ISSN:0003-2654    

DOI:10.1039/AN98106FX045

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

iv SUMMARIES OF PAPERS IN THIS ISSUE December, 1981Summaries of Papers in the IssueEffect of Experimental Design on Assay CalibrationCriteria are developed for the maximisation of the amount of informationavailable from an analytical calibration graph. For linear calibration graphs,the methods involve replicate determinations of standards at two points,which are chosen so that they occur a t the extremes of the calibration range.The increase in precision that can be obtained by using an optimised designcan be considerable, particularly when the variance in the assay errors is notconstant. The methods described can be extended readily to non-linearcalibration graphs.Keywords 1 Assay design ; sensitivity ; weighted least squares ; optimised design ;Calibration graphLEON AARONSPharmacy Department, University of Manchester, Manchester, M13 9PL.Analyst, 1981, 106, 1249-1254.Plasma Source Mass Spectrometry Using an Inductively CoupledPlasma and a High Resolution Quadrupole Mass FilterThe performance of an instrument using an inductively coupled plasma asan ion source with a high resolution mass filter to analyse the extracted ionsis described.A description of the experimental system is given and thespectral characteristics of the technique are discussed. The current per-formance of the technique for simple solution analysis is described, and thefuture potential for trace element analysis and isotope ratio measurement isassessed.Keywords : Inductively coupled plasma ; atomic mass spectrometry ; highyesolution quadrupole filter ; multi-channel scaler ; trace element determinationALAN R.DATEInstitute of Geological Sciences, 64/78 Gray’s Inn Road, London, WClX 8NG.and ALAN L. GRAYDepartment of Chemistry, University of Surrey, Guildford, GU2 5XH.Analyst, 1981, 106, 1255-1267.Extractive Spectrophotometric Determination of Niobium inPyrochlore-bearing Rocks with 5,7-Dichloroquinolin- 8-01A spectrophotometric method for the determination of trace amounts ofniobium(V) based on its extraction into chloroform with 5,7-dichloroquinolin-8-01 from a hydrochloric acid medium has been developed. The maximum ab-sorbance of the extracted species occurs at 400 nm, the molar absorptivitybeing (1.28 & 0.02) x lo4 1 mol-1 cm-l. Beer’s law is obeyed over the range10-80 pg of niobiuni(V) extracted.The relative standard deviation is 0.92%and the recovery of niobium is 99.6 & 0.5%. The method is particularlysuitable for the determination of niobium in ores and rocks, as commonlyinterfering elements present in niobium ores did not interfere in this determ-ination. Results of the successful analysis of standard and synthetic niobiumores with niobium contents ranging from 0.1 to 1 .O% are given.A useful modification of Faye’s method for dissolution of niobium ores isreported. The special characteristics and importance of drying the organicphase in the spectrophotometric measurements are discussed.Niobium determination ; ore dissolution ; oye analysis ; extraction KeywordsspectroplaotometryA. SANZ-MEDEL and M.E. DIAZ GARCIADepartment of Analytical Chemistry, Faculty of Sciences, University of Oviedo,Oviedo, Spain.Analyst, 1981, 106, 1268-1274v1 SUMMARIES OF PAPERS IN THIS ISSUEPotentiometric Determination of Fluoride by a Combination ofContinuous-flow Analysis and the Gran Addition MethodDecember, 1981A method is described for the determination of fluoride ion by continuous-flowanalysis combined with the Gran addition method using an ion-selectiveelectrode.The Gran addition method is a reliable potentiometric technique and anadvantage of its use is that, in addition t o the determination of the free andtotal fluoride concentration in a complexed medium, the complexation rate isalso measured. The drawback of the method is the difficulty in adjusting theperistaltic pumps, which is time consuming.However, this extra time is com-pensated for by the simultaneous determination of fluoride ion concentrationand the complexation rate.For concentrations of fluoride ion between 0.5 and 100 mg 1-l, the precisionis 5%.Keywords : Fluoride determination ; continuous-flow analysis ; ion-selectiveelectrode ; Gran addition method ; air pollutionJ . 4 1 . LANDRY, F. CUPELIN and C. MICHALService de Toxicologie Industrielle, d’Analyse de 1’Air e t de Protection Contre leBruit, Institut d’Hygibne, Case Postale 109, CH-1211 Genbve 4, Switzerland.Analyst, 1981, 106, 1275-1280.Sequential Multi-element Analysis Using Silver and CopperIon-selective ElectrodesA simple method is described for the rapid and accurate sequential deter-mination of as little as 50 pg ml-1 of Ag+, Cu2+, Ni2+, Cd2+, Pb2+, Zn2+, Fe3+,Th*+ and V4+ ions in binary, tertiary and quaternary mixtures.The methodinvolves direct titration with sodium diethyldithiocarbamate at pH 4-6 in50% V / V ethanol using silver and copper ion-selective electrodes. Potentio-metric curves with sharp consecutive inflection breaks at the equivalencepoints and a mean recovery of 99.6% (mean standard deviation 0.3%) areobtained. The method has been used satisfactorily for the determination ofthe major alloying elements in brass, motor-brass, silver-brazing and nickel-silver alloys by both direct and spiking titration techniques. The resultsagreed well with the certified values.Keywords : Ion-selective electrodes ; copper- and silver-base alloys ; sequentialdetermination of metals ; diethyldithiocarbamateSAAD S. M. HASSAN and M. M. HABIBDepartment of Chemistry, Faculty of Science, Ain Shams University, Cairo, Egypt.Analyst, 1981, 106, 1281-1287

ISSN:0003-2654    

DOI:10.1039/AN98106FP145

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

December, 1981 SUMMARIES OF PAPERS I N THIS ISSUEDifferential-pulse Voltammetric Determination of Phosphate asMolybdovanadophosphate at a Glassy Carbon Electrode andAssessment of Eluents for the Flow Injection VoltammetricDetermination of Phosphate, Silicate, Arsenate and GermanateixThe redox behaviour of molybdovanadophosphate at a glassy carbon electrodeis described and a procedure is given for the voltammetric determination ofphosphate as molybdovanadophosphate at a glassy carbon electrode in astatic system. Procedures are also given for the voltammetric flow injectiondetermination of phosphate, silicate, arsenate and germanate by the injectionof heteropolyacids pre-formed in various aqueous, aqueous acetone andaqueous ethanolic reagents into eluents consisting of the reagent blank.This procedure effectively eliminates the background signal of the blank andallows precise determinations to be made. Plots of electrode potentia1against the current obtained are given.Silicate and phosphate can be determined a t lo-' and M levels,respectively.M level, andthe precise determination of germanate is difficult owing to adsorption atthe glassy carbon electrode.Arsenate has only been determined at theKeywords ; Orthophosphate, silicate, arsenic and germanium determination ;voltammetry ; $ow injection analysisA. G. FOGG and N. K. BSEBSUChemistry Department, Loughborough University of Technology, Loughborough,Leicestershire, LE11 3TU.Analyst, 1981, 106, 1288-1295.Use of Ascorbic and Thioglycollic Acids to Eliminate Interferencefrom Iron in the Aluminon Method for Determining AluminiumThe use of ascorbic and thioglycollic acids as inhibitors for the interference ofiron in the aluminon method of Hsu have been examined. The use of ascorbicacid, as proposed by Jayman and Sivasubramaniam, has been found tochange iron interference from positive to negative causing aluminium to beunderestimated.However, the addition of 0.2 ml of a 1% V / V solution ofthioglycollic acid to solutions containing aluminium in amounts ranging from10 to 50 pg has been proved to suppress the interference from up to 900 pgof iron.Keywords ; Aluminium determination ; aluminon method ; ascorbic acid; ironinterference ; thioglycollic acidF. CABRERA, L. MADRID and P.DE ARAMBARRICentro de Edafologia y Biologia Aplicada del Cuarto, Apartado 1052, Seville, Spain.Analyst, 1981, 106, 1296-1301December, 1981 SUMMARIES OF PAPERS I N THIS ISSUEEffect of Temperature on the Structural Rearrangements ofPolyesters (LAC-series) when Used as Liquid StationaryPhases in Gas - Liquid ChromatographyxiThis study deals with the effect of temperature on the gas-chromatographicbehaviour of thiols, alkanes, branched-chain alkanes and cyclic alkanes onpolyester (LAC-446, LAC-745, LAC-772, LAC-860, LAC-886, LAC-841 andLAC-935) liquid stationary phases. Increases were observed in specificretention volumes as well as improvements in the separations a t certaintemperatures. The critical temperatures were different for the variousliquid stationary phases; these were 110 "C for LAC-446 and LAC-886 and120 "C for LAC-860 and LAC-772.Differential-thermal analysis of two ofthe liquid stationary phases indicated a change in the physical properties ofthe polymers during heating.Keywords : Gas - liquid chromatography; liquid stationary phases ; effect oftemperature ; LA C-series of polyesters ; structural rearrangementsALBERTINE E. HABBOUSH, SABRI M. FARROHA, AL'A K. ABDULAL-SADA and ABDUL MASSEH N. KITTOCollege of Science, University of Baghdad, Adhamiya, Baghdad, Iraq.Analyst, 1981, 106, 1302-1308.Rapid Automated Enzymatic Method for the Determination ofAlcohol in Blood and Beverages Using Flow Injection AnalysisAn enzymatic method for the determination of alcohol using flow injectionanalysis is described.Samples are suitably diluted and introduced directlyinto the system. Blood alcohol is analysed by the stop-flow technique at arate of 70-80 samples per hour and the results are compared with those fromheadspace gas chromatography. Alcohol in several beverages is analysed bythe continuous-flow technique a t a rate of up to 120 samples per hour. Inboth instances the result is available less than 30 s after injection.Keywords : Flow injection analysis ; alcohol determination ; enzymatic method ;whole blood ; stop-flow techniqueP. J. WORSFOLD, J. RGZIeKA and E. H. HANSENChemistry Department A, Technical University of Denmark, Building 207, DK-2800Lyngby, Denmark.Analyst, 1981, 106, 1309-1317.A Copper(1) Iodide Paper for the Detection and Determination ofthe Concentration of Mercury Vapour in the Workplace AtmosphereThe development and preparation of a test paper for mercury in air isdescribed.A measured volume of air is drawn through a silica gel-loadedfilter-paper coated with a mixture of copper(1) iodide and sodium carboxy-methylcellulose, and in the presence of mercury a pink stain is produced.The paper has been validated in the laboratory and a factory for the reliabledetermination of mercury vapour in air a t concentrations up to 100 pg m-3.Keywords : Mercury vapour ; copper(I) iodide test paper ; workplaceatmosphere ; airS. CRISP, D. W. MEDDLE, J. M. NUNAN and A. F. SMITHDepartment of Industry, Laboratory of the Government Chemist, Cornwall House,Stamford Street, London, SE1 9NQ.Analyst, 1981, 106, 1318-1325December, 1981 SUMMARIES OF PAPERS I N THIS ISSUEOpen-cell Polyurethane Foam as a Sorbent in the Extractionof Iodine-131xiiiOpen-cell polyurethane foams have proved to be effective as sorbents forinorganic and organic species and as supporting materials for varioushydrophobic organic phases.They possess outstanding sorption, mass-transfer and hydrodynamic properties, which enable them to be used a trelatively high flow-rates of aqueous solutions during column operation.Iodine-131 can be extracted from samples of water and milk using resilientopen-cell polyurethane foams in static and pulsed column beds, when in theform of a cylindrical packing impregnated with a tri-alkylamine containingdissolved inactive iodine.Pulsed column beds, which have some advantagesover static column beds, can be conveniently automated. This may lead tothe development of a new type of environmental monitoring device forradioactive iodine.Keywords Iodine- 13 1 extraction ; Polyurethane foam ; water ; milk ; sorbentS. PALAGYIInstitute of Radioecology and Applied Nuclear Techniques, P.O. Box A-41, 04061Kogice, Czechoslovakia.and T. BRAUNInstitute of Inorganic and Analytical Chemistry, L. Eotvos University, P.O. Box 123,1443 Budapest, Hungary.Analyst, 1981, 106, 1326-1333.Reactive Ion-exchange Precipitation Procedure for the Determinationof Trace Amounts of OxalateThe feasibility of using reactive ion-exchange for the determination of oxalatein aqueous solution has been demonstrated.The procedure consists of pHadjustment of the analyte solution, concentration of oxalate on a lead(I1)-loaded cation exchange mini-column in the form of lead oxalate precipitate,reactive elution of oxalate by sulphuric acid dissolution and analysis of theconcentrated eluate by permanganate titration. Almost quantitative yieldsof oxalate are obtained for concentrations typically found in human urine.The procedure represents a simple and reliable technique to concentrateoxalate for determination by classical methods.Keywords : Oxalate trace determination ; reactive ion exchange ; precipitation ;titrationSUSAN M. ANDEL, GILBERT E. JANAUER and WILLIAM E. BERNIERDepartment of Chemistry, State University of New York a t Binghampton, Bing-hampton, N.Y.13903, USA.Analyst, 1981, 106, 1334-1337.Sample Fusion at Low Temperature for the PotentiometricDetermination of Fluorine in Silicate MaterialsShort PaperKeywords : Fluorine determination ; silicates ; alkaline fusion ; fluoride ion-selective electrodeB. FABBRI and F. DONATIC.N.R., Research Institute for Ceramics Technology, Via Granarolo 6, 48018 Faenza,Italy.Analyst, 1981, 106, 1338-1341xiv SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Fluoride Ion in Bovine Milk Using a FluorideIon- selective ElectrodeDecem bey, 1981Short PaperKeywords Fluoride determination ; milk ; ion-selective electrodeCLIFFORD G. BEDDOWSSchool of Health and Applied Sciences, Leeds Polytechnic, Leeds, LS1 3HA.and DAVID KIRKDepartment of Hotel and Catering Studies, Sheffield City Polytechnic, Sheffield,S1 1WB.Analyst, 1981, 106, 1341-1344.Use of a Carbon Fibre Indicator Electrode for the PotentiometricTitration of Glycine in Glacial Acetic AcidShort PaperKeywords : Glycine determination ; potentiovnetric titration ; carbon fibreindicator electrodeV. J. JENNINGSDepartment of Applied Chemistry, Coventry (Lanchester) Polytechnic, PrioryStreet, Coventry, CV1 5FB.Analvst, 1981, 106, 1344-1347.Reagent for the Spectrophotometric Determination of Iron( 11) inAlkaline SolutionShort PaperKeywords : Iron( I I ) deterininntion ; pv~idine-2-hydrazide veagent ; spectro-photo~netryKHALID A. ABDULLAH and YOUNIS I. HASSANDepartment of Chemistry and Biology, College of Education, University of hlosul,hIosul, Iraq.and A. M. AL-DAHER and W. A. BASHIRDepartment of Chemistry, College of Science, University of Mosul, Mosul, Iraq.Analjist, 1981, 106, 1348-1351

ISSN:0003-2654    

DOI:10.1039/AN98106BP151

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

DECEMBER 1981 Vol. 106 No. 1269 Effect of Experimental Leon Aarons Design on Assay Calibration Pharmacy Department, University of Manchester, Munchester, M13 9PL Criteria are developed for the maximisation of the amount of information available from an analytical calibration graph. For linear calibration graphs, the methods involve replicate determinations of standards at two points, which are chosen so that they occur at the extremes of the calibration range. The increase in precision that can be obtained by using an optimised design can be considerable, particularly when the variance in the assay errors is not constant. The methods described can be extended readily to non-linear calibration graphs. Keywords : Assay design ; sensitivity ; weighted least squares ; optimised design; calibration graph In most instances it is necessary to calibrate an analytical instrument so that the instrumental response can be converted into a more meaningful chemical variable such as concentration.One-point calibrations are usually not sufficient and a range of calibration standards is needed. Little attention has been paid to the pattern or arrangement of the calibration standards, the usual practice being to divide up the experimental region uniformly. The number and arrangement of experiments are known as the experimental design and it is the purpose of this paper to describe the effect of experimental design on assay calibration. The importance of experimental design can be best gauged if the consequences of a poor design are examined. Consider a linear calibration in which all readings are made a t just one point.Although this design will give a good estimate of the variance in the assay, it will give no information about the intercept and slope of the calibration line. In general, the experimental design influences the precision with which the parameters of the calibration line can be estimated. As the calibration is used predictively, it is the purpose of good experi- mental design to obtain the best possible predictive power. In the next section a general theoretical framework is developed to achieve this goal and in later sections it is applied in a variety of situations. General Theoretical Approach Mainly, linear calibrations will be examined although the results are easily extended to If yi is the response observed at a point x i , then the calibration those that are non-linear. can be described by equation (1) : where ti is a random variable.weighted sum of squared deviations (WSS), which is given by The parameters co and c1 are obtained by minimising the n .. * * (2) wss = Z ( y , - co - c,xpWi * . .. i = 1 where n is the number of data points and wi is a weighting factor equal to the reciprocal of the variance of the ith experiment. Because the data are “noisy” the parameters are subject to a degree of uncertainty that can be determinedl from the variance - covariance matrix, vih by V6 ==I [ATV-lA]-l .. .. .. * . (3) where V is the variance - covariance matrix of the errors, ti, and A is given by 12491250 AARONS : EFFECT OF EXPERIMENTAL DESIGN Analyst, Vol. 106 1 A = [1 i .... .. .. .. (4) The design criteria should be to choose the (x,> such that the variance of a prediction, x , is as small as possible. As will be shown later, this design will depend on x and is not very practical for a calibration that is to be used over a wide range of x values. A more realistic criterion is to choose the (xi> to give the best estimates of the parameters. This can be achieved2 by minimising the determinant of the variance - covariance matrix with respect to the (xi>. In order to make further progress the form of the variance - covariance matrix of the errors must be known. Two particular instances will be studied. Unweighted Linear Least Squares If the errors are randomly and normally distributed, uncorrelated and have a constant variance, 02, then equation (2) reduces to the familiar unweighted linear least-squares problem.The variance - covariance matrix becomes with the determinant (Det. Ve) given by In practice, the experimental region is bounded from below and above, x(L) and x(U), respectively. The optimised design is obtained by dividing the calibration standards into two groups, one half at x(L) and the other half at x(U). If there are an odd number of standards the final experiment can be done either at x(L) or x(U). In order to assess the effect of experimental design on assay calibration, the calculation of assay sensitivity was examined. After equation (1) has been fitted to some data, the variance (Ve) of a further observation, x, is given3 by The assay sensitivity is defined* as that value of x, xo, such t h a t the coefficient of variation is equal to loyo, i.e., Equations (7) and (8) can be solved simultaneously to define xo. Clearly from equation (7) the design that is going to give the lowest value of xo is the one for which xo = 5.Hubaux and VOS~ discussed designs of this kind and suggested concentrating most of the calibration standards at x(L). Although this design will give a good estimate of the assay sensitivity it will give fairly poor estimates of the parameters and consequently poor estimates of pre- dictions around x(U). Therefore, we*prefer to adopt designs that give the best parameter estimates. Four standards were examined for which c1 = 1, o = 0.1, x(L) = 1 and x(U) = 100, and two designs, A, in which x = 20, 40, 60 and 80 and B, in which x = 1, 1, 100 and 100, were investigated.The sensitivities calculated for design A and design B were 1.56 and 1.22,December, 1981 ON ASSAY CALIBRATION 1251 respectively. In the limit, as the number of calibration standards becomes very large, xo approaches 1 and it can be seen that the optimised design behaves better than the uniform design but the gain in efficiency is relatively small and will become even smaller as more calibration standards are used. Weighted Linear Least Squares When the errors obey all the conditions of the last section except that the variance, Vi, is a function of yi, the determinant of the variance - covariance matrix can be shown to be Four variance models were studied: model 1, for which V(yi) cc yi2; model 2, for which V(yi) m y i ; model 3, for which V(yi) CK yi-l; and model 4, for which V(yi) oc ya2.The meaning of the first model is that the confidence in an observation, yi, decreases as the value of yi increases. Before considering the optimised designs for these models it is interesting to investigate the consequences of neglecting weighting in least-squares analyses. To this end we generated some simulated data to which random noise of mean zero and variance proportional to yi2 was added. Both unweighted and weighted least-squares analyses were carried out and the results of these studies are shown in Table I. The parameters were equally well predicted by the two methods, whereas the sensitivity predicted by the unweighted analysis was considerably greater than in the (correct) weighted analysis.Thus, although weighting may be neglected for the purposes of determining the regression parameters of a calibration line, the calculation of assay sensitivity and the standard error of a prediction may be severely affected by this neglect. Of these models, probably the first two are the most common. (See also Garden et aL6) TABLE I EFFECT OF WEIGHTING IN LEAST-SQUARES ANALYSES The data were generated using an equation y = x . Ten equally spaced data points, between 0 and 1, were used and random noise of mean, zero and variance equal to 0 . 0 1 ~ ~ was added to the data. The values given in paren- theses are standard errors. The results in the table are the means of ten runs. co C1 XO* Unweighted least squares .. . . -0.0005 (-&0.0022) 1.0029 (f0.0064) 0.0625 (i0.0224) Weighted least squarest . . . . -0.0004 (&0.0005) 1.0027 (Lt0.0065) 0.0116 (&0.0001) * Estimated sensitivity. t See Appendix for details. The optimised designs for the four models described above all involve the performance of replicate experiments at two values of x. Taking again x(L) = 1 and x(U) = 100 the optimised designs are as follows: model 1, x1 = 1 and xz = 100; model 2, x1 = 1 and xz = 100; model 3, x1 = 33 and x 2 = 100; and model 4, xl. = 50.5 and x2 = 100. We again looked a t the calculation of assay sensitivity using four calibration standards, comparing design A and design B of the last section, in the instance where the errors are distributed according to model 2, i.e., the variance is equal to ky.The limiting sensitivity for large values of n (see Appendix for details) is xo (lim.) = 100k/c12 . . * . .. and for k = 0.01 and c1 = 1, xo (lim.) = 1. The sensitivities calculated for design A and design 13 were 6.18 and 1.36, respectively. The gain in precision obtained by the use of an optimised design is considerable, in contrast to what we saw in the last section where the error variance was constant. It should be pointed out that the weights, wi, needed to perform the analysis are generally unknown to the analyst. In The four models discussed represent idealised situations.1252 AARONS : EFFECT OF EXPERIMENTAL DESIGN Analyst, Vol. 106 practice the analyst must determine the variance structure of the calibration graph by performing replicate determinations over the working range of the assay.Once the variance structure is known the optimum design can be determined. Nonlinear Calibration Graphs The problem with the preceding analysis is that it assumes that the calibration is linear. If the calibration is not linear then the “optimised” design will not only be sub-optimum but it will give no possibility of detecting any departure from linearity. For well established assays that are known to be robustly linear, this should not be a problem. However, in circumstances where the linearity of the assay is in question precautions need to be taken. The simplest departure from linearity would be caused by the presence of a quadratic term in equation (1) : yi = co + C l X i + c2xi2 + & .. .. .. . (11) The optimised design for this model, assuming that the errors are randomly and normally distributed, uncorrelated and have a constant variance, involves replicate measurements of standards in three experimental regions, viz., at both ends of the experimental range and in the middle. Although this design may be sub-optimum, if the calibration is in fact linear it represents a good compromise for those analysts worried about the inflexibility of the optimised design. Discussion The results obtained above are applicable to any analytical assay that involves the con- struction of a calibration graph over a finite range (of concentration, for example). The main conclusion is that the assay precision is determined by the experimental design. How- ever, the gain in precision obtained by using an optimised design will become less and less as the number of calibration standards increases.Therefore, optimised design will have the most benefit where only a limited number of standards is possible or desirable. Alternatively, it is possible that the use of an optimised design will give the same information with fewer calibration standards. The gain in information, I , in an experiment is equal to the reduction in uncertainty I = H($o) - H($) .. .. .. . . (12) where H(+,) and H ( p ) are the uncertainties in the assay before and after calibration, respec- tively.’ The gain in information2 after performing the experiment is given approximately by I = -ln(det.Ve) . . .. .. .. .. (13) Assuming a normal, random distribution of errors with constant variance, the information gains calculated for four designs are shown in Table 11.TABLE I1 INFORMATION GAIN CALCULATED FOR FOUR DESIGNS Design X I A .. . . 20, 40, 60, 80 12.99 B I . .. 1, 1, 100, 100 14.58 C . . . . 20, 40, 60, 80, 100 13.90 D .. . . 1, 20, 40, 60, 80, 100 14.64 It can be seen that nearly as much information is gained from four optimally spaced standards as from six uniformally spaced standards. This would represent a saving in both assay time and expense. As we saw, if the variance in the errors is not constant then it is likely that the information gained by using an optimised design will be even more significant. Finally we should re-emphasise that all of our results depend on the calibration beingDecember, 1981 ON ASSAY CALIBRATION 1253 linear.If this is not true then the designs discussed above could be very inefficient. A possible compromise is to place a third of the calibration standards in the centre of the experimental region. However, if the calibration is known to be non-linear and the equation of the calibration graph has been established then it would be possible to design the assay optimally, using the techniques described in this paper, to obtain the maximum information from the calibration. Appendix Weighted Linear Least-squares Analysis If V(yi) is the variance associated with an observation, yi, then the weighted sum of squared deviations, WSS, is given by where .. .. (Al) Normalising the weights equation (Al) becomes WSS' = q y i - co - C1Xi)2Wi' . . .. .. .. (A2) i which is a minimum when co = 7 - c13 . * .. .. .. 1 . where 3 = (Zwi'xi)/n i and The variance of a prediction, x, can be found as follows: which for large values of n approaches V ( x ) m V(y)/c12 . . .. .. * . .. (A6) If V ( y ) = ky andy = x, for example, equation (A6) becomes V ( X ) = kx/c12 . . .. .. .. . . (A7)1254 AARONS and so from equation (8) the sensitivity, xo, can be found as follows: and therefore, xo = 10OK~c,2 * . * . .. .. .. (A8) References 1. 2. 3. 4. 5. 6. 7. Draper, N. R., and Smith, H., “Applied Regression Analysis,” John Wiley, New York, 1966, Chapter Bard, Y., “Nonlinear Parameter Estimation,” Academic Press, New York, 1974, Chapter 10. Davies, 0. L., and Goldsmith, P. L., Editors, “Statistical Methods in Research and Production,” Liteanu, C., and Rica, I., “Statistical Theory and Methodology of Trace Analysis,” Ellis Horwood, Hubaux, A., and Vos, G., Anal. Chem., 1970, 42, 849. Garden, J. S., Mitchell, D. G., and Mills, W. N., Anal. Chem., 1980, 52, 2310. Shannon, C. E., Bell Syst. Tech. J., 1948, 27, 379 and 623. 2. Longmans, Harlow, 1972, p. 205, Chichester, 1980, Chapter 8. Received May 5th, 198 1 Accepted July 16th, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601249

