ISSN:0003-2654    

DOI:10.1039/AN98005FX005

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BX007

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005FP013

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BP019

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

FEBRUARY 1980 The Analyst Vol. 105 No. 1247 Examination of Petroleum Products of High Relative Molecular Mass for Forensic Purposes by Synchronous Fluorescence Spectroscopy Part 1. Appraisal of Experimental Factors J. B. F. Lloyd Home Ofice Forensic Science Laboratory, Gooch Street North, Birmingham, B5 SQQ The luminescence of petroleum derivatives is discussed with reference to a standard technique for their differentiation based on synchronous fluorescence spectroscopy. The spectra are highly sensitive to quenching by oxygen and by halogenated compounds ; for general-purpose use, solutions of samples (in hydrocarbon solvents) must be deoxygenated. Although the quantum yield of the fluorescence is of the order of 0.15, the fluorescence is distributed over a wide wavelength range and, at the sample concentrations usually required in order to obtain satisfactory spectra, significant inner filter effects are often present and corrections must be made for these.Various spectral features that differ between samples are defined. After correction for instrumental and inner filter effects, the features are expressed numerically for evaluation and use in the pattern recognition study that will be described elsewhere in Part 11. Examples of other types of spectra are considered in relation to the standard technique. These are synchronous and conventional fluorescence contour diagrams, synchronous derivative spectra, low-temperature syn- chronous fluorescence spectra and phosphorescence spectra. Keywords : Fluorescence ; phosphorescence ; synchronous fluorescence spectro- scopy ; petroleum products ; forensic analysis Since the work of Parker and Barnes1 an appreciable number of uses of fluorescence spectro- scopy has been made in the identification and determination of petroleum oils and related materials, mainly in the fields of oil pollution and forensic science.However, because the optical characteristics of the various types of spectrofluorimeter employed have varied greatly in the extent to which the spectra they produce are distorted (some examples are given in a recent paper by Eastwood et d2), and because the fluorescence characteristics of the highly complex mixtures present in petroleums are still not well understood, the various published methods of analysis have produced results that are of necessarily restricted appli- cation.Although the choice of analytical methods remains an essentially empirical undertaking , instrumentation that produces relatively undistor ted spectra is now widely available commercially. Hence, a satisfactory degree of star da .disation has become possible for the compilation of spectral data that are sufficiently reFrolucible and stable with time for their use in the classification of unknown samples and in the calculation of evidential significance, e.g., by pattern recognition techniques. This paper reviews the experimental variables pertinent to these materials that must be defined and controlled before any attempt can be made to establish satisfactory collections of data. This will be considered in subsequent work.Experimental Instrumental Details The spectra are recorded with a standard Perkin-Elmer MPF-4 spectrofluorimeter fitted with a corrected spectra accessory that, together with the various instrumental settings for wavelength and sensi tivity, is calibrated essentially according to the manufacturer’s manuals. 97 Crown Copyright.98 LLOYD : EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, Vol. I05 Both a standard lamp and a quantum counter have been used for the emission correction, without significant differences between the results. Except where indicated otherwise, the spectra shown here are corrected on the excitation side and, therefore, represent the relative excitation efficiency versus wavelength in terms of photon units. The effect of the relatively smaller emission correction on an already excitation- corrected synchronous spectrum is indicated in Fig.1 (c). The derivative spectra are obtained by electronic differentiation of the signal from the fluorimeter with an attachment made by Mr. D. A. Collins. A similar device was described by Green and O’Ha~er.~ Samples are presented to the fluorimeter, for perpendicular illumination, in a 5-mm square- sectioned Spectrosil cell (Starna Ltd., London), in which they can be deoxygenated by a stream of nitrogen passed into and out of the cell through 0.25 mm i.d. steel tubing sealed into an air-tight cover.* