnitude of the inter- ference from iron reported previously.8 Iron interferes because it forms a non-fluorescent complex with Lumogallion, which has two results. Firstly, it reduces the length of the linear section of the calibration graph of aluminium concentration versus fluorescence intensity by reducing the effective Lumogallion concentration in solutions in which it is present. Secondly, it reduces the fluorescence yield because the complex it forms with Lumogallion absorbs more strongly at the analytical wavelength than does the Lumogallion itself (Fig.1). For a total aluminium concentration of 20 pg l-l, the normal calibration procedure can be used in the presence of iron concentrations of up to 50 pg 1-1 with a Lumogallion Concentration of 1 mg 1-1. By increasing the Lumogallion concentration to 2 mg l-l, iron concentrations of up to 1OOpg1-1 can be tolerated (Table 11). Lower iron concentrations can produce TABLE I1 EFFECT OF IRON ON THE DETERMINATION OF ALUMINIUM IN DISTILLED WATER SOLUTIONS Iron added as FeCI,/pg I-' 0 10 20 50 100 150 Apparent aluminium detected/pg 1-1 Lumogallion Lumogallion 20 20 20 20 20 20 20 20 17.6 20 14.3 16.5 A t 'r 1 mg 1-l of 2 mg 1-1 of much greater apparent decreases in the fluorescence yield as a result of the co-precipitation of aluminium from solution by iron, when solutions with iron added to them are allowed to stand overnight prior to analysis (Fig.4). No effect of precipitation was detected when the analyses were carried out immediately after the addition of the iron. For all of the tests reported here, there was no significant difference when the iron was added as iron(II1) chloride or iron(I1) sulphate. In natural waters, stable free iron ion concentrations sufficiently 0 10 20 30 40 50 Iron/pg I-' Fig. 4. Effect of increasing concentrations of iron on the fluorescence intensity produced by Lumogallion in solutions of iron and 10 p g 1-' of A1 that had been allowed to stand over- night prior to analysis.928 HYDES AND LISS: FLUORIMETRIC DETERMINATION OF LOW AnaZyst, Vol.101 high to cause interference will be rare. In some waters, such iron may be present as organic complexes or as fine particulate material, the stability of which will determine how effectively they will interfere in the aluminium analysis. Organic No accurate prior assessment can be made of the likely interference from organic matter because of the wide range of species and concentrations that occur in natural waters. Organic matter will produce interference by competitively complexing the aluminium, but it should be possible to allow for this effect by using a spike calibration. Problems may also arise when the organic matter has a natural fluorescence which overlaps with that of the Lumo- gallion - aluminium complex.These problems can be dealt with by measuring the fluorescence of a duplicate sample to which no Lumogallion hits been added, as discussed under Method. Only one series of field samples collected during the course of this work contained sufficient organic matter to cause interference. This water came from the River Beaulieu, Hampshire, which directly drains the acidic peats and bogs of the New Forest. The fresh water had a brown coloration even after filtration and had a natural fluorescence at the analytical wavelength equivalent to 3 pg 1-1 of aluminium in distilled water. The concentration of organic matter in the fresh water was such that calibration increments were complexed to such an extent as to make them undetectable. For estuarine samples, the fluorescence from the aluminium increments increased with salinity and was equivalent to that in sea water at 2.7% salinity. Although no fluorescence from the aluminium increments was detectable, the fluorescence of the fresh-water samples after reaction with Lumogallion was considerably higher than the natural fluorescence of the water, suggesting that some of the organic material reacts with the reagents to produce a second source of non-aluminium fluorescence.An apparently similar type of interference can be reproduced in the laboratory. Humic acid (Fluka) at a concentration of 5 mg 1-I produces a 25% reduction in the fluorescence yield from a solution containing 2Opg1-1 of aluminium plus 2 mgl-1 of Lumogallion.It also has a strong natural fluorescence centred at 520 nm, which overlaps the Lumogallion - aluminium fluoresence, giving an increase in the apparent aluminium concentration of 15 pg 1-l. After the analyses had been attempted in the normal way, it was possible to carry out a limited number of experiments with the rernaining samples in order to assess whether or not destruction of the organic matter by photo-oxidation using an ultraviolet irradiator would make the aluminium available for determination by Lumogallion. Measurement of Aluminium in Water with a. High Content of Organic Material An ultraviolet irradiation apparatus similar to that described previouslyg was used. The samples were placed in quartz tubes of capacity 100 ml surrounding a central 1000-W ultra- violet tube (Hanovia) and irradiated for various periods of time.Treatment of samples Sample A was filtered through a 0.45-pm membrane filter and irradiated for 15 h. A brown precipitate formed and was filtered off. Samples B, C and D were filtered through 0.45--pm membrane filters, 0.5 ml of 20-volume hydrogen peroxide was added to each tube, followed by irradition for 7.5 h. A brown pre- cipitate formed and was dispersed by ultrasonic shaking before the samples were analysed. Results The results of the irradiation of the four samples are given in Table 111. Irradiation with and without peroxide completely destroyed the natural fluorescence present in the untreated samples. The recovery of added aluminium was much improved but in none of the samples was it restored to the level in distilled water.This result may be due either to the presence of resistant organic complexing species that are riot broken down by the irradiation or to the production of secondary species (e.g., inorganic phosphate) that will compete with Lumo- gallion for aluminium. The cause of this interference needs to be identified and further work should be carried out in order to find the optimum irradiation times and amount of oxidising agent needed ; also, the possibility that the brown precipitate converts aluminium into an unreactive form should be checked before any definite procedure can be recommended for the use of an ultraviolet irradiator to remove organic interferents.Decenzber, 1976 CONCENTRATIONS OF DISSOLVED ALUMINIUM IN NATURAL WATERS 929 TABLE I11 RESULTS OF USING ULTRAVIOLET RADIATION TO BREAK DOWN ORGANIC MATTER IN RIVER BEAULIEU WATER Natural Fluorescence fluorescence Fluorescence yield Salinity, Peroxide before after (20 p g 1-l A1 Apparent Sample % added analysis analysis spike) Al/pg 1-1 A : before 0 11 80 7 197 after No 0 58 38 30.5 B : before 0.17 9 36 16 31.1 after Yes 0 80 54 29.6 C : before 0.05 10 65 28 39.3 after Yes 0 186 88 42.3 D : before 0.15 13 63 21 47.6 after Yes 0 124 78 31.8 Distilled water - - - 110 - Speciation of Aluminium in Natural Waters Size Fractionation and Fluorimetric Analysis For some natural waters, differences were observed in the amount of aluminium detected when samples were filtered through membrane filters with different pore sizes, and between sub-samples that had been allowed to stand overnight at laboratory temperature in order to allow the fluorescence to develop and those which had been treated in the standard way by heating on a water-bath (see Table IV).For the sea-water sample, there was no change in the amount of aluminium determined with either treatment or filter size. For fresh water from the River Conway, both heating and passage through a filter having a larger pore diameter increased the aluminium determined by a small, but significant, amount. Similar effects of pore size of the filter on apparent aluminium concentrations have been observed by other ~ o r k e r s . 8 ~ ~ ~ ~ ~ ~ TABLE IV COMPARISON OF INFLUENCE OF REACTION CONDITIONS AND PORE SIZE OF FILTER ON THE APPARENT ALUMINIUM CONCENTRA4TION O F WATERS FROM DIFFERENT SOURCES Sample source River Conway .. Sea water (Conway Bay) Treatment Heated Stood Heated Stood Heated Stood Heated Stood Filter pore size/pm 0.45 0.45 0.1 0.1 0.45 0.45 0.1 0.1 Aluminium detectedlpg 1-1 - Sample No. 1 22.3 20.8 14.8 17.8 13.7 14.3 - 12.7 4.4 6.3 4.3 6.3 4.3 6.3 4.3 6.3 Sample No. 2 No increase in the amount of aluminium detected was observed when samples were pre- treated, by heating for up to 4 h at pH values ranging from 1.5 to 10.5, before being analysed. More severe pre-treatment was not attempted because of the levels of contamination involved. It appears from these results that the standard analytical method is sufficient to detect all of the aluminiuni present in samples in a relatively soluble form.However, because of its very low solubility, fine particulate clay material that passes through the filters is unlikely to be brought into solution under the relatively mild conditions used in the analysis. In order to clarify this aspect, analyses were made of the aluminium concentrations of clay suspensions. Analysis of Clay Suspensions Procedure Suspensions of the below 2 pm fraction of the standard clay Wyoming Montmorillonite930 HYDES AND LISS: FLUORIMETRIC DETERMINATION OF LOW Analyst, VoZ. 101 H25 (API 1949) were prepared by ultrasonic dispersion and then allowed to stand overnight before filtration and analysis. The fluorimetric analyses were carried out in the standard way directly on the prepared suspensions. Atomic-absorption spectroscopic measurements were carried out USI ng an HGA-70 heated graphite atomiser fitted to a Perkin-Elmer 303 atomic-absorption spectrophotometer.When necessary, clay suspensions were diluted in order to reduce their aluminium concentrations to less than 500 pg 1-1 so that they were within the linear portion of the calibration graph of absorption against concentration. In order to improve the reproducibility of the individual atomisations, 1% of nitric acid was added to the solutions to prevent sublimation and 2% of hydrogen peroxide was added to prevent the formation of aluminium carbide. Samples of 50 or 20 p1 were injected into the tube for each atomisation, which was carried out at the maximum temperature of the tube in the gas-stop mode. Results For the atomic-absorption spectroscopic analysis of the clay suspensions, the levels of aluminium detected agreed closely with those expected from the previous standard analysis of the clay.The temperature of the graphite tube is sufficient to break up the very stable clay structure and hence makes all of the aluminium present in these suspensions detectable. When atomic-absorption spectroscopic analysis was applied to clay suspensions that had been filtered through membrane filters of pore size 0.45 and 0.1 pm, the amount of aluminium originally present a t the 5 mg 1-1 level was reduced to 3.9 and 0.2 mg 1-1, respectively. Fluorimetric analysis detected only a small proportion of the aluminium present in the clay suspensions. The amount detected fluorimetrically in an unfiltered suspension is much smaller than the amount detected by atomic-ahsorption spectroscopy with graphite tube atomisation on a sample passed through a 0.1-pm filter (12.7 pg 1-1 compared with 200 pg 1-1 of aluminium).Filtration also causes a reduction in the amount of aluminium detected fluorimetrically (from 12.7 to 0.8 pg 1-1 of aluminium after filtration through a 0.1-pm membrane filter). For suspensions containing different amounts of clay, the amount of aluminium measured is not proportional to the concentration of clay and is also affected by the concentration of Lumogallion in the solution under analysis (see Table V). Heating the solutions under analysis for periods longer than '1.5 h does not increase the amount of alu- minium detected. Therefore, dissolution of the clxy structure itself by the analytical reagents does not appear to take place. FZuorimetric analysis.Atomic-absorption spectroscopic analysis with graphite tube atomisation. TABLE 14 FLUORIMETRIC ANALYSES OF ALUMINIUM IN CLAY SUSPENSIONS OF DIFFERENT CONCENTRATIONS Results are concentrations of aluminium detected ( p g 1-l). Lurnogallion in solution under andysis/mg 1-' Al present in ----*-, claylmg 1-' 0.5 1.0 6.0 5.5 12.7 0.5 2.0 2.4 0.1 0.6 0.5 The results in this and the previous section sbow that the fluorimetric analysis detects only aluminium present in the sample as very simple species, the analytical conditions being too mild for attack on stable structures, such as clays, to occur. However, aluminium adsorbed on the surface of particles is detected, producing the observed filtration effects.Competition between the Lurnogallion and clay surface produces the non-linear results shown in Table V. Analysis of River Waters by Fluorimetry and Atomic-absorption Spectroscopy with Graphite Tube Atomisation The results of the analyses of three different river waters (Table VI) show that the probable changes in the speciation of aluminium in the wa.ters with their pH and origin are reflectedDecembey, 1976 CONCENTRATIONS OF DISSOLVED ALUMINIUM I N NATURAL WATERS 931 TABLE VI RESLJLTS OF ANALYSIS OF RIVER WATER BY FLUORIMETRY AND ATORIIC-ABSORPTION SPECTROSCOPY WITH GRAPHITE TUBE ATOMISATION Results are concentrations of aluminium detected ( p g 1-l). Atomic-absorption Fluonmetric analysis analysis r- - 0.45-pm 0.1-pm 0.45-pm 0.1-pm Sample PH filter filter filter filter Fischbach .. . . . . 