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, December, 1981, Vol. 106, pp. 1255-1267 Plasma Source Mass Spectrometry Using an 1255 Inductively Coupled Plasma and a High Resolution Quadrupole Mass Filter Alan R. Date and Alan L. Gray Institute of Geological Sciences, 64/78 Gray's I n n Road, London, WC1X 8NG Department of Chemistry, University of Surrey, Guildford, G U2 5XH The performance of an instrument using an inductively coupled plasma as an ion source with a high resolution mass filter to analyse the extracted ions is described. A description of the experimental system is given and the spectral characteristics of the technique are discussed. The current per- formance of the technique for simple solution analysis is described, and the future potential for trace element analysis and isotope ratio measurement is assessed.Keywords : Inductively coupled Plasma ; atomic mass spectrometry ; high resolution quadrupole filter ; multi-channel scaler ; trace element determination Conventional methods of multi-element trace analysis such as atomic mass spectrometry and plasma emission spectrometry have great potential in many branches of analytical chemistry. Atomic mass spectrometry is a very sensitive analytical tool, but in a multi-element context, with spark or secondary ion sources, it suffers from several well known disadvantages] which have restricted its routine use. Plasma emission spectrometry is proving invaluable as a multi-element solution technique offering good precision with high sensitivity, but is limited by the inability of modern optical spectrometers to resolve the complex spectra developed, particularly in geochemical and environmental samples.The use of ion lines in plasma emission spectrometry confirms the presence of atomic ions in the tail flames of many plasma sources. The application of a d.c. plasma (capillary arc plasma, CAP) as an ion source for atomic mass spectrometry was pioneered by Grayly2 who demonstrated the potential of the technique. This work also drew attention to the limitations of the d.c. plasma in terms of freedom from inter-element and matrix interferences] and recommended a change to the more useful inductively coupled plasma (ICP).3 The application of the ICP as an ion source has been described by two groups, one at the Ames Laboratory of Iowa State University under the direction of F a ~ s e l , ~ t ~ which was the first to extract ions from an ICP, and another supported by the Institute of Geological Sciences and the European Economic Community at the University of Surrey under the direction of Gray.6 Published work from these two groups, in common with that described by Douglas and French,' using a microwave-induced plasma (MIP), has been limited by the use of small quadrupole filters more suitable for residual gas analysis.These small mass analysers, although simple and inexpensive, are really only capable of adequate resolution for quantitative analysis at mass numbers below 100 a.m.u. Attempts to obtain increased resolution at higher masses result in such severe loss of ion transmission that the over-all sensitivity is reduced.More sophisticated quadrupole analysers have been developed in the last few years for use with gas chromatograph -mass spectrometer systems. These use additional electrode systems to modify the fringe fields at the entrance to the quadrupole rods and enable good ion transmission to be maintained up to high mass numbers. The use of such a high resolution quadrupole mass filter is described in the current work. Experimental System Development The experimental system used in this work is shown schematically in Fig. 1, and a list of the major components appears in Table I. The system may be considered in four sections] as follows : inductively coupled plasma (ICP) ; mass spectrometer ; plasma sampling interface ; and data system.1256 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, vd.106 Mass analyser control RF power supply I l l Fig. 1. Plasma source mass spectrometer. Inductively Coupled Plasma The inductively coupled plasma used in this work is the Plasma-Therm ICP 2500A, with automatic power control and impedance matching. This system is convenient in having a separate torch box and power supply. In order to gain access to the torch, it was necessary to remove the gas-control section from the torch box. The standard torch used in this work was then re-aligned a t 30" to the horizontal, and a new coil manufactured. A new safety shield was made from copper sheet, connected during operation to mass spectrometer ground. This new torch box, with the matching and tuning circuitry, was mounted on an optical bench with adjustments for vertical and horizontal movement.All the work reported here was carried out with the plasma operating at a forward power of 1 kW (reflected power, less than 10 W), with argon flow-rates of 15 1 min-l (coolant), 0.5 1 min-l (plasma) and 1 .O 1 min-l (sample). Although a batch-type ultrasonic nebuliser, with desolvation, was developed during the early stages of the work, the introduction of a high resolution quadrupole filter, with a much greater ion transmission, permitted a change to the more convenient, but less efficient, pneumatic nebuliser system. The original design of the desolvator was retained for use with a conventional Scott-type cloud chamber. This system is shown schematically in Fig. 1 TABLE I COMPONENT DESCRIPTION Inductively coupled plasma, ICP 2500A .. . . Plasma-Therm Tnc., Kresson, N.J., USA Plasma torch, T 1.0 . . .. .. . . . . Plasma-Therm Inc., Kresson, N.J., USA Nebuliser, cross-flow, fixed geometry, 09-790 . . Sampling aperture, molybdenum disc, Type 2 . . Quadrupole mass analyser, VG 12-12F . . . . VG Isotopes Ltd., Winsford, Cheshire Detector, channel electron multiplier, Type 4870 . . Jarrell-Ash Co., U'altham, Ma., USA Agar Aids Ltd., Stansted, Essex Galileo Electro-Optics Corp., Sturbridge, Ma., USA Pre-amplifier, PAD 400 . . .. . . . . Galileo Electro-Optics Corp., Sturbridge, Ma., USA Ratemeter, Type 449 . . .. . . .. . . EG R: G Instruments Ltd., Bracknell, Data system, Canberra Series 80 . . .. . . Canberra Instruments Ltd., Faringdon, Berkshire Oxf ordshireDecember, 1981 HIGH RESOLUTION QUADRUPOLE MASS FILTER TABLE I1 1257 NEBULISER - DESOLVATOR SYSTEM Cloud chamber ..* . .. .. * . Pneumatic nebuliser (cross-flow, fixed geometry) . . Transfer to heater (ball, S13, to socket, 19/26) . . Heater tube (stillhead, recovery bends vertical) . . Transfer to condenser (multiple adaptor, one neck a t 45') . . .. . . I . .. .. Thermometer gland , . .. .. 0 . .. Condenser (jacketed coil) . . * . .. .. Transfer to plasma and drain (socket to cone, with T connection) . . 0 . .. .. * . Drain flask . . .. .. .. .. .. Heating tape . . . I . . .. .. .. Heater control . . .. . . . . .. . . Plasma-Therm Inc., Sc-2 Jarrell-Ash, 09-790 Glassblowing Unit, University of Surrey Exelo, J56/2 ; Quickfit, SH2/22 Exelo, J32/1; Quickfit, MA2/2 Exelo, JS30/2; Quickfit, ST52/13 Exelo, C22 ; Quickfit, CX6/22 Exelo, J39/5; Quickfit, MFl8/2 Exelo, F30/8 ; Quickfit, FE l00/2 Electrothermal, HT351 (240 V, 1 A max.) (Scientific Supplies Co.Ltd.) Domestic dimmer switch (250 W) and details of the nebuliser-desolvator are given in Table 11. The data presented below were obtained using a Jarrell- Ash fixed geometry cross-flow nebuliser, Type 09-790. The natural sample uptake rate, at a carrier gas flow-rate of 1.0 1 min-l, is 2.5 ml min-l. The use of a peristaltic pump for sample introduction to the nebuliser was discontinued when it was found that a pulsating signal interfered with isotope ratio measurements made under the particular recording conditions used in this work. Although provision was made to monitor the temperature of the aerosol leaving the desolvator, no attempt was made in the present work to control desolvation temperature.Mass Spectrometer The prototype equipment described by Gray1v2 was purchased for this work, stripped, cleaned and modified to provide access to the larger ICP torch box. Several improvements were made to the two-stage vacuum system for ease of operation. In addition, a fast counting channel electron multiplier, capable of counting to over 1 MHz without significant loss, was fitted. The channel electron multiplier was mounted just off the axis of the system (Fig. 1) to avoid interference from photons. During the course of this investigation the original small quadrupole filter was replaced by a large, high transmission, high resolution quadrupole filter with its associated circuitry.This quadrupole, the VG 12-12F, is of the type first proposed by Brubaker* in which the main analyser rods are preceded by a set of short rods on the same axes. These pre-filter rods, forming the so-called delayed d.c. ramp, are fed through coupling capacitors from the main rod a.c. supply, but without the d.c. component. This arrangement greatly reduces the fringe field at the end of the main analysing rods so that the entering ions require only a low velocity to traverse the rod system. The residence time of the ions in the quadrupole field is thus much longer than in the simple type of quadrupole, and the filtering process is much more efficient. In particular, the fall in transmission of higher mass ions, a t high resolution, is greatly reduced.The instrument used in this work is designed to operate over a mass range of 0-800 a.m.u., with a resolution (MIAM, 50% peak height) of better than 2M. This implies that at any particular m/x setting the response from the peak at M + 1 will be zero for perfectly tri- angular peaks. This definition of resolution is inappropriate for the simple atomic mass spectra obtained from a plasma source. The most important consideration in this instance is the degree of overlap on, or contribution to, adjacent peaks from the precursors and tails of the peaks, which occur at a few percent. of the peak height. In practice, only the mass range from 0-300 a.m.u. is used and the peak shape is such that the response falls virtually to zero on either side of the maximum, well within unit mass width.In terms of adjacent mass channel interference, the most convenient expression of performance is the abundance sensitivity, the contribution at one mass number from a peak one mass unit either side, expressed as a ratio. With this instrument the abundance sensitivity is typically 10-6 or less. The ion mass transmitted by a quadrupole filter is a function of electrode geometry and the electric fields applied, and is usually set by a d.c. control potential. In this instrument1258 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, VOl. 106 a range of 0-10 V corresponds to the mass range 0-800 a.m.u. The range may be scanned repeatedly by supplying a repetitive d.c. ramp of the required amplitude.The starting mass, scan width and scan rate may be readily set. The instrument therefore lends itself to microprocessor control or integration with a data system. Plasma Sampling Interface The plasma system and mass spectrometer used in this work are slightly modified standard items of equipment. The plasma sampling interface, however, which occupies the centre of the field of view in Fig. 1, represents the main region of uncertainty in this research field. This area is shown in detail in Fig. 2, with a diagrammatic representation of the plasma tail flame under operating conditions. The configuration of plasma torch and sampling cone is that described by Gray and Date,6 and differs from the Ames Laboratory mode1435 in having no skimmer in front of the sampling cone.In addition, the off-axis cone mounting is arranged at 20" in the to the horizontal to furcherkeduce interference from the intense light sour"ce generGed plasma. no r\ Slide valve ounting block (water cooled) erosol dissociation zone LJ Sample carrier gas Fig. 2. Plasma sampling interface. In both instances, however, the sampling aperture itself is a molybdenum disc (2.0 mm diameter, 0.5 mm thickness) with a central countersunk hole, designed originally for beam collimation in electron microscopy. Data reported in this work have been obtained using molybdenum apertures with a diameter (D) of 70 pm, and a hole length approximating to 0.5D. The aperture opens up during use but continues to operate satisfactorily to 80 pm diameter. The aperture is swaged into the tip of a water-cooled copper cone that projects into the plasma tail flame.Ions are sampled from the area of maximum ionisation in the tail flame, but have to traverse a boundary layer that forms over the sampling cone. The nature of this boundary layer is discussed in detail el~ewhere~-~ but its effect on ion sampling is relevant to this work. The delay in ion transmission through the boundary layer results in spectral degradation caused by ion - molecule reactions. Increasing the aperture diameter to induce continuum flow and break through the boundary layer increases the compression of the free electron population of the plasma and raises its energy. This not only produces an intense light source in the mouth of the aperture, but also acts as an additional ion source, increasing the ion-energy spread and degrading spectrometer resolution.Although work is in hand to overcome this problem, the research described here is limited to samples containing boundary-layer gas. The improved performance obtained with the current system is a measure of the future potential of the technique in inorganic trace analysis. Data System Operation Data reported in this work owe much to the application of a multi-channel analyser developed originally for use in y-ray spectrometry. The data system used is the Canberra Series 80, with 8192 channels and a multi-channel scaling (MCS) data acquisition module A "pinch" discharge O C C U ~ S . ~ ~ ~December, 198 7 HIGH RESOLUTION QUADRUPOLE MASS FILTER 1259 capable of counting to 20 MHz.Under normal operating conditions, the data input memory group is pre-set to 1024 channels, a dwell-time per channel of 1 ms and 60 consecutive sweeps, giving a total integration time Gf just over 1 min. All data presented here have been obtained using this operating scheme. The mass spectrometer is set to the desired first mass and mass range, and is synchronously driven by the data system sweep. The channel electron multiplier output is amplified and fed into the data system. The spectrum developed may then be displayed on the visual display unit as it builds up. At the termination of each complete scan, the data may be transferred into one of three other memory groups, recorded on cassette, or read out on to an X - Y plotter or teletype. Several program options are available for data manipulation, including peak-area measure- ment, spectrum overlap (for comparison purposes) and spectrum stripping (for background correction).Simple programs may be written for automatic data acquisition, processing and read-out. The X - Y plotter may be used in one of two modes, the first giving channel point data only, and the second providing a continuous trace. The teletype may be used to record data for all channels, or for restricted regions of interest (ROI) only. In the second instance, two strobe markers are used to set the channel limits for each region of interest in a memory group, and the regions of interest are maintained for that memory group until erased. The integral count between the limits, or the area count (taking a pre-set number of channels on either side to compute background), is printed out on the teletype.This option is used in the preparation of calibration graphs from successive solutions in a calibration series. Spectral Characteristics The basic characteristics of the technique in its current state of development, and the advantages afforded by a high resolution mass filter, are illustrated in Fig. 3, the X - Y plot of a spectrum obtained in just over 1 min from aspirating a solution of cobalt at 1 pg ml-l in 1% VjV nitric acid. The vertical scale represents the peak channel count converted into its equivalent in counts per second, which would be obtained under conditions of single ion monitoring, the technique applied in some of the earlier work in this The horizontal scale covers a mass range of 0-96 a.m.u., and the single isotope of cobalt a t 59 a.m.u.is clearly visible at a rate of 733517 counts sec-l, and well separated from the groups of back- ground peaks. Most of the elements studied in this work were observed only as monatomic singly charged ions (Mf). Some elements were observed as a distribution of metal and oxide (MO:) ions (e.g., Yf, YO+; UO$, UO+), while alkaline earth elements tended to form combinations of metal, oxide and hydroxide (MOH+) ions (e.g., Sr+, SrO+, SrOH+). Only barium and strontium, with low second ionisation potentials, were observed as doubly charged ions (e.g., Ba2+). 1 19 41 Fig. 3. Spectrum of cobalt at 1 pg ml-l in 1% nitric acid showing major background ions.1260 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, F/‘ol.106 The configuration of plasma torch, sampling cone, ion lens and detector described above results in the almost complete removal of stray background caused by scattered ions and photons. The general background level is therefore very low and, in the absence of true spectral background, is typically between 0 and 10 counts s-l. This compares favourably with the background levels of 30-100 counts s-l and 200-2000 counts s-l reported by Houk et aZ.4 and Douglas and French,’ respectively. There are several major background peaks between 17 and 41 a.m.u., and at 80 and 81 a.m.u. At least The peaks numbered in Fig. 3 are provisionally identified in Table 111. TABLE 111 MAJOR BACKGROUND IONS, ORIGIN AND COINCIDENT ISOTOPES Mass number Probable identity Origin* Coincident isotope Alternative isotope 17 OH+ BL 1 7 0 (0.04%) 1 6 0 (99.7%) 18 OH,+ BL ‘ 8 0 (0.2%) 1 6 0 (99.7%) 19 OH3+ BL I9F (1000/0) None 29 N,H+ BL 29Si (4.7%) 28Si (92.2%) 30 NO+ BL 30Si (3.1%) 2sSi (92.2%) BL 32S (95%) .34S (4.2%) BL 33s (0.8%) 34S (4.2%) 32 33 40 Ar+ PA 40Ca (97%) 44Ca (2.1%) 41 ArH+ BL 41K (6.9%) 39K (93%) 80 Ar.Ar+ EX *OSe (50%) 78Se (23%) 81 Ar.ArH+ EX 81Br (49%) 79Br (50%) 37 OH,+. (H,O) AT 37ci ( 2 4 ~ ~ ) 35c1 (75%) %L+ * BL AT PA EX = condensation reaction in gas expansion through aperture = boundary layer ion - molecule reaction ; = catalysed attachment of H,O molecule t o ion; = primary ionisation in plasma; and four independent production processes are involved, but ion - molecule reactions occurring in the cool boundary layer predominate.In terms of line overlap, however, the background peak interferences are not as serious as they first appear, fluorine being the only element that cannot be deterrnined. In view of its very high first ionisation potential (17.42 eV), fluorine is unlikely to be detected as a positive ion a t ICP temperatures in the presence of an excess of argon (ionisation potential, Vi = 15.76 eV). Excellent resolution is achieved under these conditions, with the signal falling to the base line between the peaks at 40 (Ar+) and 41 a.m.u. (ArH+). Interference between adjacent mass numbers is reduced to negligible proportions (as compared with atomic-emission spectrometry), and, as noted above, an abundance sensitivity of or less is often achieved.Although it would appear that the spectrum illustrated in Fig. 3 is relatively free from 21 29 20 z., m/z Fig. 4. Scale expansion of cobalt spectrum, mass range 0-48.December, 1981 HIGH RESOLUTION QUADRUPOLE MASS FILTER 1261 interference between 41 and 80 a.m.u., and above 81 a.m.u. the reduction of random back- ground, allied with the introduction of a high transmission mass filter, allows the true spectral background to be assessed. The spectrum illustrated here has been subjected to data system vertical scale expansion ( x 64), and split into two groups of 512 channels. The X - Y plots for the two spectra are shown in Figs. 4 and 5. Many small background peaks are revealed Fig. 5. Scale expansion of cobalt spectrum, mass range 49-96. by this technique.Although some are due to sample contamination (e.g., sodium-23, copper-63 and copper-65), it is possible that a significant proportion of the peaks are attri- butable to ion - molecule reactions in the boundary layer and will be absent when true bulk plasma sampling can be achieved. It is interesting to note that the spectrum is free from interferences caused by ionisation of residual gases characteristic of in V ~ C U O ionisation sources, and that significant peaks a t fractional mass numbers are absent. The very low non-spectral background is illustrated in Table IV, the channel count record for part of a spectrum obtained from a solution containing 20 pg ml-l each of chromium, cobalt, copper and zinc. The copper-63 peak occupies seven channels with the peak centre a t channel 674, equivalent to a rate of 673633 counts s-l. The background drops to zero between the peaks, and the position occupied by mass number 62 has an integral count of zero.It must be noted that the sensitivity obtained under these conditions is inferior to that obtained with the spectrum of cobalt described above. Current Performance Under Boundary Layer Sampling Conditions The current performance of the technique under boundary layer sampling conditions may be illustrated with three examples. In the first example, ten blank spectra (1% VjV nitric acid) were recorded, followed by a calibration series of solutions containing chromium, cobalt, copper and zinc. In practice, each solution in the calibration series was aspirated for a time equivalent to one complete scan before a spectrum was recorded, so that the system could reach equilibrium.Allowing for time taken in sample changing, the total analysis time per sample was less than 2.5min, a very short time compared with conventional methods of atomic mass spectrometry. The mass filter was set to cover a mass range of 0-100 a.m.u. in 1024 analyser channels. The total spectrum recorded is similar to that shown in Fig. 3, except that the third memory group of 256 channels covering the mass range approximating to 50-75 a.m.u. conveniently includes all the isotopes of the four analytes. The X - Y plot of this section of the spectrum obtained from a solution containing each analyte a t the 0.1 pg ml-l level is shown in Fig. 6.In this instance the X - Y plotter was used to record channel point data only, and the lines between the peak points have been added manually, for clarification. No vertical scale is Each spectrum was recorded in just over 1 min.1262 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, VoZ. 106 shown, but the cobalt peak channel represents 1192 counts. All of the isotopes are clearly visible a t this level of concentration, together with three background peaks at 56, 57 and 58 a.m.u. In order to plot calibration graphs using these data, a region of interest (ROI) is set around each analyte peak. The series of spectra stored on cassette are sequentially transferred into the data system memory and the integral count for each ROI is recorded on the teletype.TABLE IV CHANNEL RECORD FOR SPECTRUM OF CHROMIUM, COBALT, COPPER AND ZINC DuTell time = 1 ms; 60 sweeps; 20 p.p.m. of chromium, cobalt, copper and zinc. Channel 528 536 544 552 560 568 576 584 592 600 608 616 624 632 640 648 656 661 672 680 688 696 704 712 720 728 736 744 752 Data A I \ 1 0 0 355 0 0 1412 0 0 0 66 0 0 1812 7 0 0 0 11 175 0 0 17 860 203 0 0 3 761 0 0 2 0 0 0 14 677 0 0 66 1 0 0 0 79 0 0 1 13 1 0 0 32 444 0 0 1616 2 338 1 0 2 320 19 0 0 Isotope 52Cr “CU 59c0 642n 0 0 0 38 957 0 0 2 1 0 0 21 3 0 0 8 0 0 0 40418 37 0 2 3 803 2 0 74 150 0 0 258 0 0 33 051 92 0 0 3 1 2 0 1 9 25 0 0 2 0 0 0 533 2 668 1 0 3 472 42 1603 0 0 932 2 322 0 0 0 22 0 0 19 1225 0 0 1 0 0 35 5 887 7 0 198 358 2 0 0 0 176 111 1 59 1 0 Summary Peak channel Count 554 38 987 630 54 636 674 40418 685 6 873 2 285 0 0 0 5 166 3 0 0 99 0 0 4 29 233 1 0 1 0 0 0 6 873 344 0 0 59 1 2 1 5 176 1 350 0 0 0 3 473 56 0 0 115 0 0 1 54 636 0 0 0 0 1 0 1098 9279 0 0 460 553 0 1 224 0 Count rate/ counts s-l 649 783 910600 673 633 I14 550 0 0 3 0 58 839 0 0 35 9 0 1 45 043 0 0 0 0 98 0 2 17215 5 1 21 2 775 0 1 96 4 The calibration graphs are plotted manually.In this example, it was found convenient to use 9 channels in each ROI, leaving a single channel (often with a count of zero) between adjacent peaks. With a dwell time of 1 ms per channel, and 60 consecutive sweeps, the total time allocated to each isotope was 0.54 s. The calibration graphs illustrated in Fig. 7 should be considered in this light. The calibration graphs for chromium-52 (abundance, 83.76%) and cobalt-59 (abundance, 100%) are almost linear down to 0.001 pg ml-l.The calibration for copper-63 (abundance 69.09”/b) is subject to a high blank reading, and is erratic below 0.002 pg ml-l. The graph for zinc-64 (abundance, 48.89%) begins to flatten off a t 0.01 pg ml-l, reflecting a lower sensitivity for zinc and the tendency for volatile elements to show a serious memory effect in the desolvation system.December, 1981 HIGH RESOLUTION QUADRUPOLE MASS FILTER 1263 The means (Z) of the integral counts obtained for ten runs of the blank solution and the standard deviations (a) may be summarised as follows: chromium-52, Z = 5.1, a = 2.2; cobalt-59, Z = 10.4, a = 3.0; copper-63, Z = 102.2, a = 10.5; zinc-64, Z = 44.5, (T = 3.9. Detection limits (30 blank), under these far from ideal conditions, are 0.4 ng ml-l (chromium), 0.2 ng ml-1 (cobalt), 2.0 ng rnl-l (copper) and 3.0 ng ml-l (zinc). 52Cr "CU "cu Fig.6. Spectrum (256 channels) of chromium, cobalt, copper and zinc at 0.1 pg ml-l in 1% nitric acid (cobalt-59 peak channel, 1 192 counts). The calibration graphs*_for-alL-four analytes begin to flatten off above 1.0 pg rnl-l. This effect may result partly from loss of counts caused by a change of the working point on the counter plateau at higher concentrations. The principal reason for this behaviour, however, is gradual blocking of the sampling aperture caused by condensation in the cool boundary layer. 1 o6 1 o5 +- c 5 104 - I CT) S Y a +- .- 103 1 o2 10' 45" ,/' I 0.001 0.01 0.1 1 .o 10.0 Element total concentrationipg ml- Calibration graphs for : A, chromium-52 ; 0, cobalt-59; 0, copper-63; and 0, zinc-64.Fig. 7 .1264 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, Vd. 106 In the second example, the mass filter was set to cover a smaller mass range approximating to 105-117 a.m.u. Ten blank spectra were again recorded, followed by a calibration series for silver and cadmium. The X - Y plot obtained from aspirating a solution containing each analyte at 0.01 pg ml-l in 1% VlV nitric acid is shown in Fig. 8. In this instance a plot of channel point data is superimposed over a continuous plot, for clarification. All of the isotopes of silver and cadmium are clearly visible a t this level, with excellent resolution. I I "'Cd 1 I '"Cd 1 . I m/z Fig.8. Spectrum of silver and cadmium a t 0.01 pg ml-l in 1% nitric acid (silver-107 peak channel, 180 counts). No vertical scale is shown, but the peak channel represents 180 counts. The two silver isotopes have total integrated peak counts (50 channels) of 3127 (silver-107) and 2951 (silver-109). The isotopic abundance for silver-107 is therefore 51 .asyo, which compares well with the accepted value of 51.82%. The error is within that expected from counting statistics. For cadmium, the cadmium-108 peak (abundance, 0.88%) is clearly larger than the cadmium-106 peak (abundance, 1.22%). The presence of a background peak at mix 108, similar to that at m/x 105, is suspected. The existence of silver-108 hydride cannot be excluded, although there should then be more evidence of the corresponding silver-1 10 hydride interference on cadmium-1 10.The determination of cadmium isotope ratios is clearly not feasible under these conditions. In order to plot calibration graphs using these data, each region of interest was allocated 50 channels, corresponding to a total analysis time per isotope of 3 s. The calibration graphs obtained for silver-107 (abundance 51 .82y0), cadmium-1 14 (abundance, 28.86%), cadmium-1 16 (abundance, 7.58%) and cadmium-106 (abundance, 1.22%) are shown in Fig. 9. The calibration graph for silver-107 is linear down to 0.001 pg rnl-l, and the detection limit (3a blank reading) for silver using this isotope is 0.06 ng ml-l. The calibration graphs for cadmium tend to Batten off at low concentrations, reflecting a memory effect in the desol- vation system similar to that exhibited by zinc.The detection limit for cadmium, using cadmium-114, is 0.2 ng ml-l. The drop in signal common to both analytes above 1.0 pg ml-l is again the result of partial blocking of the sampling aperture, although there is some evidence of counting losses in that the graphs are straighter a t low count rates. In the third example, the mass spectrometer was set to cover a range approximating to 204-209 a.m.u. Only one blank solution spectrum (1% V/V nitric acid) was recorded, followed by a calibration series for lead (prepared from Johnson Matthey Specpure grade lead oxide). The X - Y plot obtained for a solution containing lead at 0.001 pg ml-1 isDecember, 1981 105 c 13 8 5 104 - 6) c .- Y m 2 103,- HIGH RESOLUTION QUADRUPOLE MASS FILTER - - ,, 1265 1 06 P’ q‘ /A/’ 0 (I ‘ 0 0 0 D O 0 .0 0 0 1 o2 If:= O o 0 ~ . 0 o 0 , , , 10’ 0.001 0.01 0.1 1.0 10.0 Element total concentration/yg mi-’ Fig. 9. Calibration graphs for silver and cadmium. A, Silver-107 (51.8%); B, cadmium-114 (28.9%) ; C , cadmium-116 (7.58%); and D, cadmium-106 (1.22%). shown in Fig. 10. In this example, a plot of channel point data is again superimposed over a continuous plot, for clarification. The peak lead-208 channel represents 27 counts, and the vertical scale deflection is such that plateaux may be seen at levels of 0, 1 and 2 counts, a good illustration of the low background level achieved in this work. ’OaPb f . . . . . . ... . . . ry.. ...... 206Pb ’O’Pb .. , ..s .. . :: : 1:.. ...... . -.. .. . . . ...... . -. , .. (. *.. .... . .5..-*..** . . * . . I .-. .. .-a .. .. ’C . - . . . . . ........ ... .... . . . . - . - * 4 . b . a . c .-. I--.: ...a . . . . .. . - .. .... .-L. 204Pb * ... . . . . . . . . . . . . . . . . ..- . . .-* .. ..., .. .C . .. .... .. .. .. 0.v- .... - .. --. - - .- 0.. .b* ..-.. --I--- .- - -. -.. -4- ..rC .c - .. - .. .. ...- *. --- . - ’II --. ,. .- .-- . .- -- ~ ~~ m/z Fig. 10. Spectrum of lead a t 0.001 p g ml-l in 1% nitric acid (lead-208 peak channel, 27 counts). Calibration graphs, prepared by assigning 150 channels to each region of interest, may be seen in Fig. 11. The graphs for lead-206, lead-207 and lead-208 are linear down to 0.001 pg ml-1, and that for lead-204 shows a progressive decrease in signal to the same level.The single blank spectrum recorded gave a count of 378 at the lead-208 position, suggesting a detection limit below 1 ng ml-l. The fall in signal at high concentrations is again apparent in this example.1266 DATE AND GRAY: PLASMA SOURCE MS USING AN ICP AND A Analyst, V d . 106 No serious attempt has been made to assess the technique in terrns of isotope ratio deter- mination. It is interesting to note, however, that the isotope abundances calculated for the centre point of the calibration (0.1 pgml-l) are very close to frequently accepted values for natural lead. The calculated values, with accepted values in parentheses, are lead-204 1.48% (1.48y0), lead-206 24.2% (23.6y0), lead-207 22.1% (22.6%) and lead-208 52.2% (52.3%). I 106 - c.’ 5 105 - 0 A A 0.001 0.01 0.1 1.0 10.0 Total lead concentrationiug ml-’ Fig.11. Calibration graphs for lead isotopes. x , Lead-208 (52.23%) ; 0, lead-206 (24.21%) ; 0, lead-207 (22.08%) ; and A, lead-204 (1.48%). Conclusions This application of a high resolution quadrupole mass analyser to plasma source mass spectrometry illustrates the true spectral characteristics of the technique under boundary layer sampling conditions, and provides an insight into the future potential of the method for multi-element trace analysis. Although still in its infancy, the technique described in this work has reached a stage .where reproducible plasma sampling may be extended over several hours. This period of time, although short by contrast with conventional methods of atomic mass spectrometry, is significant in terrns of solution analysis in that sample changing takes only a matter of minutes. The calibration point data presented in this paper represent single determinations for each solution, with no correction for background signal.The technique is shown to have high selectivity and high sensitivity. Detection limits are better than 1 ng ml-l for a wide range of elements. Recent data, in some instances obtained under compromise conditions (e.g., chromium, cobalt, copper and zinc in the same spectrum), are shown in Table V. In its current state of development, the technique is limited in two respects. Ionisation suppression is of the type described by Houk et a1.,4 and is a function of the cool boundary layer “plasma” conditions.The signal loss noted at high analyte concentrations is caused primarily by salt condensation at the sampling aperture, another feature of boundary layer sampling. Further development work is in hand to limit or remove the boundary layer so that true bulk plasma sampling may be achieved. The increase in signal expected under such ideal conditions should, with the consequent reduction in spectral background, lead to a viable method of analysis combining the advantages of plasma-emission spectrometry andDecember, 1981 Element Magnesium . . Aluminium . . Vanadium. . . . Chromium. . .. Manganese .. Iron .. . . Cobalt . . .. Copper . . .. Ion 2%Mg 2 7 ~ 1 6’VO 52Cr 55Mn 56Fe 63Cu 5 9 ~ 0 HIGH RESOLUTION QUADRUPOLE MASS FILTER TABLE V DETECTION LIMITS Detection limit1 pg ml-l 0.0002 0.0003 0.0003 0.000 1 0.00006 0.000 5 0.000 1 0.0003 Element Zinc . I .. Rubidium . . . . Silver . . .. Cadmium .. .. Barium . . .. Mercury . . .. Lead .. .. Uranium . . .. Ion G4Zn B5Rb 114Cd 138Ba 202Hg 208Pb 107Ag 2 7 0 ~ 0 , 1267 Detection limit\ pg ml-’ 0.002 0.0002 0.00006 0.0002 0.0002 0.002 0.0002 0.000 05 atomic mass spectrometry, but free from the disadvantages of the complex spectra observed in atomic-emission spectrometry, and the slow speed of analysis characteristic of atomic mass spectrometry. This work is supported by the Institute of Geological Sciences and the European Com- munities Research and Development Programme on Uranium Exploration and Extraction (Contract No. 012-79-4 EXU UK). The authors thank Mr. P. J. Moore and Miss E. Waine for critically reading the manuscript. The paper is published with the approval of the Director, Institute of Geological Sciences (N.E.R.C.). References 1. 2. 3. 4. 5. 6. 7. 8. Gray, A. L., PYOC. SOC. .4nal. Clrzenz., 1974, 11, 182. Gray, A. L., Anal-vst, 1975, 100, 289. Gray, ,4. L., in Price, D., and Todd, J. F. J., Editors, “Dynamic Mass Spectrometry,” Volume 5, Houk, R. S., Fassel, V. A4., Flesch, G. D., Svec, H. J., Gray, A. L., and Taylor, C. E., Anal. Chem., Houk, R. S . , Fassel, V. A., and Svec, H. J., in Price, D., and Todd, J. F. J., Editors, “Dynamic Mass Gray, A. L., and Date, A. R., in Price, D., and Todd, J. F. J., Editors, “Dynamic Mass Spectro- Douglas, D. J., and French, J. B., Anal. Chem., 1981, 53, 37. Brubaker, W. M., Adv. Mass Spectrom., 1968, 4, 293. Heyden, London, 1978, pp. 106-113. 1980, 52, 2283. Spectrometry,” Volume 6 , Heyden, London, 1981. metry,” Volume 6, Heyden, London, 1981. Received May I l l h , 1981 Accepted August 20th, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601255

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

1268 Analyst, December, 1981, Vol. 106, pp. 1268-1274 Extractive Spectrophotometric Determination of Niobium in Pyrochlore-bearing Rocks with 5,7=Dichloroquinolin=8=oI A. Sanz-Medel and M. E. Diaz Garcia Department of Analytical Chemistry, Faculty of Sciences, University of Oviedo, Oviedo, Spain A spectrophotometric method for the determination of trace amounts of niobium(V) based on its extraction into chloroform with 5,7-dichloroquinolin- 8-01 from a hydrochloric acid medium has been developed. The maximum ab- sorbance of the extracted species occurs at 400 nm, the molar absorptivity being (1.28 f. 0.02) x lo4 1 mo1-I em-’. Beer’s law is obeyed over the range 10-80 pg of niobium(V) extracted. The relative standard deviation is 0.92% and the recovery of niobium is 99.6 f.0.5%. The method is particularly suitable for the determination of niobium in ores and rocks, as commonly interfering elements present in niobium ores did not interfere in this deter- mination. Results of the successful analysis of standard and synthetic nio- bium ores with niobium contents ranging from 0.1 to 1 .Oyo are given. A useful modification of Faye’s method for dissolution of niobium ores is reported. The special characteristics and importance of drying the organic phase in the spectrophotometric measurements are discussed. Keywords : Niobium determination ; ore dissolution ; ore analysis ; extraction spectrophotornetry The complex composition of most niobium ores and especially the special features of the wet chemistry of this metal when mixed with elements such as tantalum, titanium, zirconium and uranium (which usually accompany niobium in rocks and ores) are responsible for the involved and time-consuming wet methods for the determination of niobium in natural materials.Moreover, published methods (classical and spectrophotometric) show large discrepancies in the resu1ts.l This lack of accuracy and precision is common in niobium ore analysis,2 and it follows that new, improved analytical methods for niobium must be developed. Extractive spec tropho t omet ric determinations are most promising and organic extract ion reagents proposed for the direct spectrophotometric determination of niobium(V) include neutral chelate-forming reagents, e.g., quinolin-8-01 (oxine) ,3,4 and charged chelate-forming dyestuffs, e.g., 4-(2-pyridylazo)resorcino15 or bromopyrogallol red (BPR) ,6 in connection with an “onium” type counter-ion.Previous work on the behaviour of several dihalogen derivatives of quinolin-8-oP showed that the 5,7-dichloro- and 5,7-dibromo-derivatives could be of great value for the direct extraction spectrophotometric determination of niobium(V) . Thus, 5,7-dibromoquinolin-8-01 has been proposed for the determination of niobium in steels and al10y.s.~ A careful strdy of the extraction and spectral characteristics of the complexes of many cations with the above two derivatives3p7 r8 indicated that the use of the 5,7-dichloro-derivative could provide a superior method for the determination of niobium in ores because typical elements associated with niobium in rocks and ores, e.g., iron, manganese, titanium, tantalum and uranium, should not interfere at reasonable concentration levels.Therefore, a detailed study of 5,7-dichloroquinolin-8-01 as spectrophotometric reagent for niobium(V) was undertaken and, as a result, a new method for niobium ore analysis has been developed. Concerning the sample dissolution process, it is worth noting that the method proposed was applied to Canadian niobium ores of the pyrochlore type (in a matrix of calcite) and two dis- solution techniques were assayed : the classical potassium hydrogen sulphate fusions and Faye’s method using hydrochloric - hydrofluoric - orthophosphoric acid mixtures.1° The former technique proved to be unreliable, but Faye’s method was satisfactory. However, a signifi- cant amount of an unattacked black residue of sulphides was obtained in the last technique.Therefore, a modification of Faye’s method is proposed here, consisting in the addition of con-SANS-MEDEL AND D ~ A Z GARC~A 1269 centrated hydrochloric acid and hydrogen peroxide to the paste (obtained by evaporating until fumes of orthophosphoric acid were obtained) before final dissolution of the sample in 1 M tartaric acid. The advantages obtained with this modification were as follows: (a) virtually total dissolu- tion of the sample, which avoids possible niobium losses without introducing an excessive content of alkaline salts (as occurs when using a fusion technique); (b) the hydrofluoric acid, which interferes in most niobium determinations, is more easily removed, as the hydrochloric acid - hydrogen peroxide mixture facilitates its evolution from the orthophosphoric acid paste ; and (c) lower temperatures can be used for elimination of hydrofluoric acid and thus PTFE beakers were employed in our work (this eliminates the need for expensive platinum ware).Experimental Apparatus recording of the spectra and 1-cm silica cells, was used. Spectrophotometer. A Perkin-Elmer, Model 124, spectrophotometer with automatic Mechanical shaker. Heron. p H meter. Separating funnels, 100 ml. All of the reagents and acids used were of analytical-reagent grade. Niobium( V ) stock solution, 100 pg ml-l. Niobium( V ) standard solutions. 5,7-Dichloroquinolin-8-01 solution, 1 yo m/V in chloroform. Hydrochloric acid, concentrated and 5.2 M.Orthophosphoric acid, concentrated. Hydrogen peroxide, 30% m/V. Hydrojuoric acid, 48% m/V. Tartaric acid solutions, 1 M and 4% mlV. Metrohm, Model E-516, with a glass electrode and a saturated calomel reference electrode. Reagents Prepared as described elsewhere.ll All standard niobium solutions were prepared freshly by appropriate dilution of the stock solution with 2% tartaric acid. Procedure Dissolution of niobium ores Place in a 250-ml PTFE vessel an appropriate amount of finely ground niobium ore (0.1- 2.0 g, depending on the niobium content of the sample) and add 10 ml of concentrated hydro- chloric acid, 15 ml of 48% m/V hydrofluoric acid and 2 ml of concentrated orthophosphoric acid. Gently heat the mixture at about 300 “C on a hot-plate until evaporation results in a paste with the orthophosphoric acid, and smoky fumes of the orthophosphoric acid are evolved.Allow the vessel to cool, then add 5 ml of concentrated hydrochloric acid and 1 ml of hydrogen peroxide. Evaporate carefully until a phosphoric paste is again obtained. Repeat the above addition once more if a significant amount of black residue is still un- attacked. Cool the solution and dilute to 100 ml with 1 M tartaric acid in a calibrated flask. If necessary, filter off any residue through Albet 242 filter-paper. For maximum accuracy, it is advisable to prepare simultaneously a “blank” ore consisting, in our case, in containing as above 0.1-2.0 g of calcium carbonate and amounts of titanium and iron roughly equivalent to the content of these elements in the niobium ore.General spectrophotometric procedure Pipette x ml of the sample solution (containing 15-65 pg of niobium) into a separating funnel and add 10-x ml of 4% tartaric acid solution and 10 ml of 5.2 M hydrochloric acid. Then add 10 ml of 1% m/V 5,7-dichloroquinolin-8-01 solution in chloroform. Extract the niobium by constant mechanical shaking for about 30 min and allow the phases to separate. Remove any moisture in the organic layer by pouring it on to about 1 g of anhydrous sodium sulphate and measure its absorbance at 400 nm against a blank reagent [prepared by extracting all of the reagents in the absence of niobium(V)]. Dissolve the cooled paste by warming it in about 50 ml of 1 M tartaric acid.1270 SANS-MEDEL AND DIAZ .GARCIA SPECTROPHOTOMETRIC Analyst, VOZ.106 Prepare a calibration graph by taking independent portions of dilute standard niobium In the ore analysis, extract 2 ml of the “blank” ore solution as the blank reagent (absorbance solution and extracting them by the procedure described. = 0.05 & 0.01 against chloroform). Results and Discussion Spectral Characteristics The absorption spectrum of the niobium(V) - 5,7-dichloroquinolin-8-01 complex extracted by following the above general procedure exhibits a maximum at 400 nm when measured against a similar blank, as shown in Fig. 1. 0.7 0.6 0.5 0.4 e a 2 0.3 0.2 0.1 Fig. 1. Absorption spectra of: A, reagent blank extracted into CHCI,, measured against CHCI, as reference; B, Nb(V) - 5,7-dichloro- quinolin-8-01 complex extracted into CHC1, measured against CHCI, as reference; C, Nb(V) - 5,7-dichloroquinolin-8-01 complex extracted into CHC1, measured against reagent blank. Concen- tration of Nb(V) in the aqueous phase, 2.14 x 10-5 M.390420 450 480 51 0 Wavelengt h/n m Effect of the Acidity of the Aqueous Phase A series of several solutions, each containing 56 pg of niobium(V), was treated with 47” tartaric acid solution to ensure a final concentration of 2y0 in this acid and enough acid (hydrochloric acid) or base (ammonia solution) to fall within the pH range 0.5-11. Then the extraction was carried out by shaking for 20 min and the absorbance of the organic extracts was measured by the procedure mentioned above; pH measurements were made on the aqueous layer after extraction. Two more analogous series were carried out, increasing the mechanical shaking to 8 and 28 h, respectively, in order to ascertain the extraction equilibrium conditions. The results are given in Fig.2(a), which shows the kinetic character of the initial minimum extraction observed (different extraction rate, depending on the pH) and the possibility of extraction of niobium(V) at acidities higher than tested. Thus, the influence of increasing amounts of three mineral acids (hydrochloric, sulphuric and perchloric acids) was studied by the general procedure (changing the nature and concentration of the mineral acid added). The results are shown in Fig. 2(b), which indicate that a 2-3 M hydrochloric acid medium ensures a higher degree of extraction and then is adequate for the spectrophotometric determination of niobium(V).Absorption spectra were recorded throughout and it was demonstrated that their shape and the maximum absorption were identical for any of the three acids studied. Therefore, the extracted species is probably the same in all instances. According to Vaezi-Nasr et aZ.,12 the extraction of niobium(V) with 5,7-dichloroquinolin-8-01 from 3 M hydrochloric acid proceeds according to the extraction equilibrium (assuming C1- ions are absent from the chelates) : NbO(OH),Cl,- + 3C1,0zH+ NbO(C1,OZ), + 2C1- + Hf + 2H20 Our own experiments (precipitation, purification and elemental analysis of the niobium - dichloroquinolin-8-01 chelate) confirm the formula NbO(Cl,O,) ,.December, 1981 DETERMINATION OF NB IN PYROCHLORE-BEARING ROCKS 1271 0.8 0.6 al C m n 0.4 z 0.2 (a) 1 2 3 4 5 6 7 8 9 1 0 1 1 DH (bf 5 4 3 2 1 Concentration of acid/M Fig.2. (a) Rate of extraction and pH effect on the extraction of the Nb(V) complex: A, 20 min of mechanical shaking; A, 8 h of mechanical shaking; 0, 28 h of mechanical shaking. Concentration of Nb(V) in the aqueous phase, 2.49 x M . Wavelength, 400 nm. (b) Effect of the acidity on the extraction of the Nb(V) complex: 0, from HC10, medium; 0, from H,SO, medium; x , from HC1 medium. M. Concentration of Nb(V) in the aqueous phase, 2.99 x Reagent Blank Extraction One of the main problems initially encountered was the lack of reproducibility and high value (absorbance = 0.22-0.32) of blanks when drying the organic phase by filtering it through different types of filter-paper (or drying molecular sieves).This fact was not observed in the extraction of niobium(Vi) from sulphuric acid medium with 5,7-dichloroquinolin-8-o13 and 5,7- dibromoquinolin-8-o17 and it therefore appears to be related to the use of hydrochlochloric acid in the extraction. On the other hand, Duplessis et a1.l3 claimed that a charge-transfer ion pair of the type Cl,O,H,+ C1- is extracted into chloroform when using 5,6-dichloroquinolin-8-ol and hydrochloric acid in the extraction. This complex exhibits an absorption band in chloroform at A,,,. = 400 nm, which disappears when using lithium chloride (0.1 M) together with the hydrochloric acid.13 In our case, however, addition of lithium chloride did not affect the absorbance values of the extracted reagent blank.Only when the chloroform layer was dried by drawing it off into a screw-capped tube containing 1 g of anhydrous sodium sulphate was the problem overcome (the absorbance decreased to a reproducible value of 0.05). This fact may be explained in terms of an anionic interchange between chloride and sulphate ions ' according to the reaction and the possible sulphate ion pair formed should not absorb in the 400-nm region. Rate of Extraction and Stability of Colour The low rate of extraction observed at certain pHs [Fig. 2(a) J forced us to study the influ- ence of the shaking period and the order of addition of reagents on the extraction of 50 pg of niobium(V) by the general procedure. The results indicate that equilibrium is achieved after 20 min of mechanical shaking (30 min were used in subsequent work) and that with this shak- ing period the order of addition of the reagents is unimportant.The colour produced in the chloroform layer remained constant for at least 48 h when pro- tected from direct exposure to sunlight. Effect of Reagent Concentration The effects on niobium(V) extraction of the concentration of 5,7-dichloroquinolin-8-01 in chloroform were studied by extracting 56 pg of niobium(V) with 10 ml of chloroform contain- ing increasing amounts of the reagent and following the general spectrophotometric procedure. The corresponding absorbances obtained are plotted in Fig. 3 against percentage of the reagent1272 SANS-MEDEL AND D ~ A Z GARC~A : SPECTROPHOTOMETRIC Analyst, VoZ. 106 in chloroform.As can be seen, the use of a 0.75% solution of reagent is sufficient to obtain a virtually constant absorbance. In order to ensure an adequate excess, a 1% concentration was eventually selected. 0.8 0.6 S m 0.4 s a 0.2 D 0.25 0.50 0.75 1.0 Sat. Concentration of 5,7-dichloroquinolin-8-ol, % m/V Fig. 3. Effect of 5,7-dichloroquinolin-8-01 concentration on the extraction of the Nb(V) complex. Concentration of Nb(V) in the aqueous phase, 2.99 x 1 0 - 5 ~ . Wavelength, 400 nm. Calibration Graph : Sensitivity and Precision The sensitivity of the determination, expressed in terms of molar absorptivity, was (1.28 & 0.02) x lo4 1 mol-l cm-1. The precision for ten replicate determinations of 40 pg of standard niobium(V) was 0.92%. The extent of extraction of niobium(V) under the conditions of the recommended procedure was determined by analysing the niobium(V) extracted into the organic phase with diphenyl- glyoxal bis(2-hydroxybenzoy1)hydrazone (BSHB)14 and was 99.6 j, 0.5%.Beer's law is obeyed over the range 10-80 pg of total niobium(V) extracted. Interference Studies The effect of various foreign metals on the determination of niobium(V) (special attention was paid to those elements most frequently associated with niobium in ores containing rare earths) is shown in Table I. Table I1 shows the influence of some common niobium(V) masking anions. TABLE I EFFECT OF FOREIGN IONS ON THE DETERMINATION OF 40 pg OF NIOBIUM(V) WITH 5,~-DICHLOROQUINOLIN-8-OL Foreign ion Ca, Al, Bi(III), Si(IV), Th(IV), Ce(III), Ni(I1) . . .. .. Cr(II1) .. .. .. .. .. B(II1) . . .. .. .. .. Fe(II1) . . .. .. .. .. CO(I1) . . .. Sb(III), Ti(IV), La(II1) . . .. Ta(V) .. .. .. .. .. V(V) .. .. .. .. . * Sn(1V) .. .. .. .. .. Zr(1V) . . .. .. .. . . W(VL) . . .. .. .. .. Mo(V1) . . .. .. .. .. Mg, Pb(II), CuiiI), MniII), UiVI), Tolerance limit*/pg per 40 pg of Nb(V) Without masking With masking A 7 -7 Masking agent agent agent 4 000 2 000 1 500 1200 1000 Hydrazine (0.1 g 1-l) Absence of C1- ions EDTA (0.4 g 1-') F- (2 g 1-l) F- (3 g 1-l) 4 00 80 4 000 4000 t 800 40 t 40 t * The tolerance given corresponds to the concentration level at which the inteferent causes an error in the t Masking and demasking procedures are given in the text. absorbance of not more than &2%.December, 1981 DETERMINATION OF NB IN PYROCHLORE-BEARING ROCKS TABLE I1 TOLERANCE LIMITS FOR COMMON NIOBIUM-MASKING ANIONS 1273 Tolerance limit*/mg per 40 pg of Nb(V) Without masking With masking r A > Masking anion Masking agent agent agent Tartatrate .. .. .. .. 280 Phosphate . . .. .. .. 60.0 EDTA .. .. .. .. .. 40.0 Oxalate . . .. .. .. .. 2.0 F- . . .. .. * . . . Boric acid (17.1 g 1-l) 0.8 60.0 * The tolerance given corresponds to the concentration level a t which the interferent causes an error in the absorbance of not more than f 2%. As can be seen, the proposed method is in general very selective. In particular, it is very suitable for the analysis of niobium ores and rocks because iron(III), titanium(IV), uranium(V1) and tantalum(V), so frequently associated with niobium in its natural sources (and responsible for most of the problems encountered in niobium analysis), do not interfere at the levels commonly found in natural niobium samples.In fact, only two elements, zirconium and tin, which may be present in some particular “low-grade” niobium ores (zirconium in naegite, hagatelite and eudialyte and tin in anialite or tantalocassiteritel) interfere. The other three elements extracted to a greater or lesser extent together with niobium(V) under the conditions of the procedure, i.e., molybdenum, tungsten and vanadium, are rarely found in natural niobium sources.192 When necessary, as shown in Table I, vanadium(V) interference (up to 100-fold excess) can be avoided by the addition of hydrazinium chloride. Tungsten(V1) and molybdenum(V1) interferences can be eliminated by a pre-washing step, introduced into the general spectrophotometric procedure, consisting of the addition of potassium fluoride to the aqueous phase, which prevents extraction of niobium(V), while molybdenum(V1) is extracted {and also tungsten(V1) to a lesser extent].After this washing step, boric acid is added to the aqueous phase, which demasks niobium(V), allowing it to be completely extracted with a fresh aliquot of reagent solution in chloroform [tungsten(VI) is not extracted under these conditions]. Only concentrations of molybdenum and tungsten at the same level as niobium(V) were tested and their effect could be eliminated by using about 0.18 g of potassium fluoride for masking and 0.34 g of boric acid for demasking. In niobium ores containing tin, this interference can be easily overcome by extracting niobium(V) from 2 M sulphuric acid medium instead of from hydrochloric acid, as in the absence of chloride ions tin(1V) is not extracted with 5,7-dichloroquinolin-8-ol.8 Zirconium(1V) inter- ference is avoided by masking it with EDTA, as shown in Table I.Determination of Niobium in Natural Samples synthetic samples. (niobium content approximately 0.3%). The proposed method has been applied to the determination of niobium in natural and Synthetic samples were prepared from calcite-base Canadian pyrochlores In order to reduce the niobium content the ore was TABLE I11 DETERMINATION OF NIOBIUM IN NATURAL AND SYNTHETIC ORES Sample Nb(V) expected, yo Nb(V) found, yo* Oka-1 No. 82 t * . .. a .0.36 0.352 Oka-1 No. 56 t .. .. .. 0.36 0.356 A $ .. .. .. .. .. 0.117 0.118 B : .. * . .. .. .. 0.705 0.703 c s .. .. .. * . * . 1.010 1.045 * Each value is the average of four separate determinations. t Calcite-base pyrochlore ores from Oka (Canada). Synthetic samples.1274 SANS-MEDEL AND D ~ A Z GARC~A thoroughly mixed with calcium carbonate powder and to increase niobium content it was mixed with pure niobium(V) oxide powder. The samples (0.5-1 g) were analysed for niobium following the procedures described previ- ously. The results obtained agreed very well with the expected values, as illustrated in Table 111. For samples with the lowest niobium content (less than 0.1%) negative errors (about -8%) may be obtained if insoluble salts form in the dissolution of the sample (excess- ive evaporation to fumes or orthophosphoric acid) or if the orthophosphoric acid to niobium ratio used is too large (see Table 11). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Atkinson, R. H., Steigman, J., and Hiskey, C. F., Anal. Chem., 1952, 24, 477. Gibalo, I. M., “Analytical Chemistry of Niobium and Tantalum,” Ann Arbor Science Publishers, Sanz-Medel, A., PhD Thesis, Zaragoza, 1973. Motojima, I. K., and Hashitani, H., Anal. Chem., 1961, 33, 48. Siroki, M., MariC, L., Herak, M. J., and DjordjcoiC, T. V., Anal. Chem., 1976, 48, 55. Ramakrishna, T. V., Rahim, S. A., and West, T. S., Talanta, 1969, 16, 847. Bonilla Simh, M. M., and Sanz-Medel, A., Anal. Quim., 1978, 74, 595. Sanz-Medel, A., and GutiCrrez Carreras, A. M., Analyst, 1978, 103, 1037. Moshier, R. W., “Analytical Chemistry of Niobium and Tantalum,” Pergamon Press, Oxford, 1964. Faye, G. H., Chem. Can., 1958, 10(4), 90. Sanz-Medel, A., Chmara Rica, C . , and PCrez Bustamente, J . A., Anal. Chem.., 1980, 52, 1035. Vaezi-Nasr, F., Duplessis, J ., and Guillaumont, R., Radiochem. Radional. Lett., 1979, 37, 153. Duplessis, J., Vaezi-Nasr, F., and Guillaumont, R., Analusis, 1978, 6, 446. Sanz-Medel, A., and Shnchez Uria, J. E., unpublished work. Ann Arbor, Mich., 1970. Received March 16th, 1981 Accepted June lst, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601268