4.6 200 200 200 200 Weende . . . . . . 8.5 103 94 42 39 Leine . . . . . . . . 8.1 73 73 16.4 14.0 in changes in the relative magnitudes of the amounts of aluminium detected by the two different methods. For the atomic-absorption spectroscopic analysis with graphite tube atomisation the reproducibility of the individual atomisations is much poorer than for clay suspension or aluminium sodium sulphate solutions. This result is probably due to the com- plex speciation making the very small 5O-pl samples involved poorly representative of the bulk sample. Accuracy and Precision Accuracy Not all of the aluminium present in a given sample will be detected if that sample contains clay or similarly insoluble aluminium-containing material.The actual limit of what species are detected is uncertain so that the analysis can only be termed an analysis of “reactive aluminium,” i.e., that aluminium which will be affected by changes in the chemistry of the carrying water; this includes adsorbed aluminium and aluminium weakly bound to organic matter. Precision The standard deviation for the analysis of ten replicates of a filtered river-water sample containing 1.0 pg 1-1 of aluminium was equivalent to 0.05 pg 1-1 (coefficient of variation 5%). At a higher concentration, the standard deviation for five separate measurements over a period of 3 d was 0.6 pg 1-1 of aluminium for a 22 pg 1-I sample (coefficient of variation 2.7%). Other repeated ineasurements have usually had variations within these limits. During the period of this work, D. J. Hydes was supported financially by studentships from the Science Research Council and the University of East Anglia while at that University, and by a Royal Society Fellowship at Universitat Gottingen. Thanks are due to H. Heinrichs for his assistance with the atomic-absorption measurements. For natural waters, the accuracy of the fluorimetric analysis is uncertain. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Dougan, W. K., and Wilson, A. J., Analyst, 1974, 99, 413. Gosink. T. A., Analyt. Chem., 1975, 47, 165. Barnes, R. A., Chem. Geol., 1975, 15, 177. Kahn, H. L., Int. J . Environ. Analyt. Chem., 1973, 3, 121. Heinrichs, H., Doctorarbeit, Universitat Gattingen, 1975. Sackett, W. M., and Arrhenius, G. 0. S.. Geochim. Cosmochim. Acta, 1962, 26, 955. Nishikawa, Y., Hiraki, K., Morishige, I<., and Shigematsu, T., Japan Analyst (Bmseki KagaRu), Shigematsu, T., Nishikawa, Y., Hiraki, K., and Nagano, N., Japan Analyst (Bunseki Kagaku), Armstrong, F. A. J., Williams, P. M., and Strickland, 3. D. H., Nature, Lond., 1966, 211, 481. Burton, J. D., Leatherland, T. M., and Liss, P. S., Limnol. Oceanogr., 1970, 15, 473. Sill&, L. G., and Martell, A. E., “Stability Constants of Metal-Ion Complexes,” Spec. Publs Chem. Warner, T. B., Deep-sea Res., 1971, 18, 1255. Hem, J. D., Roberson, C. E.. Lind, C. J., and Polzer, W. L., Wut.-Supply Irrig. Pug., Wash., Wagemann, R., and Brunskill, G. J., Int. J . Environ. Analyt. Chem., 1975, 4, 75. 1967, 16, 692. 1970, 19, 551. SOC., No. 17, 1964; No. 25, 1971. NO. 1827-E, 1973. Received Febmary 16th, 1976 Accepted J w e 17th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100922

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

932 Analyst, December, 1976, Vol. 101,pp. 932-938 Improved Method for the Determination of Acrylamide Monomer in Water by Means of Gas - Liquid Chromatography with an Electron= capture Detector A. Hashimoto Kitakyushu Municipal Imtitute of Environmental Health Sciences, Tobata-ku, Kitakyushu, Japan 804 A gas - liquid chromatographic method involving electron-capture detection has been developed for the determination of an acrylamide monomer present in trace amounts in water. The method is based on bromination of the double bond by means of an ionic reaction and the 2,3-dibromopropionamide pro- duced is extracted twice with 10ml of ethyl acetate after salting out with sodium sulphate. The 2,3-dibromopropionamide from the reaction mixture is then cleaned up by using a Florisil column.The calibration graph for the method is linear over the range 0-50 pg of acrylamide monomer and the limit of detection is 0.032 p g 1-' for an aqueous standard solution. The yields of the brominated compound are 85.2 & 3.3 and 83.3 f 0.9% at fortification levels of 1.0 and 5.0 pgl-l, respectively. Recoveries of acrylamide monomer from river water, sewage effluent and sea water spiked at a level of 0.20 pg per 50ml are !39.4 f 2.5, 101.3 f 3.0 and 98.0 f 3.2y0, respectively. No interferences were observed from sea water or from 8.0% of ammonium ions present as ammonium bromide. A method involving gas - liquid chromatography with flame-ionisation detection has generally been used for the determination of acrylamide monomer in polyele~trolytes.~-~ However, this method has too low a sensitivity to be used for the determination of trace amounts of acrylamide monomer in environmental samples.A method for the determination of trace amounts of acrylamide monomer was first reported by Croll and Simkins4 and then by Arirnit~u.~ Their methods enabled the determination of 0.25 pg 1-l of acrylamide monomer to be carried out by means of gas - liquid chromatography with an electron-capture detector, the double bond being brominated in a radical reaction, i.e., by irradiation with ultraviolet light. However, these methods have several disadvantages : yields of the brominated compound are low and depend on the nature of the samples; the bromination is subjected to interference in the presence of ammonium ions; use of an ultra- violet lamp is inconvenient for treating a large number of samples; and the interferences cannot be completely eliminated with the clean-up procedure used.In this paper, a method is described for the bromination of acrylamide monomer in an ionic reaction and for the effective extraction of the 2,3-dibromopropionamide formed with a small volume of ethyl acetate. These chemical processes give increased yields. Experimental Apparatus Mass spectrometer. Model JMS-OlSG-2 combined with JEOL-980 &IS data system (Nihon Denshi Inc., Japan). Gas - liquid chromatograph with nickel-63 electron-capture detector. Shimazu, Model 5AP3FE. Column for the determination of 2,3-dibromopropionamide. A glass column (2 m x 3 mm) containing 5% of FFAP (free fatty acid polyest'er) on 60-80-mesh acid-washed Chromo- sorb W.Conditions: Nitrogen carrier gas at a flow-rate of 40 ml min-l; column temperature, 165 "C; injection temperature, 180 "C; detector temperature, 185 "C. Wash one part of Florisil (60-100 mesh) with 20 parts of boiling distilled water and dry at 110 "C for 24 h. Reactivate at 620 "C for 2 h prior to use.6 Pack 5 g of the Florisil suspended in benzene into a glass column (30 x 2 cm). Shaker. Florisil cohmn for clean-up procedure. TS Type (Irie Shokai Co. Ltd., Japan).HASHIMOTO 933 Reagents Solvents. Pure Chemical Industries Ltd.). Saturated bromine water. in a refrigerator at 4 "C. Anhydrous sodium sulphate. 600 "C overnight. Sodium thiosulphate solution, 1 M. Potassium bromide. Concentrated hydrobromic acid, sp.gr. 1.48. Acrylamide monomer. Electrophoresis-reagent grade, minimum 95% purity. Use without further purification. Dimethyl phthalate, 99.0%. Prepare a solution containing 100.0 pg ml-l of dimethyl phthalate. Stock solution of acrylamide monomer. Dissolve 105.3 mg of acrylamide monomer in dis- tilled water and dilute to 100.0ml. Standard solutions of acrylamide monomer. Dilute the stock solution so as to obtain standard solutions containing 0.1-10 pg ml-l of acrylamide monomer. Stock solution of 2,3-dibromopropionamide. Dissolve 100.0 mg of 2,3-dibromopropion- amide (see below) in methanol and dilute to 100.0 ml. Standard solutions of 2,3-dibromopropionamide. Dilute the stock solution ( a ) with methanol, for the study of extraction, and (b) with ethyl acetate, for gas - liquid chromatography, so as to obtain standard solutions containing 0.005-0.50 pg ml-l of 2,3-dibromopropionamide. Nano-grade ethyl acetate, diethyl ether, methanol, benzene and acetone (Wako Shake distilled water with bromine and allow to stand for 1 h Use anhydrous sodium sulphate that has been heated at Use the aqueous phase.Specially prepared for infrared analysis. Preparation of 2,3-Dibromopropionamide Dissolve 5 g of acrylamide monomer and 7.5 g of potassium bromide in 50 ml of distilled water in a 300-ml glass-stoppered flask. Add two or three drops of concentrated hydrobromic acid and then saturated bromine water dropwise with stirring, until a yellowish colour of bromine persists. Then add a further 5 ml of saturated bromine water and allow the mixture to stand for several hours in the dark at 0 "C.Decompose the bromine that remains with 1 M sodium thiosulphate solution. Filter off the resulting white solid with suction, dry it in vacuo and recrystallise it from benzene until a constant melting-point is obtained (132- 134 "C). The crystals should be identical with those of 2,3-dibromopropionamide, as shown by mass spectrometry. Procedure Pipette 50 ml of sample into a 100-ml glass-stoppered flask. Dissolve 7.5 g of potassium bromide, with stirring, and adjust the pH so that its value is between 1 and 3 by addition of concentrated hydrobromic acid. Wrap the flask with aluminium foil in order to exclude light. Add 2.5 ml of saturated bromine water, with stirring, and set aside the flask and contents for more than 1 h in the dark at 0 "C.When the reaction is complete, decompose the excess of bromine by adding 1 M sodium thiosulphate solution dropwise. Add 15 g of anhydrous sodium sulphate, using a magnetic stirrer to effect vigorous stirring. Transfer the resulting solution into a 150-ml separating funnel. Rinse the reaction flask three times with 1-ml portions of distilled water and transfer the rinsings into the separating funnel. Extract the aqueous solution with two 10-ml portions of ethyl acetate for 2 min, using a mechanical shaker (240 strokes min-l). Dry the organic phase with 1 g of anhydrous sodium sulphate and transfer it into a 25-ml calibrated amber-glass flask. Rinse the solid phase with three 1.5-ml portions of ethyl acetate and combine the rinsings with the organic phase.Add exactly 100 pg of dimethyl phthalate and make the solution up to the mark with ethyl acetate. Interferences Inject 5-p1 portions of this solution into the gas chromatograph. Whenever interferences were observed, the samples were cleaned up as follows. Transfer the dried extract into a Kuderna-Danish evaporator with 15ml of benzene, evaporate the solvent at 70 "C under reduced pressure and concentrate the solution to about 3 ml. Add 50 ml of benzene and subject the solution to Florid column chromatography934 HASHIMOTO: DETERMINATION OF ACRYLAMIDE IN WATER BY AnaZyst, VoZ. 101 at a flow-rate of 3 ml min-l. Elute the column first with 50 ml of diethyl ether - benzene (1 + 4) at a flow-rate of 5 ml min-l and then with 25 ml of acetone - benzene (2 + 3) at a flow-rate of 2 ml min-l.Discard all of the first eluate and the initial 9-ml portion of the second eluate, and use the remainder for the: determination with dimethyl phthalate (4 pg ml-1) as an internal standard. Preparation of Calibration Graphs Treat standard solutions of acrylamide monomer in the same manner as the samples described above. Inject 5-p1 portions of the extract into the gas chromatograph in triplicate. Calculate the mean peak-height ratio for each standard and prepare a calibration graph by plotting the peak-height ratio against the concentration (micrograms per 50 ml) : Peak height of 2,3-tlibromopropionamide (mm) Peak height of internal standard (mm) Peak-height ratio = Calculations The amount of acrylamide monomer ( A pg) in the water sample is given by where Pt is the average peak-height ratio of the test solution, Ps the average peak-height ratio of the standard solution and S p g the amount of acrylamide monomer present in the standard solution.The yield of bromination product (Byo) is given by 100 x D x P, R x -?a B = where Pa, is the average peak-height ratio of the standard solution of 2,3-dibromopropionamide, P, the average peak-height ratio of the bromination product produced on the bromination of the standard solution of acrylamide monomer, D pg the amount of 2,3-dibromopropion- amide present in the standard and R pg the calculated amount of the bromination product after the bromination of acrylamide monomer. Results and Discussion Identification of Synthesised 2,3-Dibromopropionamide The white needles of the synthesised bromination product melt at 132-134 "C.The low- and high-resolution mass spectrometric data are shown in Fig. 1 and Table I. The peaks at m/e 228.874 7,230.872 7 and 232.871 4 are identical with those of M+, (M+ + 2) and (Air+ + 4) of 2,3-dibromopropionamide, respectively. Quasi-molecular ions are also observed at m/e 229.880 9 as (M + H)+, 231.878 1 as (M + 2 $- H)+ and 233.876 5 as (M + 4 + H)+.' 70 90 110 130 150 170 190 210 230 mle Fig. 1. Low-resolution mass spectrum of the synthesised 2,3-dibromopropionamide.December, 1976 GLC WITH AN ELECTRON-CAPTURE DETECTOR TABLE I HIGH-RESOLUTION MASS SPECTRA OF PARENT PEAKS OF THE SYNTHESISED 2,3-DIBROMOPROPIONAMIDE 935 Br* = Isotope of bromine. Number of Elemental Observed M+ Error unsaturation composition 228.874 7 0.8 1 C,H,ONBr, 229.880 7 1 .o 0.5 C,H,ONBr, 230.872 7 0.8 1 C,H,ONBrBr* 231.878 1 1.5 0.5 C,H,ONBrBr* 232.871 4 1.4 1 C,H,ONBr,* 233.876 5 1.1 0.5 C,H,ONBr,* Extraction of 2,3-Dibromopropionamide The extraction of 2,3-dibromopropionamide from aqueous solution involves difficulties owing to its high polarity.Croll and Simkins4 and Arimitsua reported that ethyl acetate was an effective solvent for this extraction. However, the former did not recommend ethyl acetate because it was not possible to purify it sufficiently to achieve the desired limit of detection of less than 1 pg 1-l. Arimitsu’s ethyl acetate method needed large volumes of solvent (4 x 50 ml). The use of a salting-out agent and the adjustment of pH were examined in an attempt to reduce the volume of solvent required. Sodium sulphate (30 g per 100 ml) was the best salting-out agent.Table I1 shows that good recoveries of 2,3-dibromopropionamide spiked in 50 or 100 ml of water were obtained when the water was extracted twice with 10-ml portions of ethyl acetate at pH 1-3 with shaking for 2 min. The recoveries in the diethyl ether procedure were lower than those in the ethyl acetate procedure, even in the presence of sodium sulphate, when the water was extracted three times with 50-ml portions of diethyl ether. TABLE I1 EXTRACTION EFFICIENCIES BY THE USE OF ETHYL ACETATE AND DIETHYL ETHER AS THE EXTRACTION SOLVENTS Sodium sulphate (30 g per 100 ml) was added to each sample as a salting-out agent.Solvent Solvent volume/ml Ethyl acetate 25 15 10 10 10 10 10 10 Diethyl ether 50 50 Sample volume/ml 50 50 50 50 50 50 100 100 100 100 2,3-DBPA* spiked/pg 0.50 0.50 0.50 0.50 0.50 0.50 1.00 1.00 1.00 1 .oo No. of extractions 1 1 1 2 2 2 1 2 3 3 Extraction 3.0 3.0 2.0 1.0 2.0 3.0 1.0 1.0 1.0 3.0 PH Recovery, 95.0 90.0 90.0 %t 100 100 100 79.0 97.5 75.0 86.0 * 2.3-Dibromopropionamide. t Means of at least two determinations. Select ion of Gas - Liquid Chromatographic Column The column reported by Croll and Simkin~,~ containing 10% of FFAP on acid-washed dimethyldichlorosilane-treated Chromosorb W did not permit the separation of the peaks of impurities in potassium bromide and of 2,3-dibromopropionamide. Other columns, such as 25% PEG 2011 or 20% PEG 6000, gave similar results.It was eventually found that a 5% FFAP column was useful for separating the peak of 2,3-dibromopropionamide and those of the impurities in potassium bromide under the given chromatographic conditions. A typical gas - liquid chromatogram is shown in Fig. 2. The retention times for 2,3-dibromopropionamide and dimethyl phthalate are 4.0 and 8.5 min, respectively.936 HASHIMOTO : DETERMINATION OF ACRYLAMIDE IN WATER BY Analyst, VoZ. 101 Clean-up Method with Florisil Column The eluate was collected in 1-ml fractions and a typical elution pattern is shown in Fig. 3. The impurities from potassium bromide and phthalates in the sample were eluted with diethyl ether - benzene (1 + 4). The recovery of 2,3-dibromopropionamide was more than 95% under standard conditions.Y I I I I I I I I J 0 2 4 6 8 10 12 14 16 Time/min F i g 2. Typical gas chromatograms of the bromination product obtained from aqueous acryl- amide monomer solution: A, untreated; B, with Florisil clean-up ; and BL, chromatogram of blank concentrated five-fold before gas-chromatographic analysis. Peaks : 1, 2,3-dibromopropionamide ; 2, dimethyl phthalate ; and 4-7, impurities from potassium bromide. Sample size, 100 ml; acryl- amide monomer, 0.1 pg. Volume/m I Fig. 3. Typical chromato- graphic elution pattern from a Florisil column. Effect of Potassium Bromide and Hydrobromic -cid on Bromination As the rate of bromination of acrylamide monoiner in water with bromine water is very low, presumably owing to the electron-attractive effect of its carboxyl functional group, a catalyst, e.g., bromide is needed.s In this study, potassium bromide and hydrobromic acid were used.The recovery of the bromination product as a function of the amounts of potassium bromide and hydrobromic acid is shown in Fig. 4. Potassium bromide (15-20 g per 100 ml) gave good results and therefore 15g of potassium bromide per 100 ml were used throughout the experiments. Ammonium bromide has recently been found to be a useful catalyst of the bromination reaction and ammonium ions do not seem to interf,ere with the bromination by the proposed method, provided that the ions are present as bromide. These studies will be discussed from the analytical point of view in a separate paper. The yield of 2,3-dibromopropionamide was constant in the presence of more than 3 ml of saturated bromine water per 100 ml and therefore 5 ml per 100 ml were used throughout the experiments.Effect of Reaction Time on Bromination 0 to 24 h. The yield was constant when the reaction time was more than 1 h. Fig. 5 shows the recovery of the bromination product as a function of reaction time from Effect of Initial pH on Bromination Fig. 6 shows the recovery of the bromination product as a function of the initial pH fromDecember, 1976 GLC WITH AN ELECTRON-CAPTURE DETECTOR loo; Amount of K8r/g per 50 ml I I I I I I 0 2 4 6 8 1 0 Amount of HBr/mI per 50 ml Fig. 4. Effect of (A) potassium bromide and (B) hydrobromic acid on the yield of bromination. Sample size, 50 ml; acrylamide monomer, 0.25 pg. 100 - 1 - /I 8 s s 50 U 4 1 I l l Tirne/h 0 1 2 3 4 / 24 Fig.5. Effect of reaction time on the bromination. Reaction conditions : sample, 50 ml ; acrylamide monomer, 0.25 pg; potassium bromide, 7.5 g; and saturated bromine water, 2.5 ml. Extraction conditions : sodium sul- phate, 15 g ; extraction pH, 2; solvent, ethyl acetate, 10 ml ( x 2). 937 1 to 7.35. The yield was constant within this pH range. The use of conventional buffer solutions, such as sodium acetate - acetic acid solution or phosphate solution, was disadvan- tageous as a significant decrease in the yield occurred. Calibration Graph The calibration graph was prepared by plotting the peak-height ratio against the con- centration [micrograms per 50 ml), and was linear over the range from 0.05 t o 0.25 pg of acrylamide monomer per 50 ml.As the peak-height ratio changes with change in the gas - liquid chromatographic conditions, a calibration graph should be prepared daily. Sensitivity and Precision The limit of detection, expressed as the product of multiplying 3 4 2 by the standard deviation of the blank value, was equivalent t o 0.032 pg 1-l. The blank value of this method s 9 50 E 2 0 \ \ \ 't \ \ \ \ \ 1 1 I I I r r l 0 1 2 3 4 5 6 7 8 PH Fig. 6. Effect of the initial pH on bromina- tion. Reaction and extraction conditions as in Fig. 5. The pH was adjusted t o below 3 and t o 4-5 with concentrated hydrobromic acid and dilute hydrobromic acid, respec- tively. Reaction at pH 6 was in distilled water. pH 7.35 was achieved by careful addition of dilute sodium hydroxide solution.The broken line shows the result obtained by the use of sodium acetate - acetic acid buffer solution.938 HASHIMOT0 was found to be equivalent to 0.035 r f i 0.007 5 pg 1-1 (five determinations) under standard conditions. The coefficient of variation calculated from a series of five replicate determina- tions of 0.2Opgl-1 of acrylamide monomer (in the form of an aqueous standard solution) was 4.2%. Analysis of Water Samples Spiked with Acrylamide Monomer The method was applied to the analysis of river water, secondary treated sewage effluent and sea water, which were spiked with 0.20 pg of acrylamide monomer per 50 ml. The values for the three different samples are given in Table 1I:I. The proposed procedure can be applied successfully to natural sea water. TABLE 1111 RECOVERY OF ACRYLAMIDE MONOMER FROM WATER SAMPLES AS 2,3-DIBROMOPROPIONAMIDE Number of samples = 5; sample volume = 50 ml.Over-all Acrylamide Amount of 2,3-DBPA* / p g Bromination recovery of Coefficient monomer +-- recovery, acrylamide of variation, Sample spikedlpg Calculated Found.1 %t monomer, %t yo 3.3 Standard 0.05 0.162 0.138 85.2 - 1.0 solution 0.20 0.649 0.535 82.4 - 0.9 0.25 0.812 0.677 83.3 - River water 0.20 0.649 0.531 81.8 99.4 2.5 Sewage effluent 0.20 0.649 0.542 83.5 101.3 3.0 Sea water 0.20 0.649 0.524 80.7 98.8 3.6 * 2,3-Dibromopropionamide. t Means of five replicate determinations. Conclusions Trace amounts of acrylamide monomer in water were brominated by an ionic mechanism and determined by gas - liquid chromatography with an electron-capture detector.The proposed bromination procedure gave higher yields of 2,3-dibromopropionamide (about 80%), with less variation in the yields, than the ultraviolet irradiation method. This bromination method did not depend on the nature of the samples, and was successfully applied to sea water samples, which can be analysed by the ultraviolet irradiation method only with difficulty. No interference was observed in the presence of S.OyO of ammonium ions. The author thanks Dr. T. Akiyama, Research Director of the Kitakyoshu Municipal Institute of Environmental Health Sciences, and Mr. M. Suzuki for their valuable advice and encouragement, and Mr. S. R. Shinohara and Mr. M. Koga for their co-operation in the mass spectrometry. He also thanks Professor K. Saito, Department of Chemistry, Tohoku University, for his kind assistance in the preparation of this manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. References MacWilliams, D. C., Kaufman, D. C., and Waling, B. F., Analyt. Chem., 1965, 37, 1546. Water Research Association Technical Inquiry Report, No. 171, 1968. Croll, B. T., Analyst, 1971, 96, 67. Croll, B. T., and Simkins, G. M., Analyst, 1972, 97, 281. Arimitsu, H., Suido Kyokai Zasshi, 1974, No. 473, 10. Mills, P. A., J . Ass. 08. Analyt. Chern., 1968, 51, 29. McFadden, W. H., “Te2finiques of Combined Gas Chromatography/Mass Spectrometry : Applications Arimitsu, H., Suido Kyokai Zasshi, 1975, NO. 487, 31. Conners, P. A., “Reaction Mechanisms in Organic Analytical Chemistry.” Wiley-Interscience, New Received April 28th, 1976 Accepted July 27th, 1976 in Organic Analysis, Wiley-Interscience, New York, 1973, p. 33. York, 1973, p. 346.