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, December, 1981, Vol. 106, fip. 1275-1280 Potentiometric Determination of Fluoride by a 1275 Combination of Continuous-flow Analysis and the Gran Addition Method J.-Cl. Landry, F. Cupelin and C. Michal Service de Toxicologie Industrielle, d'A nalyse de I'A i r et de Protection Contre le Bruit, Institut d'HygiBne, Case Postale 109, CH-1211 Genbve 4, Switzerland A method is described for the determination of fluoride ion by continuous-flow analysis combined with the Gran addition method using an ion-selective electrode. The Gran addition method is a reliable potentiometric technique and an advantage of its use is that, in addition to the determination of the free and total fluoride concentration in a complexed medium, the complexation rate is also measured. The drawback of the method is the difficulty in adjusting the peristaltic pumps, which is time consuming.However, this extra time is com- pensated for by the simultaneous determination of fluoride ion concentration and the complexation rate. For concentrations of fluoride ion between 0.5 and 100;mg I-l, the precision is 5%. Keywords : Fluoride determination ; continuous-flow analysis ; ion-selective electrode ; Gran addition method ; air pollution Fluorides are discharged into the air by aluminium factories.1*2 Although monitoring stations give the most reliable measure of fall-out, this method is expensive3 and, as an alternative, solid absorbent collectors can be used.4 According to I ~ r a e l , ~ ~ ~ a good correlation between fluoride levels in plants and those collected on impregnated paper samplers is found, so this method was chosen for the collection of samples.The volume of air in contact with the filter- papers is not known; instead, a fluoride flux, consisting mainly of hydrogen fluoride or fluorine molecules, is determined. The collectors are made of a vertical support on to which are fixed ten filter-papers impregnated with a solution of sodium silicate. They are protected against bad weather conditions. Every month the filter-papers are collected and the absorbed fluoride is determined in the laboratory. Frant and Ross7 constructed an ion-selective electrode for the determination of fluoride ion and automated methods for the determination of fluoride in water have been p u b l i ~ h e d . ~ ~ ~ To verify the results obtained by spectrophotometry we have devised a potentiometric method making use of an ion-selective electrode ( ISE),1° which permits the determination of fluoride ion in the concentration range 0.5-70 mg 1-I.I t is simpler and less expensive than the spectro- photometric method, which gives no information about the complexing power of the medium towards fluoride ion. A continuous-flow potentiometric method using the Gran addition method is described, which allows the determination of fluoride ion in the concentration range 1-100 mg 1-I. It is possible to determine complexed fluoride, free fluoride and total fluoride ion concentrations and, in addition, the complexation rate can be measured simultaneously, which compensates for the difficult and time-consuming task of adjusting the peristaltic pumps.Theoretical The potential ( E ) of the fluoride ion-selective electrodes, with respect to the reference electrode, is given by the equation .. - - (1) E = E" + Ej + slog[ayFC'(F)] 0 . where E" is the standard electrode potential, Ej is the junction potential, yF is the activity coefficient of fluoride ion, C(F) and C'(F) are the free and total concentrations of fluoride ion, respectively, and s is the response of the potential to log(fluoride ion concentration). The complexing ratio ( K ) is given by1276 LANDRY et UE. : POTENTIOMETRIC DETERMINATION OF FLUORIDE Analyst, vd. 106 .. .. * (2) cc = C(F)/C’(F) . . 1 . The junction potential (Ej) can be reduced to a minimum by the use of a 1 M solution of potassium chloride in the reference bridge. A value of 1 for the ionic strength of the buffer solution (pH 5.3) keeps the value of the activity coefficient constant.In the continuous-flow addition method, a known flow (4) of concentration C(F) is added to an initial flow (4i) of the solution of unknown concentration, Ci(F). This is buffered by a solu- tion of flow 4t. Under equilibrium conditions Rearrangement of equation (3) and substitution of E’ = E” + Ej gives (4t + 41 + $)exp(E/s) = yFa$C‘(F)exp(E‘/s) + rFx$ici’(F)exp(E’/s) - (4) This is of the form Y = P X + q * . (5) * . * * (6) where 9 yFa$exp(E’/s) .. .. .. and = 3/Fa$iC’i(F)exp(E’/s) - * .. The measurement of at least two sets of values [for E and C’(F) J allows the calculation of the equation of the straight line y = f(x) and the slope p for a given complexation rate.Assuming a particular set of complexation conditions a value 9, can be determined from .. .. (7) Pl = alYF+XP(E’IS) .. Then, by changing the complexation conditions, it is possible to measure a new set of values 9, and a2 from .. .. ’ (8) .. * * (9) P2 = a,yF#exp(E’/s) .. By eliminating y&exp(E’[s) from equations (7) and (8), the following equation is obtained: $la2 = P,a, * - .. * . The equilibrium constant (K8) of the hydrogen fluoride - fluoride ion system is given by .. .. .. .. . . (10) yHPc ( HF) Ks = 3/FC(F)4H) where a ( H ) is the activity of the protons and 3/HF = 1. ion can be written as The total concentration of fluoride C‘(F) = C(HF) + C(F) . . .. * . . . (11) a = I/yvKsa(H) + 1 .. .. .. ,.* (12) Rearrangement of equations (10) and (11) gives Introduction of this expression for cf into equation (9) for two successive values of gives (92-91) = yFKS [91al(H)-p2a2(H)l * . .. . . (13) and using this equation, a straight line graph of slope KsyF is obtained. Experimental Preparation of Samples bottle. aliquot is filtered into the measuring cuvettes. The ten filter-papers, after sampling, are transferred into a 400-ml wide-necked polyethylene After agitation, an Water (250 ml) is added and the bottle is left to stand overnight.December, 1981 BY CONTINUOUS-FLOW ANALYSIS AND GRAN ADDITION METHOD 1277 Method of Additions and Measurement by Ion-selective Electrode The Carlo Erba continuous-flow analyser consists of an SD3 sample distributor, Model 1512, and a PP20 proportioning pump, Model 1512j20. Three minimicro 2/6 Ismatec peristaltic pumps are coupled to three Selectron SPR 32 relays, which, under the given impulse of the sampler, makes three additions of reference solutions at pre-selected times.All reagents used are of analytical-reagent grade. The determination is performed by means of a Beckman ion-selective electrode, No. 39600, with a Metrohm EA 425 silver - silver chloride electrode saturated with potassium chloride as the reference. The potential is measured by means of a Metrohm €2500 digital voltmeter. Buffer solution Standard solutions 0.5 . -- -. -- a Mixing 0.8 Potentiometer Fig. 1. Basic experimental arrangement for continuous analysis by the Gran addition method: PI, Pp and P,, additional peristaltic pumps ; flow-rates in mi min-l; ISE, ion-selective electrode; and RE, reference electrode.Manifold and Procedure The manifold (Fig. 1) draws a sample and adjusts its pH and ionic strength by means of 1 M acetic acid - acetate buffer solution of pH 5.3. Three peristaltic pumps perform the additions of the buffered (pH 5.3) standard solutions (1,2 and 3). The timer allows the setting of the three peristaltic pumps, P,, P, and P,, in order to introduce the standard solutions 1, 2 and 3 successively into the sample. It is necessary to measure the flow of each solution before 1 > E . - m Q, P .- c c Time/min -b Fig. 2. Response time of ISE for increasing (a) and decreasing (b) concentrations: 1, 0.5; 2, 5 ; and 3, 50 mg 1-1 of F-.1278 LANDRY et d.POTENTIOMETRIC DETERMINATION O F FLUORIDE Analyst, vd. 106 performing a series of measurements. log(fluoride ion concentration) must be known. The slope of the line relating electrode potential and Results and Discussion Electrode Response The response time of the ion-selective electrode is such that if it is used in continuous flow, the equilibrium potential is no longer reached when passing from a concentrated to a dilute solution. Fig. 2 shows the response of the ion-selective electrode in terms of time for (a) an increasing concentration and (b) a decreasing concentration of fluoride ion. The measured potential is no longer the same as the equilibrium potential. Under such circumstances, quantitative analysis is impossible as the error can be as high as 37%.Buffle et aZ.11 showed that the adsorption of fluoride ion at the solution - membrane interface reduces the sensitivity of the ion-selective electrode. We have shown that a rapid return to the base-line potential is possible by complexing the adsorbed fluoride.l* Ringbom12 showed that of the cations lanthanum(III), chromium(III), thorium(IV), aluminium(II1) and beryllium(I1) , beryllium( 11) is the most efficient complexing agent. Experimentally, for 10-3 M solutions of the above-mentioned cations and a 2 mg 1-1 solution of fluoride ion, beryllium(I1) proves to be a better complexing agent than aluminium(II1). The stability constants of the complexes decrease in the orderberyllium(I1) > aluminium(II1) > thorium(1V) > chromium(II1) > lanthanum( 111) > water.Beryllium(I1) being weakly hydrolysed makes it easier to work at a pH that is not very acidic. Experience shows that during the rinsing of the ion-selective electrode membrane the response time is considerably reduced as the concentration of beryllium(I1) increases. The base-line potential is recovered in 3.2 min when 0.1 M beryllium(I1) solution is used. Fig. 3 shows the manifold used for the potentiometric determination of fluoride ion by continuous-flow analysis, with the adjunction of the complexing agent for rinsing the electrode. The length ( N ) of the mixing coil is adjusted so that the complexing agent reaches the elec- trode's crystal when the equilibrium potential of the sample has been attained. In the part of the graph labelled A, the electrode responds to the fluoride solution.Then at B, the rinsing solution and the 0.1 M beryllium sulphate solution reach the electrode surface simultaneously. Finally, at C, only the rinsing solution of pH 5.3 flows over the membrane. There is hardly any change in the base-line potential. Under these conditions the calibration graph is linear up to a value of 1 mg 1-1 of fluoride ion with a slope of 63 mV. I t is possible to extend the linearity of the calibration graph by performing measurements at arbitrary times. The results obtained with this manifold are shown in Fig. 4. Peristaltic pump - Potentiometer Fig. 3. The basic arrangement for continuous analysis of F- by potentiometric method (the broken Flow-rates in ml min-1; line shows the position of the pumping arm when it has revolved through 90").ISE, ion-selective electrode; RE, reference electrode; and N, length of mixing coil. Gran Addition Method determined over two ranges : Under the experimental conditions used, the concentration of fluoride ion in solution can be (i) 0.5-10 mg I-1 with standard solutions of 5 , 10 and 15 mg 1-1 of fluoride ion; (ii) 5-100 mg 1-l with standard solutions of 50, 100 and 150 mg 1-1 of fluoride ion.December, 1981 BY CONTINUOUS-FLOW ANALYSIS AND GRAN ADDITION METHOD 1279 There is excellent agreement between experimental and theoretical values, with a correlation coefficient of 0.99 and an error at the origin of 0.2 mg 1-1 of fluoride ion. With a view to reducing the analysis time, the response kinetics of the ion-selective electrode were studied. w A 5 min - I Time/min-+ Fig.4. Electrode response obtained after rinsing with BeSO, solution A, Fluoride solution; for increasing concentrations of fluoride ions (mg 1-I). B, 0.1 M BeSO, solution; and C , pH 5.3 buffer solution. Partha~arathyl~ showed that the response of the ion-selective electrode can be represented by a hyperbolic function B E-El E-Ei A ~- Et - Eiti __ E,, (”) - - where E,, corresponds to the equilibrium potential, E to the potential a t a given time t, Ei to the potential at the beginning of the time ti and A and B are numerical constants. Using this equation as in the continuous-flow determination, straight-line graphs with slopes greater than the slopes corresponding to arbitrary time graphs (Fig. 5) were obtained.The concentration C’,(F) is not influenced by this phenomenon. 24 I 1 X’ r I C .g 20 - E % 16 B + 12 B + a 8 2 : : 4 . I - Q 40 20 0 40 80 Ci(F)/mg I-’ C(F)/mg I-’ Fig. 5. Representation of the line obtained by the Gran automated method: solid line, calculated equilibrium potential ; and broken line, instantaneous potential. Ct(F) : A, 40; B, 20; and C , 10 mg 1-1 of F-.1280 LANDRY, CUPELIN AND MICHAL Comparison of Results Obtained by the Gran Method with Other Methods Table I shows the results obtained for simultaneous measurements of solutions prepared from sodium silicate impregnated filter-papers that had been exposed over a period of 30 d in the R h h e valley between Steg and Martigny. Standard deviations, calculated for ten deter- minations, are given in milligrams per litre.The measurements by simple potentiometry give systematically lower values than the spectrophotometric method,l* while remaining within the confidence level limits defined by the standard deviation. The spectrophotometric and direct potentiometric methods do not take into account the complexation rate, which may therefore explain these differences. TABLE I COMPARATIVE RESULTS OF THE DETERMINATION OF FLUORIDE ION BY CONTINUOUS-FLOW SPECTROPHOTOMETRIC, POTENTIOMETRIC AND GRAN ADDITION METHODS Sample No. 1 2 3 4 5 6 7 8 9 10 Fluoride ion concentration/mg 1-1 Spectrophotometric Potentiometric Gran A I > 14.8 f 0.4 15.0 f 0.5 15.5 f 0.5 9.4 f 0.2 9.0 f 0.5 9.1 f 0.3 8.4 f 0.2 8.3 f 0.5 8.3 f 0.3 5.5 f 0.2 5.2 f 0.5 5.8 f 0.2 4.5 f 0.2 4.2 & 0.5 4.9 & 0.2 2.0 f 0.2 1.8 f 0.5 2,3 f 0.2 21.2 f 0.4 21.0 f 1.0 22.6 f 1.0 37.2 f 0.4 36.0 f 2.0 37.2 f 1.0 127.0 f 0.8 125.0 f 4.0 124.0 f 4.0 45.8 f 0.4 43.0 f.2.0 43.3 f 2.0 Conclusion The application of the Gran addition method has been extended to various media not requir- ing mineralisation. Excellent results have been obtained for soft drinks, wine, alcohols, milks, etc. Toothpastes with and without fluoride have also been examined. These results are not given here as they have no bearing on what has been demonstrated. Continuous-flow measurements have been made since June 1976 and the fluoride ion con- centrations of silicate-impregnated filter-papers have been monitored since 1965 at different localities of the Valais. No discrepancies have been found between the old and existing met hods. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References “Dossier Fluor,” CP 350, Association De Defense Contre Les €?manations Novices, Martigny, 1976. Garrec, J . P., and Battail, C., Pollut. Afmos., 1977, 74, 149. Powell, R. A., and Stockes, M. C., Atmos. Environ., 1973, 7 , 169. Wilson, W. L., Campbell, N. W., Eddy, L. D., and Poppe, W. H., Am. Ind. Hyg. Assoc. J., 1967, 27, Israel, G. W., Atmos. Environ., 1974, 8, 159. Israel, G. W., Atmos. Environ., 1977, 11, 183. Frant, M. S., and Ross, J , W., Jr., Science, 1966, 154, 1553. Erdmann, D. E., Environ. Sci. Technol., 1975, 9, 252. Bystrova, L. F., Stradomskii, V. B., and Nazarova, A. A., Gidrokhim. Matev., 1976, 63, 145. Landry, J.-Cl., Cupelin, F., Mitt. Geb. Lebensmittelunters. Hyg., 1978, 69. Buffle, J . , Parthasarathy, N., and Haerdi, W., Anal. Chim. Acta, 1974, 68, 253. Ringbom, A., “Les Complexes en Chimie Analytique. Parthasarathy, N., Thesis, UniversitC de GenBve, 1978. 254. Table des Constantes,” Dunod, Paris, 1967. Received December 1 Ith, 1979 Accepted May 6th, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601275