ISSN:0003-2654    

DOI:10.1039/AN9760100932

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, December, 1976, Vol. 101, pp. 939-942 Elimination of Ionic Interferences in the 939 Determination of Sulphates in Water Using the Lead-sensitive Ion-selective Electrode Adam Hulanicki, Ryszard Lewandowski and Andrzej Lewenstam Institute of Fundamental Problems in Chemistry, University of Warsaw, 02-093 Warsaw, Poland In the titration of sulphate ions with lead(I1) perchlorate solution using a lead-sensitive electrode for end-point determination, the addition of 75% V/V of methanol permits titration of M samples. Under these conditions, a 200-fold excess of nitrate does not interfere, but chloride a t this concentration introduces a small systematic error of 4%. Significant errors are caused by the presence of calcium, which co-precipitates as calcium sulphate with lead sulphate.In titrations carried out rapidly, the negative error increases with the amount of calcium and the true equilibrium may be obtained after a relatively long time. Increasing the ionic strength of the solution by addition of sodium perchlorate eliminates the interference by calcium, but for concentrations of sulphate below M it is advantageous to remove calcium by using a cation-exchange resin in the hydrogen form. The method has been applied successfully to the determination of sulphate in water. Methods for the determination of sulphates in waters, in the range from 4 x M, based on precipitation or complexometric titration, cannot be used below concentrations of M because of the solubility of lead or barium sulphates, co-precipitation of interfering ions and diffuse colour changes of indicators.Solid-state lead electrodes can be used for detection of the end-point in titrations with Such electrodes are sensitive to sulphate ions but, unless the solution is buffered with lead sulphate, the potential is not reproducible and direct potentiometric measurement of sulphate concentration is not possible. In this work, the interference of some ions and the effect of an indifferent electrolyte on sulphate determination has been investigated. Apparatus and Reagents module was connected with an ABU 13 Autoburette. chalcogenide membranes were used as sensors. calomel electrode with a 1 M sodium nitrate electrolyte bridge. added at a rate of 0.062 5 ml min-l. The titrations were carried out in glass vessels. distilled water.Effect of Solvent Composition The solubility of lead sulphate in water (pK,, = 7.8) limits the titration procedure to concentrations of sulphate of not less than M. Solvents of low dielectric constant such as 1,4-dio~an,~s~y~ methanol3 or propan-2-olg decrease the solubility of lead sulphate and permit titrations of smaller concentrations. 1,4-Dioxan is a relatively unstable reagent and its decomposition products have an adverse effect on the lead electrode. Methanol, which is readily available and cheap, was used; no effect on the membrane or on the electrode body has been reported with this solvent. The solubility product of lead sulphate in 75% V/V methanol is approximately 1O-l1, which allows the titration of samples containing 25 pg of sulphate a t a concentration of about 5 x 1 0 - 5 ~ .The concentration of sulphate in the titrated solution is reduced 4-fold by dilution with the added solvent and the total potential change between the beginning of titration and lOOyo excess of titrant is 60 mV (Fig. 1). The precipitation of lead sulphate is to 4 x or barium'98 salts. Experimental and Results A Radiometer pHM 64 digital pH meter with a Servograph REC-61 recorder and REA 100 The lead-sensitive electrode was an Orion 94-82A and home-made electrodes with lead The reference electrode was a saturated The solutions were stirred with a magnetic stirrer at 325 rev min-l and the titrant was All reagents were of analytical-reagent grade and the solutions were prepared in triply940 HULANICKI et al. : ELIMINATION OF INTERFERENCES IN SULPHATE Analyst, VoZ.101 slower in the low-polarity medium1* and, in addition, co-precipitation and super-saturation effects increase. The possible use of sulphate titrations for the analysis of waters and other solutions depends on the presence and concentration of interfering ions, which can react with the titrant or with sulphate or increase the total ionic strength of the solution, thus increasing the solubility of lead sulphate. Effect of Anions M, nitrate up to 5 x M. If phosphate is above this level, it must be removeds by addition of lanthanum nitrate.6 The effect of hydrogen carbonate and other ions of weak acids can be eliminated easily by acidifying the sample to pH 4. The effect of chloride and nitrate is significant in 1,kdioxan - water s o l ~ t i o n s , ~ ~ * ~ ~ but for 10-3 M sulphate in 75% methanol a 200-fold excess of nitrate has no influence on the end-point of the titration.In order to eliminate the chloride error, the best procedure is to pass the sample through a cation-exchange resin in the silver form and then through one in the hydrogen Effect of Calcium In addition to ions that have large selectivity coefficients such as Ag+, Hg2+, Cu2+ and Fe3+, which are present in only very small concentrations, calcium, usually present in waters at levels above M, can cause a negative error in sulphate determination under some con- ditions. M sulphate in 75% methanol at pH 4.0 are shifted towards more positive potentials in the presence of increasing amounts of calcium, the total potential change is smaller and the end-point appears too early (Fig.2). The end-point cannot be determined in the presence of a 20-fold excess of calcium. When the calcium is present in a 4-fold excess, the concentration product [SO,][Ca] = 6.25 x 10-8, but no The concentration of anions in surface waters is variable: chloride ranges up to M and phosphate up to 2 ><: The titration graphs of 1.25 x 1 2 Equivalents of titrant Fig. 1. Titration graph of 1.25 x M sulphate in 75% V / V methanol a t pH 4.0. Titrant, 1 x M lead perchlorate solution. 1 - 1 2 Equivalents of titrant Fig. 2. Effect of cal- cium on determination of 1.25 x W 4 M sulphate in 75% V / V methanol. Titrant, 1 x 1 0 - 2 ~ lead perchlorate. Calcium t o sulphate ratio: A, 0; B, 1; C , 2; D, 4; E, 8 ; F, 10; and G, 20.December, 1976 DETERMINATION USING AN ION-SELECTIVE ELECTRODE 941 visible precipitate of calcium sulphate occurs.However, the early inflection may suggest that part of the free sulphate has been removed from the solution, possibly by co-precipitation of calcium sulphate together with lead sulphate, the reaction being more significant the greater the excess of calcium. A comparison of solubility data of both sulphates as well as the dissociation of calcium sulphatell suggests that the equilibrium should finally be shifted towards the formation of lead sulphate. This is confirmed experimentally by a slow drift of the potential before the equivalence point is reached. When the potential is allowed to reach equilibrium after addition of each aliquot of the titrant (for M sulphate this equilibration can take up to 20 min) the calcium sulphate is transformed into lead sulphate and the titration graph approaches that obtained in the absence of calcium.When the titrant is added continuously, equilibrium is not established and the excess of lead ions in solution is responsible for the distortion of the titration graph (Fig. 3). These effects can be suppressed by addition of an indifferent electrolyte, i.e., by increasing the ionic strength or by decreasing the amount of organic solvent. The latter is disad- vantageous, however, because the potential end-point break is decreased significantly. In these investigations , much better results were obtained by addition of sodium perchlorate to increase the ionic strength.In spite of the smaller change in potential during the titration (Fig. 4), the addition of 0.05 M sodium perchlorate eliminated interference from a 4-fold excess of calcium, and 0.1 M sodium perchlorate that from a 20-fold excess of calcium at the sulphate level of M. - 1 2 Equivalents of titrant Fig. 3. Effect of titra- tion rate on the titration graph of 1.25 X M sulphate. Calcium to sul- phate ratio: 4. A, Con- tinuous titration a t the rate of 0.0625 ml min-l; B, the addition of titrant was stopped a t points 1, 2, 3 and 4 to allow establishment of a steady potential. 1 2 Equivalents of titrant Fig. 4. Titration graphs of 1.25 x 10-4 M sul- phate in the presence of 0.1 M sodium perchlorate. Titrant, 1 x 1 0 - a ~ lead perchlorate solution.Cal- cium to sulphate ratio : A, 0 ; B, 2; C, 4; D, 8; E, 10; and F, 20. When the expected sulphate concentration is smaller than M and the calcium excess Fig. 5 is greater, it is advisable to pass the solution through the cation-exchange resin.942 HULANICKI, LEWANDOWSKI AND LEWENSTAM shows titration graphs for tap-water samples containing 60.40 mg 1-1 of SO,% without any pre-treatment, after passing through Amberlite HP 120 in the hydrogen form and in the presence of sodium perchlorate. In the latter two instances, the precision of determination is within &2%, and the results agree well with tholse obtained by gravimetric determination of sulphate as barium sulphate.12 0 0.5 1 .O 1.5 2.0 2.5 Amount of 0.01 Iv1 lead (I I) perchloratdml Fig. 5. Titration of a 200-ml sample of tap water containing 60.40mgl-1 of sulphate.A, Sample end-point a t 0.746 ml equivalent to 35.8 mg 1-’ of sulphate; B, sample after cation exchange; end-point a t 1.27 ml equivalent to 60.9 mg 1-1 of sulphate ; C, sample in the presence of 0.1 M sodium perchlorate solution; end-point a t 1.255 ml equivalent to 60.2 mg 1-1 of sulphate. In all titrations: solvent, 75% V / V methanol; pH, 4.0; titrant, 1 x lo-* M lead perchlorate solution. Procedure To a 100-ml beaker add 10 ml of the water sample and 2 ml of 1 M sodium perchlorate. Stir the solution and adjust the pH to 4.04.4 using dilute perchloric acid or sodium hydroxide. Add 30 ml of methanol and introduce the lead-sensitive and reference electrodes incorporating a sodium nitrate electrolyte bridge.Titrate with a 0.01 M solution of lead(I1) perchlorate in 75% V/V methanol a t a rate not greater than 0.1 ml min-l, recording the titration graph. The end-point can be located from the graph. When the sulphate concentration is below 1 0 - 4 ~ , pass the solution through a cation- exchange resin in the hydrogen form or add 1-2 g of resin directly to the solution and, after stirring for 15 min, separate it on a sintered-glass filter. Titrate the sample as described above but without addition of sodium perchlorate. The authors are indebted to the Institute of Meteorology and Water Economics for financial support in the course of this study. References 1 . 2. 3. 4. 5 . 6. 7 . 8. 9 . 10. 1 1 . 12. Ross, J. W., Jr., and Frant, M. S., Analyt. Chewr., 1969, 41, 967. Selig, W., Mikrochim. Acta, 1970, 168. Goertzen, J. O., and Oster, J. D., Proc. Soil Sci. SOC. A m . , 1972, 36, 619. “Titration of Sulphate Ion with the Lead Electrode,” Orion Research Inc., Applications Bulletin Mascini, M., Analyst, 1973, 98, 325. “Analytical Methods Guide,” Seventh Edition, Orion Research Inc., Cambridge, Mass., 1975. Rechnitz, G. A., Lin, 2. F., and Zamochnik, S. EL, Analyt. Lett., 1967, 1, 29. Hartzdorf, C., 2. Analyt. Chem., 1972, 262, 167. Robins, C . W., Carter, D. L., and James, D. W.), Proc. Soil Sci. SOC. Am., 1973, 37, 212. Ashbrook, A. W., and Ritcey, G. M., Analyst, 1!361, 87, 740. Nakayama, F. S., and Rasnick, B. A., Analyt. Chem., 1967, 39, 1022. Belcher, R., and Nutten, A. J ., “Quantitative Inorganic Analysis,” Butterworths, London, 1960. No. 11, Cambridge, Mass. Received April 27th. 1976 Accepted June 24th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100939