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, December, 1981, Vol. 106, pp. 1281-1287 1281 Sequential Multi-element Analysis Using Silver and Copper lon-selective Electrodes Saad S. M. Hassan” and M. M. Habib Department of Chemistry, Faculty of Science, A i n Shams University, Cairo, Egypt A simple method is described for the rapid and accurate sequential deter- mination of as little as 50 pg ml- 1 o f Ag+ , Cu2+, Ni2+, Cd2+, Pb2+, Zn2+, Fe3+, Th4+ and V4+ ions in binary, tertiary and quaternary mixtures. The method involves direct titration with sodium diethyldithiocarbamate a t pH 6 6 in 50% V / V ethanol using silver and copper ion-selective electrodes. Potentio- metric curves with sharp consecutive inflection breaks at the equivalence points and a mean recovery of 99.6% (mean standard deviation 0.3%) are obtained.The method has been used satisfactorily for the determination of the major alloying elements in brass, motor-brass, silver-brazing and nickel- silver alloys by both direct and spiking titration techniques. The results agreed well with the certified values. Keywords : Ion-selective electrodes ; copper- and silver-base alloys ; sequential determination o j metals ; diethyldithiocarbamate There has been increasing concern about the use of EDTA and related compounds as titrants for the determination of heavy metal ions using ion-selective electrodes. The commercially available copper, cadmium, lead and calcium electrodes have been used to monitor the titration of their respective i0ns.l These electrodes have also been utilised for the titration of other metal ions either by addition of excess of EDTA or 1,2-diaminocyclohexane-NNN’N’- tetraacetic acid (DCTA) and subsequent measurement of the unreacted complexing agent,2s3 or by displacement of the electrode metal ion from its EDTA complex followed by titration.*-* On the other hand, there is a dearth of satisfactory procedures for simultaneous multi-element titration without prior separation. The only two papers have described the simultaneous titration of binary mixtures containing copper ions with EDTA and tetraethylenepentamine (TEPA) using the copper electr~de.~t~ A detailed investigation with many other organic reagents known to form complexes andlor precipitants with a variety of metal ions revealed that sodium diethyldithiocarbamate (NaDDC) offers at least three particular advantages over EDTA and similar titrants.NaDDC is a thiol-containing compound detectable by metal sulphide membrane electrodes, it reacts stoicheiometrically with many metal ions under the same conditions and at the same pH, and the solubility products or stability constants of many of its metal complexes vary over several orders of magnitude.QJo This suggests its usefulness for the direct titration of many metal ions and also for simultaneous multi-element titration, as sequential binding of the metals with diethyldithiocarbamate ions during titration may involve stepwise titration curves without overlapping of the end-point breaks. The work described here shows that NaDDC is advantageous for use with solid-state silver and copper ion-selective electrodes for the simultaneous determination of several metal ions at concentrations down to 50 pg ml-l in various mixtures and alloys.Experimental Apparatus All potentiometric measurements were made with an Orion Microprocessor Ionalyzer, Model 901, using silver sulphide (Orion 94-16) and copper (Orion 94-29) ion-selective elec- trodes in conjunction with a double-junction reference electrode (Orion 90-02) containing 1004 potassium nitrate solution in the outer compartment. Reagents was used throughout. versity of Delaware, Newark, Del. 19711, USA. All reagents were of analytical-reagent grade unless stated otherwise, and de-ionised water * To whom correspondence should be addressed. Present address : Department of Chemistry, Uni-1282 HASSAN AND HABIB : SEQUENTIAL MULTI-ELEMENT ANALYSIS Analyst, Vol.106 Solutions of silver, copper, lead, nickel, cadmium, zinc, iron, thorium and vanadium nitrates and zirconyl chloride were prepared and standard- ised with EDTA using visual titration procedures.ll Mixtures of copper - zinc, silver - copper - zinc, copper - lead - zinc and copper - nickel - zinc containing 0.5 mg ml-l of each metal were prepared from the nitrate salts in 50% V/V ethanol. These mixtures were standardised by titration with 0.1 M NaDDC solution using a silver ion-selective electrode in conjunction with a double-junction reference electrode. A solution of NaDDC was pre- pared in 50% VjV ethanol and standardised by potentiometric titration with 0.01 M standard copper nitrate solution using a copper ion-selective electrode in conjunction with a double- junction reference electrode.Stock standard metal solutions, 0.1 M. Standard sodium diethyldithiocarbamate solution., 0.01 M. Procedure Transfer an accurately weighed amount of the alloy (approximately 0.1 g) into a covered 50-ml beaker and add 5 ml of 10 M nitric acid. Heat the solution gently on a sand-bath at about 250 "C to dissolve the alloy and then carefully evaporate it nearly to dryness. Repeat the addition of acid and evaporation until the sample is completely dissolved. Dissolve the residue in 2 ml of 0.1 M nitric acid and add 20 ml of de-ionised water. Adjust the pH to 4-6 by addition of 0 . 0 5 ~ sodium hydroxide solution. Dilute the solution to 100ml with de-ionised water in a calibration flask.Transfer a 5.00-ml volume of the alloy solution, or a volume containing 0.5-5 mg of each metal, into a 50-ml beaker. Add 5 ml of 96% ethanol and insert the copper or silver ion- selective electrode in conjunction with a double-junction reference electrode in the solution (a silver electrode is used with silver- and copper-base alloys whereas a copper electrode is used with copper-base alloys). Titrate the solution potentiometrically with 0.01 M standard NaDDC solution. Add 1.00 ml of a standard synthetic mixture of copper - zinc, silver - copper - zinc, copper - lead - zinc and copper - nickel - zinc containing 0.5 mg of each metal to the brass, silver-brazing, motor-brass and nickel-silver alloy solution, respectively. Add 5 ml of 96% ethanol to each and titrate potentiometrically with standard 0.01 M NaDDC solution using a silver or copper electrode.Carry out a blank run without the alloy solutions and compare the titres at each inflection for both the standard and the standard plus the alloy. The difference in the titres a t each inflection is equivalent to the metal in the alloy solution. The constituents of various metallic combinations are determined by similar procedures. Known amounts (about 0.5-1 mg) of silver or copper are added before titration of mixtures that do not contain these metals. Alternatively, transfer 5.00-ml volumes of the alloy solutions into 50-ml beakers. Results and Discussion A preliminary investigation showed that NaDDC can be detected by the metal sulphide membrane electrodes. It was found that the response of the solid-state silver (silver sulphide membrane) and copper [silver sulphide - copper( I I) sulphide membrane] ion-selective elec- trodes at 25 "C in 50-75% V/V ethanol is Nernstian towards NaDDC solutions of concentra- tions down to 1 0 - 5 ~ with an average anionic slope of 55mV per concentration decade.Titration of solutions containing 50-700 pg ml-l (approximately 1-10 mM) of Ag+, Cu2+, Ni2+, Pb2*, Cd2+, Zn2+, Fe34, Th4+, V4+ and Zr4+ with NaDDC at pH 4-6 in 50% VjV ethanol using either the silver or copper ion-selective electrode resulted in sharp inflection breaks at the equivalence points, ranging from 100mV for Zn2+ to 700mV for Ag+. The results obtained showed a mean recovery of 99.5% with a mean standard deviation of 0.2%. The stoicheiometries of the reactions are exactly 1 : 1, 1 : 2 and 1 : 3 (metal: DDC) with mono-, di- or tetra-, and trivalent metal ions, respectively. Zirconyl chloride, however, reacts in a molar ratio of 1 : 1, which is in good agreement with recent findings by other workers.l2 The accuracy and precision of the procedure were substantiated by comparing the results of ten repetitive titrations carried out on three separate stock solutions of Pb2+, Cd2+ and Cu2+ (100 pg ml-l) using both the present procedure and visual titration with EDTA using xylenol orange (urotropine buffer, pH 4.6), Eriochrome Black T (ammonium chloride - ammonia buffer, pH 10) and Murexide (ammonium chloride - ammonia buffer, pH lo),December, 2981 USING SILVER ,4ND COPPER ION-SELECTIVE ELECTRODES 1283 respectively.The mean recoveries obtained by the present method (mean standard devi- ation 0.2%) and the EDTA titration (mean standard deviation 0.6%) were almost identical and the discrepancies never exceeded 1%. These results show that metals for which no electrodes are available, and which are determined by indirect or displacement procedures,2-6 can be determined satisfactorily by direct titration with NaDDC using commercially available silver or copper ion-selective elxtrodes. Sequential Multi-element Determination When binary and tertiary mixtures containing Cu2+ with Cd2+, Zn2+, Pb2+, Ni2+, V4+ and Th4+ are titrated with NaDDC at pH 4-6 in 500/, Y/V ethanol using silver and copper elec- trodes, potentiometric curves with consecutive sharp inflection breaks a t the equivalence points of the constituent elements of these mixtures are obtained (Figs.1 and 2). The 200 > 100 E .- c o o 5 -100 . - c c -200 2 -300 m -0 c - $ -400 W - 500 2 4 I 1 I 1 I 1 2 4 2 4 2 4 Volume of 0.010 M NaDDChl Fig. 1. Typical potentiometric titration curves for some 1 + 1 binary metallic mixtures with NaDDC at pH 4-43 using (10) silver and (0) copper electrodes; 1 ml of a solution 0.01 M in each metal ion in a total volume of 10 ml of water - ethanol ( 1 +- 1 ) . 200 100 > 1 € 0 6 --I00 - co .- c c L1 -0 Q) -200 2 +-’ o -300 - W -400 - 500 ‘ 4 8 4 8 4 8 Volume of 0.010 M NaDDCiml Fig. 2 . Typical potentiometric titration curves for some 1 + 1 + 1 tertiary metallic mixtures with NaDDC a t pH 4-6 using (G) silver and (0) copper electrodes; 1 ml of a solution 0.01 M in each metal ion in a total volume of 10 ml of water - ethanol ( 1 + 1 ) .sequence of inflections is in the order Cu2+, Ni2+, Pb2+, Cd2+ and Zn2+, which is the same as the sequences of the solubility products and stability constants of the DDC derivatives of these rnetal~.9,~* The results obtained (Tables I and 11) with both silver and copper elec- trodes with various mixtures and different metallic ratios are almost identical. The mean recovery was 99.7% with a mean standard deviation of 0.3%. However, mixtures con- taining those metal ions whose stability constants or solubility products are of the same order of magnitude (e.g., Pb2+, Cd2+ and Ni2+) gave only one unresolvable inflection break. Thus, mixtures containing copper with such metals can be assessed only for copper.TABLE 1 SEQUENTIAL DETERMINATION OF SOME HEAVY METAL IONS IN BINARY MIXTURES BY TITRATION WITH NaDDC USING SILVER AND COPPER IOK-SELECTIVE ELECTRODES fir, M, Altxture w \ I - +-T KO. of Added/ Standard Added/ Standard RI, &II analyses (n) w6 ml-’ Recovery, yo deviation, % 1~6 ml-l Recovery, Yo deviation, % Ag c u 6 99-198 99.0-101.lj 0.6 63-1 89 100.0-101.6 0.3 Crr XI 6 6:3-127 100.0-101.6 0.4 52-211 100.0-101.9 0.5 c u Pb 8 63-125 98.4-100.0 0.4 205-414 99.0-99.5 0.2 cu Cd 8 63-127 9s. 5-100.0 0.4 112-224 98.2-100.0 0.5 Cu Zn 8 63-127 97.9-98.5 0.2 68-135 97.1-98.5 0.4 Cu Th 6 63-12i 101.1-101.6 0.1 208-417 98.3-99.0 0.2 c u v 8 63-12i 100.0-101.6 0.4 45-91 9i.8-98.9 0.31284 HASSAN AND HABIB : SEQUENTIAL MULTI-ELEMENT ANALYSIS A.utaZyst, VoZ.106 Binary and tertiary mixtures containing both silver and copper ions show inflection breaks at 75 & 1% of the expected titre for silver and 125 & 1% of that expected for copper. The results are fairly consistent over a wide range of metal proportions. It is apparent from the potentiometric titration curves for pure solutions of silver and copper that the equivalence points of silver are always located at negative potential values and copper starts to form a soluble complex with NaDDC a t positive potential values corresponding to the precipitation of 75% of silver. This resulted in complexation of about 25% of copper before complete precipitation of silver. The overlap is presumably due to the close values of the solubility product of AgDDC and the stability constant (PI) of Cu(DDC),.S By multiplying the titre at the first inflection due to silver by a factor of 1.33, the exact titre for silver is obtained, which, on subtraction from that at the second inflection, gives the exact titre for the copper ion.The results obtained (Tables I and 11) for some mixtures containing silver (e.g., silver - copper, silver - copper - zinc, silver - copper - lead and silver - copper - cadmium) show a mean recovery of 99.6% with a mean standard deviation of 0.30/6 for all of the metals in these mixtures. In general, the use of the silver electrode with mixtures containing silver gives slightly better results than those obtained with the copper electrode. Quaternary silver - copper - lead - zinc, silver - copper - cadmium - zinc and silver - copper - zinc - nickel mixtures were also titrated with NaDDC using both silver and copper electrodes.Potentiometric curves with four sharp inflection breaks due to the constituent metal ions of these mixtures were obtained (Fig. 3). Table I11 gives results obtained for the determina- tion of the constituents of some of these mixtures where the correction factor referred to above was used. The mean recovery of all the metals was 99.9% with a mean standard deviation ’of 0.3%. Silver - nickel - cadmium - zinc and silver - copper - lead - cadmium mixtures displayed only three inflections due to silver, nickel - cadmium and zinc, and silver, copper and lead - cadmium, respectively. These results show that silver and zinc can be determined simultaneously, in the presence of nickel and cadmium, and silver and copper can be determined simultaneously in the presence of lead and cadmium.Copper and zinc can also be determined in the presence of nickel and cadmium. In the absence of both silver and copper ions, binary, tertiary and quaternary metallic mixtures of all the other metals show only one unresolvable inflection break a t their total titres. 400 , > 300 E =s 200 .- 5 100 +- o n 0 2 tj -200 -300 - 400 Q, -0 -100 - 2 4 6 8 2 4 6 8 I 2 1 1 1 1 2 4 6 8 Volume of 0.010 M NaDDCimI Fig. 3. Typical potentiometric titration curves for some 1 + 1 + 1 + 1 quaternary metallic mixtures with NaDDC at pH 4-6 using (0) silver and (0) copper electrodes; 1 ml of a solution 0.01 M in each metal ion in a total volume of 10 ml of water - ethanol (1 + 1).Sequential Determination of Major Elements in Alloys The major elements in some silver- and copper-base alloys were determined simultaneously by titration with NaDDC solution using the silver and copper electrodes after acid decompo- sition of the alloy samples and adjustment of the pH of their solutions to 4-6, The standard known-addition “spiking” titration technique was also used for the analysis of these alloys.P 0 3 TABLE I1 SEQUENTIAL DETERMINATION OF HEAVY METAL IONS IN TERTIARY MIXTURES BY TITRATION WITH NaDCC USING SILVER AND COPPER Mixture --,*------, NO. of M, M, M, analyses (n) Cu Cd Zn 8 Cu Pb Zn 6 Cu Ni Zn 6 Ag Cu Cd 6 Ag Cu Pb 8 Ag Cu Zn 8 A= pg ml-l 3 1-95 3 1-95 31-95 99-198 99-198 99-198 M, I ON-SELECTIVE ELECTRODES Recovery, % 98.4- 1 00.0 98.9- 100.0 100.0-1 0 1.6 98.5-99.1 99.0- 1 0 1.0 98.0-99.1 Standard deviation, % 0.5 0.3 0.5 0.2 0.5 0.3 r------ Added/ pg ml-l 1 12-224 20 7-4 14 52-105 63-127 63-127 63-127 TABLE I11 M2 Recovery, 98.1-99.1 99.0-99.4 100.0- 10 1.9 98.4-98.9 98.4-98.9 98.4-100.0 % 7 Standard deviation, % 0.3 0.1 0.5 0.2 0.2 0.5 7 Added/ pg ml-l 65-130 65-130 65-130 55-1 11 112-224 65-130 31, ".-----A- Recovery, 98.5-1 00.0 % 98.5-100.0 97.0-1 00.0 98.2-98.8 98.2-99.1 97.9-98.5 --i -~ Standard deviation, yo 0.4 0.4 0.3 0.2 0.3 0.2 SEQUENTIAL DETERMINATION OF SOME HEAVY METAL IONS IN QUATERNARY MIXTURES BY TITRATION WITH NaDDC USING SILVER AND Mixture r-------- -\ NO.of M, M, M, M, analyses (n) Ag Cu Pb Zn 3 Ag Cu Cd Zn 3 Ag Cu Ni Zn 3 Ag Cu Pb Cd 3 Ag Ni Cd Zn 3 Mixture r P A - - \ NO.of M, M2 M, M, analyses (n) Ag Cu Pb Zn 3 Ag Cu Cd Zn 3 Ag Cu Ni Zn 3 Cu Pb Cd 3 $ Ni Cd Zn 3 COPPER ION-SELECTIVE ELECTRODES Ml Added/ pg ml-l 99-2 14 54-198 99-214 99-161 99-198 Recovery, yo 100.9-101.5 101.2-10 1.8 100.5-10 1 .O 98.1-98.9 10 1 .O-10 1.5 M3 1 Standard deviation, % 0.2 0.2 0.2 0.3 0.2 , Added1 pg ml-1 112-168 118-177 55-83 112-168 118-177 Recovery, % 98.2-99.4 98.9-99.2 98.2-98.8 - - 1 Standard deviation, yo 0.5 0.1 0.2 - M, Added/ Standard p g ml-l Recovery, yo deviation, yo f \ 63-127 98.9-1 00.0 0.4 6 3-9 5 9 8.4-9 8.6 0.1 63-95 10 1.6-102.1 0.2 63-95 98.4-98.9 0.2 56-84 - - M4 Added/ Standard pg ml-l Recovery, yo deviation, yo 65-130 10 1.5-1 02.3 0.3 65-130 100.0-101.5 0.5 65-130 10 1.5-102.3 0.5 177-358 - - 65-130 98.5-100.0 0.51286 HASSAN AND HABIB : SEQUENTIAL MULTI-ELEMENT ANALYSIS Analyst, VoZ.106 This involved addition of aliquots of standard synthetic copper - zinc, silver - copper - zinc, copper - lead - zinc and copper - zinc - nickel mixtures containing 0.5 mg of each metal to a known volume of the brass, silver-brazing, motor-brass and nickel-silver alloys, respectively, before titration. The difference between the titres for the standard and the standard plus alloy sample at each inflection break is due to the metals in the alloy sample. With the brass, motor-brass and nickel-silver alloys, the first and third inflection breaks are due to copper and zinc, respectively. The three inflection breaks obtained with the silver-brazing alloy are due to silver, copper and zinc, respectively.The titre for silver is multiplied by a factor of 1.33 to obtain the exact titre, then subtracted from the next inflection to obtain the exact titre for copper. The results obtained by direct and spiking titrations (Table IV) are in good agreement and compare favourably with the certified values. TABLE IV SEQUENTIAL DETERMINATION OF THE MAJOR ALLOYING ELEMENTS OF SOME SILVER- AND COPPER-BASE ALLOYS BY TITRATION WITH NaDDC USING A SILVER ION-SELECTIVE ELECTRODE Direct titration of metal Spiking titration of metal Certified L L Major content, rFound,* Recovery, Standard’ ‘Found,* Recovery, 01 deviation, Standard’ yo 0 Alloy elements % % % deviation, % % Brass .. .. .. Cu 59.9 60.1 100.3 0.4 59.9 100.0 0.3 Zn 40.1 39.9 99.5 0.3 39.6 99.2 0.2 Silver brazing .. . . Ag 66.9 66.5 99.4 0.5 66.7 99.7 0.3 c u 20.1 20.4 101.5 0.4 19.9 99.0 0.3 Zn 13.0 13.1 100.8 0.2 12.8 98.5 0.2 Motor brass .. . . Cu 61.9 61.5 99.4 0.3 62.1 100.3 0 4 Pb 4.0 4.1 102.5 0.4 4.1 102.5 0.3 Zn 34.1 34.4 100.9 0.2 34.4 100.9 0.3 Nickel silver . . . . Cu 55.1 55.4 100.5 0.4 55.3 100.4 0.2 Ni 18.2 17.9 98.4 0.4 18.1 99.5 0.2 Zn 26.4 26.7 101.1 0.3 38.8 101.5 0.3 * Average of three measurements. Conclusion Sequential potentiometric titration for multi-element analysis without overlapping requires: (a) a titrant that reacts with the various metal ions in the mixture to form com- plexes or precipitates with at least 2 decades difference in their stability constants or solubility products; (b) an electrode sensitive to the titrant used; and (c) the presence of an “electrode metal ion” in the test solution that should form the highest stable or the least soluble derivative of the titrant.If the titrant used reacts differentially with the metals and the electrode used is not sensitive to it, binary mixtures containing the electrode metal ion can only be sequentially titrated (e.g., copper - zinc mixture with EDTA using the copper elec- trode). The convenience of NaDDC compared with EDTA is due to the close similarity of the stability constants of the EDTA complexes of many metals. The usefulness of the solid-state silver and copper ion-selective electrodes, on the other hand, is attributed to their sulphide membranes, which can sense the thiol group of NaDDC.It can be seen that NaDDC and a silver or copper ion-selective electrode are advantageous for the sequential determination of several metal ions in mixtures. This approach proved to be satisfactory for the determination of the following: (a) binary mixtures of silver or copper with lead, cadmium, zinc and nickel; (b) tertiary mixtures of silver and copper with lead, cadmium, nickel or zinc, or of copper and zinc with lead, cadmium or nickel; and (c) quaternary mixtures of silver, copper and zinc with lead, cadmium or nickel. Other metallic combinations in alloys, pharmaceuticals and biological solutions may also be determined. The accuracy and precision of the method compare favourably with those obtained by standard procedures and the time required for the assay of four metals in mixtures is only 15 min.The present procedure provides an effective means for routine multi-element analysis without prior separation or manipulation steps. It is much more simple and rapid than those in current use or recently rep~rted.ll+l~-~~December, 1981 USING SILVER AND COPPER ION-SELECTIVE ELECTRODES References 1287 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Vesely, J ., Weiss, D., and Stulik, K., “Analysis with Ion-selective Electrodes,” Ellis Horwood, Meer, J., Boef, G., and Linden, W., Anal. Chim. Acta, 1975, 79, 27. Sikhs, L., and Suchknek, M., Anal. Lett., 1970, 3, 613. Baumann, E., and Wallace, R., Anal. Chem., 1969, 41, 2072. Chang, F., and Cheng, K., Anal. Chim. A d a , 1975, 76, 177. Olson, V., Carr, J., Hargens, R., and Force, R., Anal. Chem., 1976, 48, 1228. Meer, J., Boef, G., and Linden, W., Anal. Chim. Acta, 1975, 76, 261. Ross, J., and Frant, M., Anal. Chem., 1969, 41, 1900. Hulanicki, A., Talanta, 1967, 14, 1371. Stary, J., and Kratzer, K., Anal. Chim. Ada, 1968, 40, 93. Vogel, A., “A Text Book of Quantitative Inorganic Analysis,” Third Edition, Longmans, London, Kaushik, N., Bhushan, B., and Chnatwal, G., J . Inorg. Nucl. Chem., 1980, 42, 457. Elwell, W., and Scholes, I., “Analysis of Copper and Its Alloys,” Pergamon Press, Oxford, 1967. Morelli, B., Analyst, 1980, 105, 396. Pabalkar, M. A., Naik, S. V., and Sanjana, N. R., Analyst, 1981, 106, 47. Thomas, L., and Chamberlin, G., “Colorimetric Chemical Analytical Methods,” Ninth Edition, Lewis, C., and Ott, W., “Analytical Chemistry of Nickel,” Pergamon Press, Oxford, 1970. Chichester, 1978. 1961. Tintometer, Salisbury, 1980. Received A p i l 8th, 1981 Accepted July 2nd, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601281

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

1288 Analyst, December, 1981, Vol. 106, ++. 1288-1295 Diff erential-pulse Voltammetric Determination of Phosphate as Molybdovanadophosphate at a Glassy Carbon Electrode and Assessment of Eluents for the Flow Injection Voltammetric Determination of Phosphate, Silicate, Arsenate and Germanate A. G. Fogg and N. K. Bsebsu Chemistry Department, Loughborough University of Technology, Loughborough, Leicestershire, L E 11 3T U The redox behaviour of molybdovanadophosphate at a glassy carbon electrode is described and a procedure is given for the voltammetric determination of phosphate as molybdovanadophosphate at a glassy carbon electrode in a static system. Procedures are also given for the voltammetric flow injection determination of phosphate, silicate, arsenate and germanate by the injection of heteropolyacids pre-formed in various aqueous , aqueous acetone and aqueous ethanolic reagents into eluents consisting of the reagent blank.This procedure effectively eliminates the background signal of the blank and allows precise determinations to be made. Plots of electrode potential against the current obtained are given. Silicate and phosphate can be determined at lo-' and 10-6 M levels, respectively. M level, and the precise determination of germanate is difficult owing to adsorption at the glassy carbon electrode. Arsenate has only been determined at the Keywords : Orthophosphate, silicate, arsenic and germanium determination ; voltammetry ; pow injection analysis Previously, orthophosphate was determined as 1 2-molybdophosphate in aqueous acidic molybdate solution by linear-sweep voltammetry and differential-pulse voltammetry at a glassy carbon electrode in a static system.l Further, the solution conditions of four spectro- photometric methods based on the formation of the p-heteropolymolybdates of ortho- phosphate, silicate, arsenate and germanate, stabilised by acetone, were used in the deter- mination of these ions by similar voltammetric methods.2 In this paper a procedure is given for the differential-pulse voltammetric determination of orthophosphate as molybdovanado- phosphate at a glassy carbon electrode in a static system, and an assessment is made of the various blank solutions used in the static systems as eluents for the flow-injection deter- mination of the pre-formed heteropolyacids of orthophosphate, silicate, arsenate and germanate.Fujinaga et aL3 have reported recently the use of an aqueous ethanolic molyb- date eluent for the voltammetric flow injection analysis of orthophosphate. From details given in their paper it is clear that the blank becomes a problem when solutions containing orthophosphate at about the 2.5 p.p.m. level are injected. The aqueous ethanolic eluent used by them has also been assessed here. Experimental Voltammograms were obtained using a PAR 174 polarographic analyser (Princeton Applied Research) with three-electrode operation (glassy carbon electrode, platinum counter electrode and calomel reference electrode). For both linear-sweep and differential-pulse voltammetry, a sweep rate of 5 mV s-l was used: for the latter technique a pulse height of 50 mV and a pulse frequency of 0.5 s were applied.Between scans the glassy carbon electrode was cleaned as described previously with 1 M sodium hydroxide solution,l and occasionally, as required, by repolishing with alumina and ethanol on a polishing cloth. With molybdovanado- phosphate solutions at the lo-* M level, a short polish (10 s) between scans gave improved results. Flow injection analysis was applied in a similar manner to that used in a study of food colouring matters.* A Metrohm detector cell (EA 1096) was used with a Metrohm glassy Voltammograms were recorded on a Gould HR2000 X - Y recorder.FOGG AND BSEBSU 1289 carbon electrode (EA 286) but it was found less problematic for long term working not to insert the counter and reference electrodes but to partially immerse the cell in electrolyte (0.01 M sulphuric acid), contained in a beaker.Contact between the electrolyte to the same counter and reference electrodes used for the static work was made by means of salt bridges. Flow of eluent (approximately 2.5 ml min-l) was produced by means of the Metrohm pressure bottle system (EA 1101) working at 0.8 bar. Eluents were degassed by means of a vacuum pump: the use of an ultrasonic bath in conjunction with this would be advantageous. Injections were made by means of a Rheodyne sample injection valve (5020) with a loop capacity of 0.1 ml. Unless otherwise indicated, the glassy carbon electrode was cleaned with 1 M sodium hydroxide solution daily or when changing eluents.A good method of assessing the condition of the glassy carbon electrode is to make a differential-pulse voltam- metric scan of a molybdophosphate solution in a static system. The electrode should be polished until well formed voltammetric peaks are obtained in this static system. The potential of the detector cell was controlled by means of the PAR 174 polarographic analyser ; current peaks were recorded on a Tarkan 600 Y - T recorder. Voltammetric Determination of Orthophosphate as Molybdovanadophosphate in a Static System The solution conditions of the colorimetric molybdovanadate method5 were found to be readily adaptable to the voltammetric method. Linear-sweep voltammetry from 0 V to more positive potentials gave an initial combined cathodic - anodic peak, which corresponded to a differential-pulse voltammetric peak a t +0.13 V [see Fig.1 (a) and ( b ) ] . The anodic portion arises from re-oxidation of material reduced earlier in the scan. Seven and a half times the amount of molybdovanadate reagent used in the colorimetric method was required in order to give reasonable rectilinearity for orthophosphate concentrations of up to 4 x As with 12-molybdophosphate in aqueous acidic molybdate solution,l the height of the molybdovanadophosphate differential-pulse voltammetric peak a t +O. 13 V was found to increase slightly when the glassy carbon electrode was left in the solution on open circuit. However, for molybdovanadophosphate, even when a strictly controlled timing procedure was adopted coeficients of variation of about 7% were typical a t the 1 x lo-* M level and also at the 1 x M level where a smaller amount of molybdovanadate reagent was used in order to reduce the size of the blank.Differential-pulse voltammetric scans from +0.6 V, in the negative-potential direction, showed an initial peak at +0.50 V that did not increase in height when the glassy carbon electrode was left in the solution on open circuit. The height of this peak was very repro- ducible and a voltammetric procedure was developed based on the measurement of this peak. Linear-sweep and differential-pulse voltammetric scans from +0.6 to 0 V are shown in Fig. 1 (c) and (d). The linear sweep voltammogram shows that the electrochemical processes are wholly cathodic. M. Reagents Dissolve 0.408 g of analytical-reagent grade potassium dihydrogen orthophosphate in water and dilute to 1 1 in a calibrated flask.Prepare less concentrated standard solutions by dilution. Molybdoaanadate reagent. Dissolve 0.50 g of ammonium metavanadate in warm distilled water. Add a solution of ammonium molybdate (10 g) in water and dilute to 1 1 with distilled water. Standard orthophosphate solution, 3 x M (285 pg ml-l of PO,3-). This solution is 3 x 10-3 M in orthophosphate. Cool and add 125.0 ml of nitric acid (70% m/m). Procedure Transfer by pipette aliquots of standard orthophosphate solution, containing 25-2000 pg of orthophosphate, into a series of 50-ml calibrated flasks. To each add 7.5 ml of molybdo- vanadate reagent and dilute to 50 ml with water, except for solutions containing less than 200 pg of orthophosphate, when only 1.0 ml of molybdovanadate reagent should be added.Transfer each solution in turn into a voltammetric cell. Place the clean, dry, glassy carbon electrode in the solution. After 10 s close the cell circuit with the initial potential at +0.8 V in the differential-pulse mode. After a further 20 s, when the current has stabilised, obtain1290 FOGG AND BSEBSU : DP VOLTAMMETRIC DETERMINATION OF Analyst, ‘vd. 106 200 f . c t f! 0 100 3 +0.2 +0.4 +0.6 0 PotentialiV +0.2 +0.4 +0.6 Potent i a IN 0 & 5. 200 . c c 2 3 0 7 00 0 +0.6 +0.2 -0.2 +0.6 +0.2 -0.2 PotentialiV PotentialiV Fig. 1. Linear-sweep and diff erential-pulse voltam- mograms of molybdovanadophosphate a t a glassy carbon electrode in a static system: A, 2 x lo-* M molyb- dovanadophosphate ; and B, molybovanadate blank.(a) Linear-sweep voltammogram, and (b) differential- pulse voltammogram, scanned in the direction of positive potential; (c) linear sweep voltammogram, and (d) diff erential-pulse voltammogram, scanned in the direction of negative potential. a differential-pulse voltammogram between +0.8 and +0.2 V. scans as indicated earlier. Clean the electrode between Results graphs, are shown in Fig. 2. 10-4 M ranges are rectilinear and coefficients of variation at 2 x were 3%. Typical diff erential-pulse voltammograms, obtained for the production of calibration and 0 . 5 4 x and 2 x M levels Calibration graphs for both the 0 . 5 4 x Assessment of Eluents for the Flow Injection Voltammetric Determination of Orthophosphate, Silicate, Arsenate and Germanate In this study the heteropolyacids were formed using the procedures described above (for molybdovanadophosphate) or previo~sly.l-~ The pre-formed heteropolyacids were then injected into a flowing stream of eluent that had the same composition as the blank solution. This compensated for the background current of the reagents.December, 1981 Standard determinand solutions PHOSPHATE AND ASSESSMENT OF ELUENTS FOR FI VOLTAMMETRY 1291 Standard orthophosphate solution, 285 pg ml-1.Standard arsenic( V ) solution, approximateZy 0.24 mg ml-l. Prepare as above. Weigh accurately about 2 g of disodium hydrogen arsenate heptahydrate, dissolve it in water and dilute to 1 1 in a cali- brated flask. Weigh accurately about 0.05 g of quartz into a platinum (or nickel) crucible and fuse it with about 2 g of sodium carbonate.Cool the melt, dissolve in water, dilute to 1 1 in a calibrated flask and store in a polyethylene bottle. Standard germanium solution, approximately 0.22 mg ml-l. Dissolve about 0.16 g, accurately weighed, of germanium( IV) oxide in dilute ammonia solution (silica free), neutralise with dilute sulphuric acid and dilute to 500 ml in a calibrated flask. Standard silica solution, approximately 0.05 mg ml-l. 120 Q, . c c 80 f?? 3 0 40 0 +0.8 +0.6 +0.4 Potentia I/V 0- +0.8 +0.6 +0.4 PotentialiV Fig. 2. Differential-pulse voltammograms for obtaining ( a ) A, 0; ( b ) A, calibration graph using recommended method. B, 1; C, 2 ; D, 3 ; and E, 4 x lo-* M in phosphate. 0 ; B, 1 ; C, 2 ; D, 3; and E, 4 x M in phosphate.Reagents sulphuric acid to 200 ml of water. solution and dilute when cold to 250 ml with water. i4cidic molybdate solution, 2y0 mlV. Add 35 ml of analytical-reagent grade concentrated Dissolve 5 g of ammonium molybdate in the resulting Molybdovanndate reagent. Prepare as above. Ammonium molybdate solution, 87' mjV. ~4mVZOniUm molybdate - sztlphuric m i d solution. Prepare and store this in a polyethylene bottle. Dissolve 40 g of ammonium paramolyb- date heptahydrate in 500ml of water in a polyethylene beaker and add 500 ml of 1 M sulphuric acid solution. Sodium molybdate - tartaric acid - hydrochloric acid solution. Dissolve 15 g of sodium molybdate dihydrate and 12 g of tartaric acid in water, add 45 rnl of concentrated hydro- chloric acid and dilute to 500 ml.A brown colour develops in the reagent solution after a few days so it should be discarded after 3 d. Sodium molybdate - sulphuric acid solution. Carefully add 80 ml of concentrated sulphuric acid to about 100 ml of distilled water. Cool, then dissolve 25.75 g of sodium molybdate in the mixture and dilute to 250 ml with distilled water. Procedures for formation of heteropolyacids and preparation of eluents The eluents are prepared exactly as described in the procedures given for the formation of Iieteropolyacids except that phosphate or other determinands are omitted and ten times the volume is usually prepared. Mix and store in a polyethylene bottle.1292 Analyst, Vol. 106 Procedure for Phosphate (aqueous method) (eluent A ) .Transfer by pipette aliquots of standard orthophosphate solution containing 5-1 500 pg of orthophosphate into a series of 50-ml calibrated flasks. To each add 5 ml of acidic molybdate solution and dilute to about 20 ml with water. After allowing to stand for 15 min, dilute to 50 ml with water. Procedure for phosphate (molybdovanadate method) (eluent B) . Described above. Use 1.0 ml of molybdovanadate reagent for the lom5 and M ranges: in this work, 7.5 ml were used for the M range. Procedure for phosphate (aqueous acetone method) (eluent C). Place 26 ml of acetone and 10 ml of sodium molybdate - tartaric acid - hydrochloric acid solution in a 50-ml calibrated flask, add the sample solution containing 0.05-1.5 mg of immediately dilute to volume and mix. Place 10 ml of the ammonium molybdate - sulphuric acid solution and 5 ml of acetone in a 50-ml calibrated flask. Add the sample solution containing 0.3 pg-1.2 mg of silica.Rinse down the inside of the neck of the flask, mix the solution and allow it to stand for 15min. Add 10ml of 10% m/V mannitol solution, dilute to volume and mix. Leave for 15 min. Pipette an aliquot of sample solution containing 0.061.5 mg of arsenic(V) into a 50-ml calibrated flask. Add 7 ml of 4 M sulphuric acid, 7 ml of 8% ammonium molybdate solution and 16 ml of acetone, dilute to volume with water and mix. In a 50-ml calibrated flask place 6 ml of ammonium molybdate solution, 10 ml of acetone and 3 ml of 1 M sulphuric acid. Add an aliquot of sample solution containing 0.4-1.5mg of germanium, dilute to volume with water, mix and allow to stand for 5 min.Mix 5.0 ml of sodium molyb- date -sulphuric acid solution and 10ml of ethanol in a 50-ml calibrated flask. Add by pipette an aliquot of phosphate solution containing 5-1 500 pg of phosphate, mix and dilute to 50 ml with water. FOGG AND BSEBSU : DP VOLTAMMETRIC DETERMINATION OF Procedure for silicate (aqueous acetone method) (eluenj D). Procedure for arsenic (aqzceous acetone method) (eluent E). Leave for 30 min. Procedure for germanium (aqueous acetone method) (eluent F ) . Procedure for Phosphate (aqueous ethanol method) (eluent G). Assessment of eluents Pre-formed heteropolyacid solutions prepared as described above were injected into the de-gassed eluent, which had the same composition as the blank.Linear-sweep voltammo- grams obtained in a static system give an indication of suitable potentials to apply to the detector cell. Voltammograms obtained by scanning in the direction of positive potential are generally a combination of cathodic and anodic waves, and for this reason voltammograms obtained by scanning in the direction of negative potential, starting at a sufficiently positive potential, are a more reliable guide. Two main methods of choosing suitable potentials were adopted here using static and flowing systems and these gave essentially the same results. With the static system, the currents obtained when a clean, glassy carbon electrode was placed in the appropriate heteropolyacid solution and the cell closed at various potentials were determined.The electrode was cleaned between measurements and the currents obtained were plotted against potential. With the flowing systems, a fixed volume of hetero- polyacid solution was injected and the current noted. Generally, three successive injections were made at each potential in order to check that the signal was constant and that no loss of signal was occurring due to contamination of the electrode surface. For germanomolybdate solutions at positive potentials of less than +0.12 V, loss of signal was observed: in this instance the detector cell was dismantled and the electrode cleaned after the third injection before studying the current signals at other potentials. Results Graphs of current obtained at the detector cell for the injection of 4 x l O U 4 ~ solutions of the heteropolyacids into the various eluents are shown in Fig.3. These graphs can be used to select suitable potentials to apply to the glassy carbon electrode when making deter- minations in a particular eluent. Eluent B (molybdovanadate) and eluent G (aqueous ethanol for phosphate) give distinct plateaux in the potential ranges 0.25-0.37 V and 0.19- 0.25 V, respectively. Highly precise results have been obtained with all systems at a wide range of potentials, although adsorption remained a problem with the germanomolybdateDecember, 1981 1293 system. The graphs shown in Fig. 3 indicate the relative sensitivities obtained with the various eluents. PHOSPHATE AND ASSESSMENT OF ELUENTS FOR FI VOLTAMMETRY +0.2 +0.4- PotentialN +0.6 +0.2 +0.4 Potent ia l/V Fig.3. Currents obtained a t various potentials for flow injection of 100 p1 of pre-formed heteropolyacids (4 x Plots are lettered according to the eluent as in the text and Table I, vzz.: A, aqueous phosphate; B, molyb- dovanadate; C, aqueous acetone for phosphate; D, aqueous acetone for silicate; E, aqueous acetone for arsenate; F, aqueous acetone for germanate; and G, aqueous ethanol. M) into eluents consisting of +he reagent blank. The rectilinearity of calibration graphs at various concentrations is indicated in Table I. The aqueous acetone eluent for silicate is particularly sensitive, and useful signals were obtained at the ~O-'M level (see Fig. 4). The effect of overloading at the ~ O - * M level of silica, and the otherwise good rectilinearity with good base lines, can be seen clearly in this example.For phosphate determinations the molybdovanadate, aqueous and aqueous ethanolic eluents gave useful signals at the Signals obtained with the molyb- dovanadate eluent are shown in Fig. 5. M level. TABLE I CHARACTERISTICS OF CALIBRATION GRAPHS IN VARIOUS CONCENTRATION RANGES Concentration of pre-formed determinand A 7 Eluent 0.5-4 X w 4 M 0.5-4 X w 5 M 0.5-4 X 1Ow6M 0.5-4 X 10-'M A, Aqueous phosphate B, Vanadomolybdate . . C, Aqueous acetone for phosphate . . D. Aqueous acetone for silicate . . ., E, Aqueous acetone for arsenate . . . . I?, Aqueous acetone for germanate . . G, Aqueous ethanol for phosphate . . . . Non-rectilinear Rectilinear Rectilinear -* .. Slightly Rectilinear Rectilinear -* non-rectilinear ..Slightly Rectilinear -* non-rectilinear -* . . Non-rectilinear Rectilinear Rectilinear Rectilinear Slightly Rectilinear -* non-rectilinear -* . . Non-rectilinear? Rectilinear -* -* . . Rectilinear Rectilinear Rectilinear -* * Base line deteriorates. t Signal loss due to adsorption. Linear sweep voltammograms run previously on the blank solutions in static systems1J indicate that the background currents associated with the eluents are not insignificant, despite their being effectively Compensated for in the present procedures. In Fig. 6, the1294 FOGG AND BSEBSU: DP VOLTAMMETRIC DETERMINATION OF Analyst, VoZ. 106 Fig. 4. Flow injection signals for determination of silicate at various concentrations: (a) A, 4 and B, 2 x lo-’ M ; (b) A, 4 and B, 2 x 1W6 M ; (c) A, 4 and B, 2 x M.Glassy carbon electrode held a t $0.18 V. M ; and (d) A, 4 and B, 2 x signals obtained on injecting the blank reagents (the normal eluents) into 0.005 M sulphuric acid eluent are shown. The molybdovanadate, aqueous acetone for arsenate and aqueous acetone for phosphate eluents give significant background currents over the potential ranges used for measurement. Where zero current is indicated in Fig. 6 for the other eluents, the base line did not necessarily remain undisturbed on making an injection but often consisted of a small combined negative and positive signal resembling a first derivative trace. There- fore, at the present stage of our studies it is recommended that for all systems used here injection is made into the blank reagent and not into 0.005 M sulphuric acid.After approximately 6 months of work with the original glassy carbon electrode it was found that the blank differential-pulse voltammogram for molybdovanadate had a steeper slope than that shown in Fig. 2. This deterioration of the surface of the electrode was not as apparent when the electrode was used in the flow system, but it may be advisable to change the electrode when this occurs. It may be possible to regenerate the electrode by polishing with a fine diamond paste. 0.1 PA T c.. P 3 0 b’ I’ pLA A C) A 10 pA f---- Time Fig. 5 . Flow injection signals for determination of phosphate a t various concentrations using molyb- dovanadate as the eluent: (a) A, 4 and B, 2 x 10-6 M ; (b) A, 4 and B, 2 x 10-5 M ; and (c) A, 4 and B, 2 x 1 0 - 4 ~ .Glassy carbon electrode held at $0.32 V. Discussion The solution conditions of the colorimetric molybdovanadate method for determining phosphate were shown to be applicable directly to the differential-pulse voltammetric deter- mination of phosphate a t a glassy carbon electrode in a static system, and a recommendedDecewzber, 1981 1295 procedure is given for obtaining voltammograms by scanning in the negative-potential direction. Determination can also be made at a glassy carbon electrode under flow conditions by injecting pre-formed molybdovanadophosphate solution into the reagent blank serving as eluent. Blank solutions used in static voltammetric methods for phosphate, silicate, arsenate and germanate reported previously were also shown to be useful eluents in flow-injection met hods.The molybdovanadate, aqueous and aqueous ethanol methods proved to be equally sensi- tive for determination of phosphate on adaptation to flow injection procedures and the results obtained were extremely reproducible and precise despite the high background currents associated particularly with the molybdovanadate reagent. The methods are reliable at the M level. The aqueous acetone reagent is particularly sensitive for silicate, satisfactory measurements being made at the lo-’ M level. PHOSPHATE AND ASSESSMENT OF ELUENTS FOR FI VOLTAMMETRY ~LI level and coefficients of variation are typically less than 1% a t the 4-0.2 +0.4 Po tent i a IN +0.2 +0.4 PotentialiV Fig. 6. Currents obtained a t various potentials for flow injection of 100 p1 of reagent Plots are blanks (Z.e., the usual eluents) into an eluent consisting of 0.005 M sulphuric acid.lettered according to the eluent as in Fig. 3. These studies have been restricted mainly to the injection of pre-formed heteropolyacids into the reagent blanks. Fujinaga et aZ.3 injected phosphate solutions directly into aqueous ethanolic reagent, in which rapid formation of molybdophosphate occurs, and were able to determine 1 x Injection of pre-formed heteropolyacid solutions, although not as convenient, is more reliable and gives sharper peaks and lower detection limits. A brief study has been made by us of the injection of phosphate solutions directly into molyb- dovanadate eluent: after passing through 50 cm of tubing (1.58 mm i.d.) after injection, peaks similar to those obtained by Fujinaga et aZ.3 were obtained. Injection of water, however, gave a negative signal owing to the background current of the eluent. This negative signal is apparent in the signal traces used as illustrations by Fujinaga et aZ.3 More detailed studies are being made of the direct injection of phosphate, silicate, arsenate and germanate solutions into reagent streams and of phosphate reagents into a phosphate stream. This latter study is of interest for on-line determination of phosphate, in hydroponic fluid, for example. The choice of eluent composition and potential applied to the detector will be important in reducing the background current levels whilst maintaining good sensitivity. One of us (N.K.B.) thanks the people of the Socialist People’s Libyan Arab Jamahiriya for financial support and leave of absence from El-Fatah University, Tripoli. M phosphate. References 1. 2 . 3. 4. 5. Fogg, A. G., and Bsebsu, N. K., Analyst, 1981, 106, 369. Fogg, A. G., Bsebsu, N. I<., and Birch, B. J., TaZanZa, 1981, 28, 473. Fujinaga, T., Okazaki, S., and Hori, T., Bunseki Kagaku, 1980, 29, 367. Fogg, A. G., and Bhanot, D., ,4nalyst, 1981, 106, 883. \Villiams, \V. J., “Handbook of Anion Determination,” Butterworths, London, 1979, p. 469. Received March lath, 1981 Accepted June 22nd, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601288