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, December, 1976, Vol. 101, pp. 943-948 Automatic Apparatus for the Determination of pH 943 and Nitrate in Soils D. Goodman National Vegetable Research Station, Wellesbourne, Warwick, C V35 9EF An apparatus is described that automatically extracts and analyses batches of up to 60 soil samples. Analysis is performed electrochemically by means of either a pH or ion-selective and reference electrodes. The electrode output is amplified and recorded on a chart recorder, thereby allowing the electrode performance t o be monitored. Results of tests with both pH and nitrate ion-selective electrodes are reported. Good agreement was obtained between automatic and manual methods. Difficulty was experienced with the operation of an Orion nitrate electrode but tests using a Corning electrode on the apparatus gave excellent recoveries and reproducibility. Agricultural analysts spend much of their time measuring nutrient levels in soil samples in order that advice may be given about fertiliser practices.The introduction of ion-selective electrodes for the measurement of nutrient ions in soil suspensions promises to speed this process, but their full potential cannot be realised until an automatic machine for using them has been devised. The apparatus described here was designed for this purpose. Unlike a similar apparatus designed by Baker1 for the automatic measurement of soil pH, it uses only one measuring electrode, with or without an associated reference electrode, and can be fitted with an ion-selective or pH electrode. It also differs from Baker’s design firstly in being a “batch” process, where a batch may be any number of samples between 6 and 60, and secondly in stirring the soil suspension with motor-driven glass paddles, a system capable of breaking the lumps of fresh soil that are necessarily used when nitrate is to be measured.Apparatus A plan of the general arrangement is shown in Fig. 1. It consists of a rail-mounted carriage, Instrument table Pressurised over carriage electrode wash Trolley track water bottles traverse motors Control box s t 7 b o t t l e 1 , C o m ? r e s s ; A , , , I Water dispenser recorder head sample beakers Fig. 1. Layout of apparatus. Arrows show direction of movement of carriage (A) and water and electrode trollies (B) and (C), respectively. (Not to scale.)944 GOODMAN : AUTOMATIC APPARATUS FOR THE Analyst, Vol.101 which carries rows of sample beakers past three ‘stations” where each sample receives an aliquot of extractant, usually water, is thoroughly stirred and has an electrode or electrodes lowered into it. The electrical output from the e:lectrode(s) is passed through an amplifier to a 25-cm flat-bed recorder. Control of the sequence of operations is completely automatic involving a system of three interlocking motor-driven cam timers, thereby ensuring that each sample receives identical treatment. - Construction The apparatus is built around a rigid main frame constructed from Dexion square-section tubular steel, 200 cm long x 38 cm wide x 13 cm high. Attached to this frame and running longitudinally is a set of guide rails supporting a travelling carriage that carries the samples.This carriage bears two removable aluminium trays, each holding thirty 50-ml sample beakers in rows of five. Drive to the carriage is effected by two continuous cords, which pass over pairs of pulleys at both ends of the apparatus, one set of pulleys being driven by a geared induction motor. Positive indexing of the carriage is achieved by a longitudinal cam fixed to the side of the carriage, which engages microswitches attached to the main frame. At the first station, a measured volume of water (usually 25ml) is dispensed into each beaker in turn by a dispenser head mounted on a screw-driven trolley. This trolley steps across the carriage above the row of beakers in direction B (Fig.1). Water is delivered to the head by a syringe-type water dispenser actuated by a double-acting pneumatic cylinder. Compressed air for this cylinder and other operations on the apparatus is drawn from a small diaphragm compressor supplying air at 69 x lo3 N m-2 (10 lbf in-2). At the second station, the row of beakers is positioned beneath a movable beam carrying a set of five glass stirring paddles. The beam is lowered by motor-driven lea-d screws so that the paddles thoroughly stir the contents of the beakers. The paddle speed can be pre-set and is constant irrespective of the load, control of the 6-V d.c. driving motor being effected by a small electronic circuit.2 At the measuring station, a screw-driven trolley, similar to that used for the water &s- penser, is stepped across the carriage above the row of beakers in direction C, Fig.1. This trolley carries two identical pairs of parallel, ve:rtically swinging arms, the outer ends of which support a vertical plate that carries the electrodes, a miniature 2-V electric stirrer motor and paddle and the pipes supplying wash water to the electrodes. A double-acting pneumatic cylinder mounted on the trolley lowers the plate so that the electrodes and stirrer paddle dip into the extract, raising it again when measurement is complete. pH or ion-selective electrodes are connected directly to a Pye Unicam, Model 291, pH meter, the recorder output from which is in the form of a varying current. Recorder scale-expansion control is obtained by tapping off the voltage drop as this current is passed down a 10-turn helical potentiometer and using it to drive a Linseis 25-cm flat-bed potentiometric recorder.The main controls and indicator lamps are mounted on the sloping front panel of a free- standing control box (Fig. 1). Inside, a top colmpartment contains the switching relays while the bottom compartment houses the d.c. power supply and three constant-cycle sequence timers (cam timers) used for timing the sample carriage movement, water application and the electrode sequences. Connection to the apparatus is effected by means of a 56-way multi-pole connector and cable. The principal electrical circuits are shown diagrammatically in Fig. 2. All major circuits are operated at 240 V, 50 Hz a.c., so that microswitches can be used for controlling the inductive loads of the eight asynchronous electric motors.Operation The carriage is loaded with the necessary beakers, samples and calibrating solutions before the start buttons are depressed. All subsequent operations are under the control of a system of three electrically interlocked cam timers (Fig. 2). Over-all control is carried out by the main cam timer, which steps the sample carriage forward one row of beakers every 15 min; thus a full batch of 12 rows of five beakers takes 4 h to be processed. The sequence of operations is as follows: (a) Movement of the sample carriage brings a row of beakers to the first station and starts the water cam timer. Water is delivered into the first beaker of the row and the dispenser re-fills. The dispenser trolley then moves in direction B (Fig.1) untilDecember, 1976 DETERMINATION OF PH AND NITRATE I N SOILS 945 it is over the next beaker. The cycle of delivery, re-fill and traverse is repeated until each beaker in the row has been dealt with. The water cam timer then stops and the trolley returns to its starting position to await the end of the 15-min cycle. After the sample carriage has stepped forward one row, the main cam timer signals for the stirrer paddles to be lowered into position. Stirring proceeds until nearly the end of the 15-min cycle, when the paddles are again raised clear of the beakers. After being stirred, the extracts are allowed 15 min for the coarser suspended matter to settle out before moving to (d). The arrival of a row of beakers at this station starts the electrode cam timer’s 2-min cycle.The electrodes are immediately lowered into the extract, which is then gently stirred for 9 s to hasten equilibration; 20 s later recording of the pH meter output commences. After a further 70 s, the recorder switches out, its input terminals being short-circuited to return the recorder pen to zero. The electrodes are raised and washed, the washings falling into the beaker just measured. The electrode trolley then steps on to the next beaker in direction C (Fig. 1) and the cycle is repeated. On completion of the row, the electrode cam timer stops and the trolley returns to its starting position to await completion of the 15-min cycle. When the whole batch has been processed, the direction of the sample carriage reverses and it returns to the starting position.Here it trips a limit switch, which disconnects the power supply and stops the apparatus. I I I Start i I cu t-ou t - Main -. Carriage repeat *I----- Mains - supply - end cam timer 4 step ------ Safety interlock circuits I I I I I I I $q Measuring and reference electrodes Amp1 if ier (pH meter) I ~’ Chart recorder Fig. 2. Block diagram of main electrical circuits. For processing batches of less than 60 beakers, a stop-pin is inserted after the last beaker. This pin trips microswitches on the water and electrode trolleys, returning them to their starting positions and reversing the sample carriage after the end of the current 15-min cycle. A switch on the control console allows a complete run to be repeated with or without the addition of a further aliquot of water.Press buttons are provided for re-starting the apparatus manually after a break in the main power supply. Safety Interlocks electrode trolley traverse signal. If a fault causes the electrodes to fail to rise clear of the beakers, a microswitch blocks the Similarly, if either water or electrode trolleys have not946 GOODMAN : AUTOMATIC APPARATUS FOR THE Analyst, Vol. 101 returned to their starting positions by the time the carriage step is signalled, a safety interlock prevents further movement of the carriage, holds the main cam timer and illuminates a warning light on the control console until the water and electrode sequences have been completed. Ion-selective Electrodes An Orion 92-07 nitrate ion-selective electrode operating through a Pye Universal pH meter was used in the early work.It was fitted with a fine cross-wire over the tip. This wire trapped a droplet of liquid when the electrode was raised, which it retained on re-entry, thereby