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

1296 Analyst, December, 1981, Vol. 106, pp. 1296-1301 Use of Ascorbic and Thioglycollic Acids to Eliminate Interference from Iron in the Aluminon Method for Determining Aluminium F. Cabrera, L. Madrid and P. de Arambarri Centro de Edafologia y Biologia Aplicada del Cuarto, Apartado 1052, Seville, Spain The use of ascorbic and thioglycollic acids as inhibitors for the interference of iron in the aluminon method of Hsu have been examined. The use of ascorbic acid, as proposed by Jayman and Sivasubramaniam, has been found to change iron interference from positive to negative causing aluminium to be underestimated. However, the addition of 0.2 ml of a 1% V/V solution of thioglycollic acid to solutions containing aluminium in amounts ranging from 10 to 50 pg has been proved to suppress the interference from up to 900 pg of iron.Keywords : Aluminium determination ; aluminon method ; ascorbic acid ; iron interference ; thioglycollic acid The determination of aluminium in soil and clay extracts is a common problem in soil science.lM3 Aluminon reagent has been used extensively for the spectropliotometric deter- mination of small amounts of aluminium, although iron introduces an interference. One of the aluminon methods more widely used is that described by Hsu,~ according to which the colour developed is stable over a long period of time and yields reproducible results. Jayman and Sivasubramanianl tested several inhibitors to eliminate the iron interference from Hsu’s m e t h ~ d . ~ They claim that adding 1 ml of a 0.5% solution of ascorbic acid (i.e., 5000 pg of ascorbic acid) suppresses the interference from up to 500 pg of iron(II1).They also report that the use of thioglycollic acid eliminates this interference but bleaches the colour of the aluminon reagent and causes a reduction in sensitivity. When known amounts of aluminium were determined in solutions containing iron by the method proposed by Jayman and Sivasubramaniam,l we found less aluminium than expected. In this work we studied the influence of different amounts of iron(III), iron(I1) and ascorbic acid on the determination of aluminium by Hsu’s m e t h ~ d . ~ The use of thioglycollic acid instead of ascorbic acid was also investigated. Experiment a1 Reagents Dilute 120 ml of glacial acetic acid to about 900 ml with distilled water, add 24 g of sodium hydroxide, mix and dissolve 0.35 g of aluminon in the resulting solution; dilute to 1000 ml with distilled water and mix.The pH of this solution should be 4.2. When 10 ml of this reagent are added to a test solution containing 2.0-4.0 ml of 1 N hydrochloric acid, dilution with water to 50 ml gives a final pH of 3.74.0. The final pH values of all the experiments described throughout this paper were within this range. Standard aluminium solution. Dissolve 0.175 8 g of potassium aluminium sulphate (potassium alum) in distilled water and dilute to 1000 ml. This solution contains 10 pg nil-1 of aluminium. Dissolve 0.50 g of ascorbic acid in distilled water and make the volume up to 100 ml. This solution contains 5000 pg ml-1 of ascorbic acid.Dilute 1 ml of thioglycollic acid to 100 ml with distilled water. Aluminon - acetate bufer solutiofi. Ascorbic acid solution, 0.50% m/V. Thioglycollic acid solution, 1% V/V. Methods Calibration of standard graph as outlined by Hsu4 Add 2.0 ml of 1 N hydrochloric acid to each flask and place them on a steam-bath for 30 min. Add aliquots containing from 0 to 50pg of aluminium to 50-ml calibrated flasks.CABRERA, MADRID AND DE ARAMBARRI 1297 Cool and dilute the solutions to about 35 ml with distilled water, add exactly 10 ml of the aluminon - acetate buffer solution, make the contents up to the mark with distilled water, mix and allow the solutions to stand for at least 2 h. Measure the colours obtained against the reagent blank in l-cm cells at 530 nm. Calibration of standard graph as described by Jayman and Sivasubramaniaml adding the 2 ml of 1 N hydrochloric acid. conditions as above against a reagent blank containing the same amount of ascorbic acid.Proceed as above except that 1 ml of ascorbic acid solution is added to each flask before Measure the colours obtained under the same Calibration of standard graph .using thioglycollic acid as inhibitor for iron interferences acid solution instead of 1 ml of ascorbic acid solution. Proceed as for Jayman and Sivasubramaniam’s methodl but add 0.2 ml of thioglycollic Results and Discussion Ascorbic Acid as an Inhibitor for Iron Interference Jayman and Sivasubramaniam’s methodl was checked by analysing solutions containing known amounts of aluminium (10 and 30 pg) and containing increasing amounts of iron(II1) or iron(I1) (30, 100 and 500pg).Table I shows that aluminium recoveries were always lower than expected and independent of the form of iron present. As ascorbic acid reduces iron(II1) to iron(II), the incomplete recovery of aluminium seems to be related to the presence of iron(I1). TABLE I ALUMINIUM RECOVERED FROM SAMPLES CONTAINING MIXTURES OF ALUMINIUM PLUS IRON(III) OR IRON(II) BY THE METHOD OF JAYMAN AND SIVASUBRAMANIAM Aluminium present/ v g per 50 ml 10 30 10 30 10 30 Iron(II1) interference Iron(II1) present1 Aluminium recovered/ 30 3.9 30 19.7 100 4.9 100 23.4 500 6.6 500 27.1 I A \ vg per 50 ml yg per 50 ml Iron( 11) interference Iron( 11) present/ Aluminium recovered/ yg per 50 ml Vg per 50 ml 30 3.6 30 22.2 100 3.8 100 22.4 500 5.0 500 27.3 0.6 1 m -f2 0.3 Q z 0 0.2 c I I I 8 I 0 1000 2000 3000 4000 5000 Ascorbic acid concentrationipg per 50 ml Fig.1. Influence of ascorbic acid on the absorbance of solutions containing : A, aluminon alone; B, aluminon plus 30 pg of iron(II1) ; C , aluminon plus 30 pg of aluminium; and D, aluminon plus 30 pg of aluminium and 30 pg of iron(II1). Readings obtained against water.1298 CABRERA et al. : ASCORBIC AND THIOGLYCOLLIC ACIDS TO Analyst, VoZ. 106 From these results, it was decided to study the absorbance of solutions containing varying amounts of aluminium and iron and increasing amounts of ascorbic acid (up to 5000pg). The colour was developed as outlined by Hsu.~ Fig. 1, line A, shows that ascorbic acid produces a very slight bleaching of the colour of the aluminon reagent when neither aluminium nor iron is present.A similar effect is found when 1000 pg of ascorbic acid or more are added to solutions containing 30 pg of aluminium (Fig. 1, line C). Fig. 1, line B, shows that small additions of ascorbic acid (up to 50 pg) decrease the absorbance of solutions without aluminium but containing 30 pg of iron(III), but larger additions produce a definite maximum at about 500 pg of ascorbic acid, after which the absorbance decreases to a constant value (from 3000pg of ascorbicacid onwards) corresponding to the absorbance of the solution without iron. Fig. 1, lines A and B, thus indicate that 3000-5000 pg of ascorbic acid suppress the effect of 30 pg of iron(II1) on the colour of the aluminon reagent when no aluminium is present.Fig. 1, line D, shows the effect of additions of ascorbic acid to solutions containing 30pg of aluminium and 30 pg of iron(II1). The general trend of the graph is similar to that of Fig. 1, line B, but the constant value of the absorbance reached at about 3000 pg of ascorbic acid is clearly below that corresponding to 30 pg of aluminium with no iron present (Fig. 1, line C). Therefore, addition of 5000pg of ascorbic acid (the amount recommended by Jayman and Sivasubramaniaml) suppresses a positive interference of 30 pg of iron(III), but at the same time introduces a negative interference. This result is confirmed by the results in Table I. The negative interference described above has been verified in samples containing increasing amounts of aluminium (from 10 to 50 pg) and 30 pg of iron(II1) or iron(I1).Fig. 2, lines A and B, show that standard graphs obtained by Hsu’s method4 (i.e., without ascorbic acid) and by Jayman and Sivasubramaniam’s method1 (i.e., with 5000 pg of ascorbic acid) are almost identical. The addition of 30 pg of iron(II1) to samples similar to those in Fig. 2, line A (i.e., without ascorbic acid) produces higher values of absorbance (Fig. 2, line C). However, when iron is added as iron(I1) (Fig. 2, line E) the absorbance 0.6 r / c 0.5 0.4 0 l-5 .“2 0.3 2 0.2 0.1 0 10 20 30 40 50 Aluminium concentration/yg per 50 ml Fig. 2. Interference of iron(II1) or iron(J.1) when aluminium is determined by Hsu’s and Jayman and Sivasubramaniam’s methods : A, Hsu’s method without iron ; B, Jayman and Sivasubramaniam’s method with- out iron; C, Hsu’s method with 30 pg of iron(II1); D, Jayman and Sivasubramaniam’s method with 30 pg of iron(II1); E, Hsu’s method with 30 pg of iron(I1); and F, Jayman and Sivasubramaniam’s method with 30 p g of iron (11).December, 1981 1299 values are the same as those of Fig.2, line A, which means that 30pg of iron(I1) do not interfere in the determination of aluminium by Hsu’s m e t h ~ d . ~ However, the addition of 30pg of iron(II1) or iron(I1) (Fig. 2, lines D and F) to samples similar to those of Fig. 2, line B (i.e., with 5000 pg of ascorbic acid) decreases the absorbance values. I t can thus be concluded that the negative interference observed appears when aluminium, iron(I1) and ascorbic acid are present simultaneously, as the amount of ascorbic acid present is more than that necessary to reduce all the iron(II1) added.Fig. 3 shows the effect of increasing amounts of iron(I1) (up to 500 pg) on the absorbance of samples with 0 and 30 pg of aluminium and 5000 pg of ascorbic acid (z.e., Jayman and Sivasubramaniam’s method1). Absorbance values of samples without aluminium are independent of the amount of iron present (Fig. 3, line A). However, when 30pg of aluminium are present (Fig. 3, line B) absorbance values are always lower than that corre- sponding to when no iron is present (starting point of Fig. 3, line B). Therefore, the negative interference reported in this paper manifests itself for a wide range of concentrations of iron in the system. ELIMINATE INTERFERENCE FROM IRON IN THE ALUMINON METHOD 0.1 0.5 i 0 o-0-0 8 U A - “ 0.1 ‘ & I 1 0 100 200 300 400 500 lron(ll1 concentrationlpg per 50 ml Influence of iron(I1) on the colour developed by the method of Jayman and Sivasubramaniam: A, without aluminium; and B, with 30 p g of aluminium.Readings obtained against water. Fig. 3. When absorption spectra were obtained for all the solutions described above, results showed that the observed negative interference of iron with ascorbic acid at 530 nm is not due to a shift of the maximum of the spectrum but to a decrease in the actual value of such a maximum; that is, the wavelength of absorbance of the colour of the aluminon - aluminium mixture is not changed when iron and ascorbic acid are present, but the intensity is decreased.I I 0 0.5 1 .o 1.5 2.0 1 Volume of 1% VIV thioglycollic acid solution/ml Fig. 4. Effect of thioglycollic acid on the colour developed by the method of Hsu: A, B, C, D and E, 10, 20, 30, 40 and 50 pg of aluminium, respectively.1300 CABRERA et a,?. : ASCORBIC AND THIOGLYCOLLIC ACIDS TO Analyst, VoZ. 106 0.5 0.4 s -2 s n a 0.2 0.3 0.1 L 1 I I I 0 10 20 30 40 50 Aluminium concentration/pg per 50 ml Fig. 5. Loss of sensitivity (expressed as absorbance units per microgram of aluminium) due to the use of different volumes of 1 yo Y / I' thioglycollic acid solution : A, 0 ml, slope 0.007 9; B, 0.2 ml, slope 0.007 6; C, 0.5 ml, slope 0.0072; and D, 2.0 ml, slope 0.0060. Thioglycollic Acid as an Inhibitor for Iron Interference Thioglycollic acid has been used widely to suppress iron interference when determining aluminium by spect rophotometric methods using aluminon.5 9 6 Increasing amounts of 1% V/V thioglycollic acid solution (up to 2 ml) were added to samples containing known amounts of aluminium (from 10 to 50 pg) and the colour developed as described by Hsu.* Addition of up to 0.5 ml of thioglycollic acid solution causes little, if any, changes in the absorbance of samples containing 10, 20 or 30 pg of aluminium (Fig. 4, lines A, B and C). For higher aluminium contents (40-50 pg) the absorbance decreases as the volume of the added thioglycollic acid increases (Fig. 4, lines D and E). In all instances, however, additions of greater than 0.5 ml bleach the colour significantly, particularly for the larger aluminium contents. TABLE I1 EFFECT OF DIFFERENT VOLUMES OF THIOGLYCOLLIC ACID SOLUTION ON THE PLUS IRON(III) BY THE METHOD OF Hsu RECOVERY OF ALUMINIUM FROM SAMPLES CONTAINING MIXTURES OF ALUMINIUM Aluminium recoveredlpg per 50 ml Aluminium present/ p g per 50 ml 20 40 20 40 20 40 20 40 20 40 20 40 20 40 Iron(II1) present/ p g per 50 ml 30 30 100 100 300 300 600 600 700 700 800 800 900 900 I 0.2 ml thioglycollic 0.3 ml thioglycollic acid added* acid added* 20.4 20.2 40.9 40.1 20.6 20.0 40.3 39.8 19.9 18.3 40.1 39.3 19.7 18.1 38.9 36.7 21.8 - 39.4 - 20.8 c_ 39.8 - 20.5 16.5 40.2 36.1 1 0.5 ml thioglycollic acid added* 20.3 40.4 20.3 40.1 18.6 39.0 18.3 36.5 - - - - 16.2 36.2 * A 1% V/V thioglycollic acid solution.December, 1981 1301 Fig.5 shows that the slopes of standard graphs obtained using different amounts of thio- glycollic acid solution (from 0 to 2 ml) decrease from 0.0079 to 0.0060 (expressed as absorbance units per microgram of aluminium).Jayman and Sivasubramaniaml found a much larger loss of sensitivity, about 50% for the addition of 1 ml of thioglycollic acid. Table I1 shows aluminium recoveries from solutions containing 20 and 40 pg of aluminium and up to 900 pg of iron(III), using 0.2, 0.3 and 0.5ml of thioglycollic acid solution to suppress the iron interference. It can be seen that 0.2 d gives good recoveries for aluminium, better than 0.3 or 0.5 ml; 0.2 ml also causes little loss of sensitivity (about 3%, see Fig. 5). These findings lead us to recommend the use of 0.2 ml of thioglycollic acid solution to suppress iron interference in the aluminon method for the determination of aluminium. ELIMINATE INTERFERENCE FROM IRON IN THE ALUMINON METHOD Conclusions From the preceding discussion it can be concluded that Jayman and Sivasubramaniam’s method1 underestimates aluminium contents, as 5 000 pg of ascorbic acid change the positive interference from iron to a negative interference. However, this negative interference only appears when aluminium, iron and ascorbic acid are present simultaneously. Substitution of 0.2 ml of 1% V/V thioglycollic acid solution for the 5000 pg of ascorbic. acid recommended by Jayman and Sivasubramaniaml eliminates interference from up to 900 pg of iron and gives good, reproducible aluminium recoveries. References 1. 2. 3. 4. 6. 6. Jayman, T. C. Z., and Sivasubramaniam, S., AnaZyst, 1974, 99, 296. Cabrera, F., and Talibudeen, O., J . Soil Sci., 1977, 28, 259. Cabrera, F., and Talibudeen, O., CZays Clay Miner., 1978, 26, 434. Hsu, P . H . , Soil Sci., 1963, 96, 230. Robertson, G., J . Sci. Food Agric., 1950, 1, 59. Chenery, E. M., Analyst, 1948, 73, 501. Received September 9th, 1980 Accepted June 24th, 1981

ISSN:0003-2654    

DOI:10.1039/AN9810601296

出版商:RSC    

年代:1981

数据来源: RSC