preventing air bubbles from collecting and covering the sensing membrane. All later work was carried out with a Corning liquid-junction nitrate ion-selective electrode operating through a Pye, Model 291, pH meter. This electrode has a flat end incorporating the sensing membrane and gave no trouble from trapped air bubbles. A Philips R44/2/-SD/1 double-junction reference electrode, containing 0.02 M potassium chloride solution in the outer chamber, was used throughout these experiments. Sample Preparation For pH determination, the soil samples may bt: either fresh or air dried, but the deter- mination of soil nitrate should be performed on the fresh soil as soon as possible after sampling.All samples should be lightly crushed to pass a 2-mm round-hole sieve and any stones should be removed before mixing and removing a portion for analysis. Performance Data PH Using a standard combined pH electrode, the pH values of a range of 12 soils were measured on the apparatus and manually. Three replicate rneasurements were made by each method, the soils being arranged in a different random order for each set of measurements. The same amount of soil (10 g) and volume of water (251 ml) and the same electrode and pH meter were used throughout. The results are given in Table I and show that the standard deviation for an individual measurement determined with the apparatus was) 0.044 (10 degrees of freedom) compared with 0.039 pH unit for those carried out by the standard manual method.The two sets of results were highly correlated (Y = 0.999); a regression of the apparatus on the manual results gave a slope of 0.966, which was not significantly different from unity, and an intercept of 0.226, which did not significantly differ from zero. Further statistical analysis of these results failed to reveal any significant carry-over of solution from one beaker to the next. TABLE I DETERMINATION OF SOIL pH MANUALLY AND WITH THE APPARATUS Manually With the apparatus Sample 1 2 3 4 5 6 7 8 9 10 11 12 i 5.51 5.81 6.00 6.16 6.41 6.54 6.70 7.09 7.10 7.32 7.98 8.06 2 3 5.43 5.51 5.80 5.71 6.00 6.97 6.12 6.18 6.34 6.40 6.53 6.62 6.67 6.68 7.09 7.04 7.09 7.06 7.31 7.30 7.83 7.93 8.03 8.02 -7 Mean 5.48 5.77 5.99 6.15 6.38 6.56 6.68 7.07 7.08 7.31 7.91 8.03 7 1 6.47 5.78 5.93 6.20 6.43 6.60 6.75 7.00 7.12 7.32 7.79 7.91 2 3 5.49 5.49 5.76 5.76 6.09 5.99 6.23 6.13 6.39 6.36 6.65 6.59 6.72 6.68 7.06 6.98 7.20 7.07 7.37 7.26 7.88 7.83 7.95 7.90 7 Mean 5.48 5.77 6.00 6.19 6.39 6.61 6.72 7.01 7.13 7.32 7.87 7.92 Nitrate An Orion electrode was used with the apparatus to analyse 17 widely different types of soil.Nitrogen concentrations in the extracts were between 1 and 21 pg ml-1. DuplicateDecem bey, 19 76 DETERMINATION OF PH AND NITRATE I N SOILS 947 extracts of these soils were prepared by shaking with water at the same soil to water ratio as that used on the apparatus (1 g to 2.5 ml) and, after filtration, were analysed by steam distillation with alkaline titanium(II1) sulphate.3 Comparison of the results obtained by the two methods gave a correlation coefficient of 0.969 (49 degrees of freedom), the electrode results being 88% of those obtained by shaking and distillation.Under these conditions, the electrode response was linear down to 3 pg ml-1 of nitrogen in the extracts, but fell off sharply below this level. In a series of experiments to test the Corning electrode with the apparatus, a range of standard nitrate solutions were added to weighed samples of a sandy loam soil. The nitrate contents of these modified samples were then determined by three different methods: (a) extracting 20 g of soil with 50 ml of water, filtering and analysing the filtrate by Kjeldahl distillation with alkaline titanium(II1) sulphate3; (b) preparing the filtrate as in (a) and deter- mining the nitrate concentration in the extract by manually inserting the Corning electrode; and (c) extracting 10 g of soil with 25 ml of water on the apparatus and determining the nitrate content of the extract by automatic insertion of the Corning electrode into the soil suspension.The results are given in Table 11. The slopes of the regression lines show that the electrode recorded 94-96y0 of the added nitrate, the response of the electrode in soil suspension being substantially linear down to 2.5pgml-l of nitrate in the extract. TABLE I1 RECOVERY OF NITRATE-NITROGEN FROM SOIL BY THREE TECHNIQUES NO,-N addedlpg ml-l 0 10 20 30 40 50 60 70 80 90 100 Regression of found (Y) on added (X) nitrate .. . . Standard error of slope . . Standard error of intercept . . Correlation coefficient (degrees of freedom in parentheses) By shaking, filtration and distillation/pg ml-1 f C 7 1 2 3 Mean 0.7 0.4 0.9 0.7 10.0 10.9 9.7 10.2 20.8 19.8 20.7 20.4 30.7 29.9 29.7 30.1 41.3 40.4 40.4 40.7 49.0 50.7 50.0 49.9 60.9 60.0 61.5 60.8 71.3 70.8 71.0 71.0 79.1 81.9 77.2 79.4 90.4 91.8 93.3 91.8 96.6 100.3 99.2 98.7 Y = 0.996 OX + 0.544 0.006 9 0.410 0,999 2(31) By shaking, filtration and manually operated Coming electrode/wg ml-' A f > 1 2 3 Mean 1.8 1.7 1.7 1.7 By Corning electrode on automatic apparatuslpg ml-l L f > 2 3 Mean 1.5 1.4 1.4 1.4 ~~ 11.7 13.7 11.7 12.4 i i ; 7 12.2 12.2 12.0 21.0 20.8 22.0 21.3 21.0 21.6 21.7 21.4 32.5 30.5 32.0 31.7 31.3 31.2 31.6 31.3 40.8 38.8 42.0 40.5 41.0 40.8 40.0 40.6 51.5 45.8 53.0 50.1 61.0 58.0 60.5 59.8 70.0 67.5 67.5 68.3 49.5 49.8 50.8 50.0 58.5 59.1 60.5 59.4 68.0 67.8 68.5 68.1 79.0 77.0 83.0 79.7 77.5 79.0 77.0 77.8 90.0 85.0 88.5 87.8 88.0 86.0 85.0 86.3 98.0 94.0 99.8 97.3 95.0 95.5 98.0 96.2 Y = 0.952 2X + 2.447 0.010 3 0.608 0.998 2(31) Y = 0.938 5X + 2.589 0.005 0 0.297 0.999 6(31) In order to simplify the regression analysis, the random error was assumed to be inde- pendent of the magnitude of the measurements.This approximation had little influence on the line of best fit because the errors are very small. However, the errors do tend to increase with the size of the measurement and an estimate of the standard error of a single measurement based on the results obtained automatically with the Corning electrode, and quoted in Table 11, was 1.9%. Discussion The excellent performance obtained with the Corning electrode mounted on the apparatus is well illustrated by the results given in Table 11.These results show that there was good agreement between the three methods, although the strictly controlled conditions of automatic operation gave significantly more reliable results than either distillation or manual methods. The ceramic membrane of the electrode performed well in soil suspension and, being particu- larly free from interference by clay particles, it responded rapidly to changes in nitrate concentration.A logarithm of concentration parameter was used in the calculation of the regression equations for the Corning electrode instead of the more correct logarithm of ionic activity, which probably accounts for the calculated recovery being less than the theoretical maxi- mum and also for the calculated endogenous nitrate (Le., the intercept of the regression equation) being higher than that obtained by direct measurement. However, the error is948 GOODMAN small and the use of this method avoids the calculaiion of ionic activity in such a complex medium as a soil suspension. The above results contrast with the relatively poor results obtained when using the Orion electrode. Air bubbles frequently became trapped under the sensing membrane in spite of the fitting of the cross-wires, a fault which might not have occurred if the electrode had been made to enter the liquid a t an angle of 20” as the makers recommend.It was also found that suspensions of sandy loam could clog the fine pores of the sensing membrane, causing a rapid loss in sensitivity and sluggish performance during the measurement of a full batch of 60 soils and standards. Conclusions Since it was built, the apparatus described has been used to analyse more than 4 000 soil samples for either pH or nitrate and has proved to be very reliable in operation. Because of the inherent simplicity of the circuitry, minor inodifications to the original design have been carried out without difficulty and component and power failures have caused very few problems. With the first design of the apparatus the results were recorded digitally but subsequent experience showed the necessity for having some additional check on the electrode performance. It has been found that the recorder trace now obtained, while losing something in convenience, provides an essential check should the electrode response become sluggish and fail to equilibrate in the time allowed. Although the tests recorded above refer only to the determination of pH and nitrate in soils, there appears to be no reason why the apparatus should not also be used for a wide range of other electrochemical measurements, for example the conductivity of water samples or the chloride content of industrial effluents. Copies of the circuit diagram can be obtained from the author on request. The author is grateful to Dr. D. J. Greenwood for his advice and encouragement and to Mr. A. Barnes for statistical advice during the preparation of this paper. References 1. 2. 3. Baker, K. F., Analyst, 1970, 95, 885. Goodman, D., Lab. Pract., 1975, 24, 589. Bremner, J. M., in Black, C. A., Editor, “Methods of Soil Analysis, Part 2,” American Society of Agronomy, Madison, 1965, p. 1191. Received June 21st, 1976 Accepted July 27th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100943

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, December, 1976, Vol. 101, $9. 949-955 949 Determination of Molybdenum in Geological Materials by Atomic-absorption Spectrophotometry P. Sutcliffe Mount Morgan Limited, Mount Morgan, Queensland 47 14, Australia A simple, direct method for the determination of molybdenum in geological samples by atomic-absorption spectrophotometry has been developed. The interferences caused by some common metals and acids have been investi- gated. The procedure is free from interferences, and a t least as accurate as the thiocyanate colorimetric method a t molybdenum levels below 25 mg kg-l while still suitable for use a t levels of more than 1 000 mg kg-l. As early as 1960, David1 proposed the use of atomic-absorption spectrophotometry for the determination of molybdenum; however, since that time it has received far less attention than most other metals. Because of the poor sensitivity of molybdenum in atomic-absorption spectrophotometry several workers293 have preferred preliminary concentration by extraction with an organic solvent and almost all w ~ r k e r s l $ ~ - ~ have used either ammonium or aluminium chloride to inhibit interferences.The findings of these workers were not always in agreement ; for example, McIsaac4 stated that, even in the presence of aluminium chloride, iron and manganese formed complexes with the molybdenum and caused interference, while on the other hand Van Loon5 and Ramakrishna et aZ.6 claimed that interferences by iron and man- ganese were overcome by the use of aluminium chloride. In spite of, or perhaps because of, the work which has been done on the atomic-absorption spectrophotometric methods for molybdenum, many chemists consider atomic-absorption methods for molybdenum to be unreliable and prefer to use either the thiocyanate or toluene- 3,4-dithiol colorimetric methods.The thiocyanate method is the most sensitive of the colorimetric methods and although it has been used, in a variety of materials, for over 20 years by the author, it does have the following major limitations. (a) Precipitation of a copper thiocyanate complex in the presence of relatively small amounts of copper. This precipitate can adsorb some of the colour of the molybdenum thiocyanate complex. (b) Chromium, tungsten, cobalt and lead can also interfere by either precipitating or forming interfering coloured complexes with the thiocyanate.(c) The timing of the steps of colour formation, extraction by an organic solvent, usually diisopropyl ether, and taking an instrument reading, is so critical that batches of about 30 samples are the maximum that can be handled with accuracy at any one time. (d) The use of diisopropyl ether for extractions can give rise to evaporation problems, particularly in non-air-conditioned laboratories in hot climates, and its use in an open laboratory causes severe safety hazards. The object of this investigation, therefore, was to develop a simple, direct method for the determination of molybdenum in geological samples by atomic-absorption spectrophotometry. The method should be free from interferences and at least as accurate as the thiocyanate method below 25mg k g l of molybdenum while still suitable for use at molybdenum levels up to at least 1 000 mg kg-1.Experimental Interferences A nitrous oxide - acetylene flame was found to be the most sensitive and should be set in a reducing condition with the acetylene flow-rate as high as possible without the flame becoming luminous. Even under these conditions it can be seen from Table I that all of the metals tested showed either positive or negative interferences. Following the practice of previous w0rkers,~~3-~ it was decided to test the effectiveness of aluminium in suppressing these interferences. Various amounts of a solution obtained by dissolving pure aluminium foil in sufficient analytical-reagent grade hydrochloric acid to give a solution containing an excess of 10% V/V of free hydrochloric acid were added to solutions of the metal salts.The950 SUTCLIFFE : DETERMINATION OF MOLYBDENUM IN GEOLOGICAL Analyst, VoZ. 101 TABLE I ATOMIC-ABSORPTION SPECTROPHOTOMETER READINGS (SCALE DIVISIONS) CAUSED BY INTERFERING METALS Each solution contained 10 mg 1-l of molybdenum in 10% hydrochloric acid. Amount of interfering metal added*/mg 1-1 7 A \ Metal used 0 Fe . . . . 25.0 Na . . . . 25.0 c u . . . . 25.0 Pbt .. . . 25.0 Zn . . . . 25.0 Ni . . . . 25.0 MnS .. . . 25.0 Mg . . . . 25.0 Ca . . . . 25.0 K . . . . 25.0 wg . . . . 25.0 was used. keep the lead in solution. * Unless otherwise stated 5% nitric acid was used 1 000 25.0 15.0 34.0 38.0 40.0 38.0 34.0 19.8 10.2 18.0 - 2 000 25.0 15.0 35.5 40.0 42.5 41.5 31.0 22.5 10.6 15.0 -- 4 000 22.0 16.2 36.0 40.0 43.0 41.5 30.0 26.2 11.2 14.0 - 6 000 20.3 18.0 37.0 41.0 43.0 41.0 30.0 28.0 12.6 13.0 24.6 10 000 19.8 18.0 39.0 42.0 43.0 40.0 30.0 33.0 14.4 13.0 23.6 the chloride of the metal in 10% hydrochloric acid in addition to the 10% hydrochloric acid in order to $&Manganese sulphate in 10% hydrochloric acid was used.$ Owing to the high insolubility of tungsten(V1) oxide a weighed amount of tungsten(V1) oxide was added to the solvent acid and the full procedure carried out. Most of the tungsten(V1) oxide remained insoluble under these conditions. The use of soluble sodium wolframate(V1) was also tried but the tungsten(V1) oxide still precipitated out and the results were similar to those for sodium alone.results are shown in Table I1 and it was decided, at least initially, to examine the use of I000 mg 1-1 of aluminium as a possible suppressing agent. For the test, the amounts of interfering metals giving the most severe interference were selected from Table I and to them were added 10 mg 1-1 of molybdenum and 1000 mg 1-1 of aluminium. The results, shown in Table 111, indicated that the addition of aluminium was successful in suppressing the interference from those metals whether they were present individually or together. TABLE I1 EFFECT OF ALUMINIUM ON MOLYBDENUM ABSORPTION READINGS Each solution contained 10 mg 1-1 of molybdenum. Aluminium added/mg 1-l . . . . 0 250 6001 750 1000 1250 1500 2000 AAS reading (scale divisions) . . 25 37 391 43 43 43 43 43 An additional advantage of the addition of aluminium was the enhancement of the molyb- denum absorption.This enhancement, which was constant whether interfering metals were present or not, considerably improved the sensitivity of the method. Although David1 and Ramakrishna et aL6 tested many more possible metal interferents it was decided that, as all of the metals likely to be present in major amounts in geological samples had been tested, no further interference tests for metals were necessary. I t was expected that most of the silicon TABLE I11 EFFECT OF ALUMINIUM AS AN INTERFERENCE SUPPRESSOR Each solution contained 10 mg 1-' of molybdenum and 1 000 mg 1-1 of aluminium in 10% V/V hydrochloric acid. Metal tested . . .. None K Fe Na CU Pb* Zn Ni Mn Mg Ca W Amount added/mg 0 6000 10000 2000 10000 1ofioo 10000 1OOC)o 1000 10000 1000 5000 AASreading(sca1edivisionsj) 43.0 43.0 43.5 43.1 42.8 43.5 43.0 43.0 43.3 43.0 42.7 43.0 * 5% of nitric acid was also added to keep the lead in solution.t A solution of all of the above amounts of metals (i.e., containing about 10 g of solid) gave a reading of 35.0; however, a dilution of the sample by a factor of four gave a reading Of 10.6 (4 X 10.5 = 42), which indicates that the low reading was probably due to the high viscosity of the solution with such a large amount of solids presmt.December, 1976 MATERIALS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 951 would remain insoluble after digestion of the sample with acid and would cause no chemical or spectral interference.It was thought that the acid used for dissolution might have some interfering effect and several acids commonly used for this purpose were tested both with and without the addition of aluminium. The results given in Table IV indicate that aluminium is only partially successful in removing interference by nitric and perchloric acids. With sulphuric and ortho- phosphoric acids the severe depression of absorption at higher concentrations was assumed to be due mainly to the increase in viscosity of the solution. From the results shown in Table IV it was decided that the final solution should be virtually free from sulphuric, ortho- phosphoric, nitric and perchloric acids and should be approximately 10% V/V in hydro- chloric acid. TABLE IV (SCALE DIVISIONS) EFFECTS OF COMMON ACIDS ON MOLYBDENUM ABSORPTION READINGS Each solution contained 10 mg 1-l of molybdenum.Acid concentration, % Acid HCI . . HC1 . . HNO, HNO, HC10, HClO, H2SO4 H2S04 HPO4 H3PO4 A1 added/mg 1-l 0 .. 0 25.0 .. 1 000 43.0 .. 0 25.0 .. 1000 43.2 .. 0 25.0 .. 1 000 42.9 .. 0 25.0 . . 1 000 43.0 .. 0 25.0 . . 1000 43.0 5 25.0 43.0 28.0 41.8 14.0 43.0 38.0 45.5 34.5 43.0 10 25.0 42.8 24.5 41.8 11.0 41.0 38.0 46.0 35.0 42.6 15 25.5 42.2 24.5 41.8 9.0 38.0 39.0 46.5 33.5 39.5 26 24.5 43.2 23.0 41.5 8.0 35.0 38.5 46.3 30.5 34.0 Method The method of sample dissolution described under Procedure B below has been in use for many years in this laboratory for the determination of molybdenum by the thiocyanate method and was known to be efficient for dissolving all of the molybdenum in the sample.An advantage of this method was that after determining the molybdenum content by pro- cedure B an aliquot of the same solution could be used for a colorimetric determination, giving a direct comparison of the two methods. Reagents All reagents should be of analytical-reagent quality unless otherwise specified. Aluminium clzloride solution. To 10 g of pure aluminium foil add 160 ml of distilled water and 210 ml of hydrochloric acid (sp. gr. 1.18). When the vigorous reaction has ceased, heat the mixture gently to complete the dissolution, then cool and dilute to 1 1 with distilled water. Solzrent acid. To 600 ml of nitric acid (sp. gr. 1.42) add very carefully 300 ml of sulphuric acid (sp. gr. 1.84). Stock molybdenum soZution A. Dissolve 1.0 g of 99.9% pure molybdenum in hot nitric acid (sp.gr. 1.42) and evaporate the solution just to dryness on a water-bath. Re-dissolve the residue in 100 ml of hydrochloric acid (sp. gr. 1.18) and dilute to 1 1 with distilled water. 1 ml of solution = 1 000 pg of molybdenum. Stock molybdenum solution B. hydrochloric acid (sp. gr. 1.18). Dilute 100 ml of stock solution A to 1 1 with 10% VlV 1 ml of solution = 100 pg of molybdenum. Calibration solutions. Transfer by pipette 1.0-, 2.0-, 3.0-, 4.0-, 5.0-, 7 . 5 , 10.0-, 15.0- and Add 10 ml 20-ml volumes of stock molybdenum solution B into 100-ml calibrated flasks.952 SUTCLIFFE : DETERMINATION OF MOLYBDENUM I N GEOLOGICAL of the aluminium chloride solution to each and dilute to the mark with chloric acid (sp.gr. 1.18). Dissolution of Samples An.alyst, Vol. 101 10% V/V hydro- Procedure A: samples containing 0-100 mg k g 1 of molybdenum Weigh 5.0 g of sample into a 250-ml heat-resistant beaker, add 15 ml of hydrochloric acid (sp. gr. l.lS), 5 ml of nitric acid (sp. gr. 1.42) and 10 ml of perchloric acid (sp. gr. 1.54). Heat to dissolve the sample, then evaporate the solution to dryness and continue heating until the dense white fumes of the acid just disappear. Allow to cool, then re-dissolve the salts by warming gently with 10 ml of hydroch1orj.c acid (sp. gr. 1.18). After cooling, add 10ml of the aluminium chloride solution and dilute t o 100ml with distilled water. Mix thoroughly and allow to settle for 30 min before taking a reading on the atomic-absorption spectrophotometer.Procedure B: samples containing over 100 mg kg-1 of molybdenum Weigh 2.0 g of sample into a 250-ml heat-resistant beaker and add 20 ml of solvent acid. Heat to dissolve the sample, then evaporate the solution to dryness and continue heating until the dense white fumes of the acid just disappear. Allow to cool, then re-dissolve the salts by warming gently with 10 ml of hydroch1or:ic acid (sp. gr. 1.18). After cooling, add 10ml of the aluminium chloride solution and dilute to 100ml with distilled water. Mix thoroughly and allow to settle for 30 min before taking a reading on the atomic-absorption spect rophot omet er . Instrument Conditions A Varian Techtron, Model AA5, atomic-absorption spectrophotometer was used throughout with the following conditions: wavelength, 313.3 nin; slit width, 50 pm; spectral band pass, 0.2 nm ; lamp current, 12 mA (hollow-cathode lamp for molybdenum) ; flame, nitrous oxide - acetylene; and scale expansion.Details of conditions and precautions are given in Notes 1-7. NOTES 1. The nitrous oxide - acetylene flame should be set for maximum sensitivity. This peak was found to occur with the acetylene flow-rate set to the maximum possible without the flame becoming luminous. The actual flow-rates of both the acetylene and nitrous oxide were found to vary with the condition of the nebuliser in use at the time. Maximum sensitivity was found to occur when the centre of the light beam was 5 mm above the burner. When examining samples containing less than 100 mg kg-1 of molybdenum the scale expansion was adjusted so that a 5 mg 1-1 standard molybdenum solution read 50 scale divisions.This scale expansion was approximately x 7 and made each scale d.ivision equivalent to 2 mg kg-1 of molybdenum in the original soil sample when a 5.0-g sample was used. Care should be taken to avoid spitting of the sample when evaporating a 5-g sample to dryness. For the highest accuracy each sample should be read a t least twice and standard solutions should be read after every eight samples. Because of the high salt content of the sample solutions, i t is essential to aspirate distilled water between each sample and to clean out the nebuliser and burner slot after each run or after 200 samples, whichever occurs first. The calibration solutions prepared will cover the range 2-400 mg kg-' of molybdenum on a 5-g soil sample and 5-1 000 mg kg-l on a 2-g sample.For samples containing more than 1 000 mg kg-l i t is suggested that a smaller amount of sample be taken. 2. 3. 4. 5. 6. 7. Samples As no standard samples were available for testing the experimental procedure, it was decided to make our own standards using routine samples. Eleven arbitrary values up to 1 500 mg k g l of molybdenum were chosen and routine samples near to these values, as determined by the thiocyanate method, were bulked together until several kilograms of each mixture were obtained. These composite samples were then re-pulverised and re-mixed until thoroughly homogeneous samples with a particle size of less than 200 mesh were obtained. Results Table V shows a comparison of sets of results obtained by the thiocyanate method and The calculated results are the arithmetical averages of the thiocyanate by procedure B.December, 1976 MATERIALS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 953 results obtained in the routine analyses of the individual samples making up each composite.The agreement between the average results for each method is very good with the possible exceptions of samples Nos. 2 and 4; however, it is noticeable that the spread of results favours the thiocyanate method below 50 mg kg-l and the atomic-absorption spectr ophotometric method above 50 mg k g l . Sample No. 1 2 3 4 6 6 7 8 9 10 11 Sample No. 1 2 3 4 5 6 7 8 9 10 11 TABLE V COMPARISON OF ATOMIC-ABSORPTION AND COLORIMETRIC METHODS Results are molybdenum contents of samples (mg kg-l).Calculated result 0 5 10 25 50 80 130 150 250 320 1 500 Calculated result 0 5 10 25 50 80 130 150 250 320 1 500 Results obtained using colorimetric method , 0 6 15 20 80 85 130 160 240 320 1500 3 5 15 24 40 90 125 200 250 375 1 400 3 5 15 20 50 80 150 120 280 400 1 300 I - 3 6 5 7.5 - 12.5 15 12.5 20 20 50 55 80 75 - - - 130 115 110 120 125 - 250 220 240 375 340 340 1500 1600 1 550 Results obtained by AAS method f 0 0 5 5 12 15 25 20 43 50 83 85 132 125 153 162 260 240 365 375 1 520 1525 - - 16 28 50 82 120 140 235 345 1340 5 10 15 20 55 82 130 147 235 340 1440 3 - 3 12 14 18 - - - - - 84 85 135 135 157 153 240 250 350 370 1475 1 500 Average of colori- metric results 3.0 5.7 14.2 20.8 51.0 82.0 126.7 145.0 246.7 358.3 1475.0 Average of AAS results 2.0 8.0 15.0 23.3 49.5 83.5 129.5 152.0 243.3 357.5 1466.7 Three synthetic samples were each prepared by mixing 0.5 g of carefully selected molyb- denum sulphide crystals with 300 g of previously analysed barren quartz and pulverising the mixture in a Zieb pulveriser. After thorough mixing, each sample was split by riffling into eight portions and molybdenum determined in each portion.The close agreement of the eight results in each group showed the samples to be homogeneous so the eighth sample was analysed eleven times by both the thiocyanate and the atomic-absorption spectrophoto- metric methods to give some indication of the precision at high levels. The molybdenum sulphide crystals were analysed in triplicate by a gravimetric procedure and shown to be 95.82% molybdenum( IV) sulphide.Assuming no losses had occurred in sample preparation, this result gave a theoretical value for the samples of 956.3 mg k g l of molybdenum. From Table VI it is obvious that the precision of the atomic-absorption spectrophotometric method is good and far superior to that of the thiocyanate method at these levels. An attempt was made to improve the accuracy of the atomic-absorption method a t the lower levels at which the thiocyanate method normally excelled. One hundred routine samples with low molybdenum contents were selected and each was analysed by the atomic- absorption procedure B but using a 5.0-g sample in order to improve the accuracy. An aliquot was taken from each solution after taking a reading on the atomic-absorption spectro- photometer and used to determine the molybdenum contents by the thiocyanate procedure.The results are shown in Table VII. All of the samples contained less than 25 mg k g l of molybdenum and a standard deviation of the difference between the thiocyanate and atomic- absorption methods at this level was calculated from Table VII to be h1.97 mg kg1.954 SUTCLIFFE : DETERMINATION OF MOLYBDENUM I N GEOLOGICAL Analyst, T/Ol. 101 TABLE VI COMPARISON OF ATOMIC-ABSORPTION AND COLORIMETRIC METHODS FOR SAMPLES WITH HIGH MOLYBDENUM CONTENTS Molybdenum content/mg kg-l A I > Colorimetric method AAS method r A \ f A \ Result No. 1 2 3 4 5 6 7 8 9 10 11 Average . . .. Standard deviation Spread of results . . Sakple Sample Sample Sample Sample Sample No.1A No. 2A No. 3A No. 1A No. 2A No. 3A 750 750 800 800 800 750 800 800 800 800 800 7 50 800 800 800 850 800 800 900 900 850 800 850 800 900 800 900 700 900 800 800 800 800 970 920 950 960 960 950 960 960 960 960 960 940 950 940 960 930 930 910 950 920 950 930 920 940 930 910 920 930 920 880 920 910 910 786 823 823 956 937 917 f23.34 f46.71 f60.67 f 12.95 f 14.90 zk 15.66 50 150 200 60 60 60 Discussion From Tables V, VI and VII it can be seen that the proposed method gives reproducible results that are close to those given by the thiocyanate method. A notable feature, however, was that the greatest discrepancies between the methods appear to occur when the colori- metric method shows no molybdenum. This error was later found to be caused by the TABLE VII COMPARISON OF ATOMIC-ABSORPTION AND THIOCYANATE COLORIMETRIC METHODS ON 100 SOILS CONTAINING LESS THAN 25 mg k g l OF MOLYBDENUM Molybdenum content*/mg kg-l AAS 6 5 4 7 8 9 20 13 22 20 23 2 2 0 3 0 1 8 5 12 Colori- metric 3 5 9 7 9 12 22 16 22 22 26 0 3 0 0 0 1 8 0 14 AAS 6 9 12 4 2 1 4 6 1 1 3 3 6 4 4 10 4 2 4 3 Colori- metric 6 12 14 0 1 4 2 5 0 1 2 2 5 3 3 8 3 2 4 3 AAS 3 3 4 5 4 3 5 5 9 4 3 12 22 7 9 5 3 2 4 3 Colori- metric 1 2 0 4 2 2 4 4 9 5 4 16 24 7 9 4 0 0 1 5 AAS 3 5 9 12 2 3 8 12 10 9 G 6 3 4 10 9 5 2 3 5 Colori- metric 3 4 8 10 2 1 7 10 10 9 6 3 3 1 11 8 2 4 5 4 AAS 5 2 3 3 4 8 6 12 7 3 7 2 2 2 4 4 4 2 8 2 coloii- metric 3 2 3 2 2 9 10 12 6 6 7 3 1 4 1 1 0 3 9 6 * A Student’s t-test of the results showed no significant difference between the methods, giving a mean difference of 0.3 f 0.38 mg kg-l a t the 95% confi.dence level.A more realistic indication of the difference between the methods can be calculated by using the difference between each A.4S and colorimetric result as ( x - 5) in the standard deviation equation S.D. = d l ( n - 1) C ( x - z)a. This procedure gives a standard deviation of the difference between the methods of f 1.97 mg kg-1 of molybdenum.December, 1976 MATERIALS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 955 instability of the atomic-absorption spectrophotometer zero when the fairly high scale expansion ( x 7 or x 8) was used. For the greatest accuracy, care must be taken that the needle or digital read-out does not drift away from zero in the short time between releasing the auto-zero and aspirating the sample.It is recommended that when time permits each sample should be read at least twice in order to overcome this problem. During the evaluation of the modification of the method for low levels of molybdenum, serious blockages occurred in the nebuliser when spraying solutions of a 5-g sample. These blockages could not be removed fully by either chemical or physical methods and, as they occurred after aspirating only about 200 samples, the use of a 5-g sample was temporarily discontinued owing to the high cost of nebulisers. Because of the excellent results obtained by the use of a 5-g sample, further investigations were carried out at a later stage. The author thought that the blockages were being caused by some form of insoluble sulphate scale and to overcome this effect treatment of the 5-g sample with aqua regia - perchloric acid was tried.This method of acid dissolution of the sample appeared to cure the blockage problem and the measured aspiration rate of a nebuliser remained unchanged after aspirating over 1000 samples. As the results using the aqua regia - perchloric acid method agree very closely with those from the nitric acid - sulphuric acid method, the aqua regia - perchloric acid method, pro- cedure A, is recommended when using 5-g samples. Using procedure A, samples have also been read on the atomic-absorption spectrophotometer for copper, zinc and lead, giving results which agree closely with those obtained by use of our normal atomic-absorption met hod. Conclusions Owing to the lack of standard samples it is not possible t o give an absolute figure for the accuracy of the method; however, from the evidence obtained it would seem that the pro- cedure described will give results at least as accurate as those of the thiocyanate colorimetric method. Below 25 mg k g l of molybdenum a standard deviation, between the methods, of better than A1.97 mg k g l is possible. At higher levels the suggested atomic-absorption procedure is far more precise with an accuracy at least as good as, and probably better than, that of the thiocyanate colorimetric method. The author thanks Mount Morgan Limited for allowing this work to be carried out and for encouraging the author to publish this paper. References 1. 2. 3. 4. 6. 6. David, D. J., Analyst, 1961, 86, 730. Butler, L. R. P., and Mathews, P. M., Analytica Chim. Ada, 1966, 36, 319. Hutchinson, D., Analyst, 1972, 97, 118. McIsaac, C. L., Engng Min. J., 1969, 170, 155. Van Loon, J. C., Atom. Absorption Newsl., 1972, 11, No. 3. Ramakrishna, T. V., West, P. W., and Robinson, J. W., Analytica Chim. Acta, 1969, 44, 437. Received June 4th, 1976 Accepted August 2nd. 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100949

出版商:RSC    

年代:1976

数据来源: RSC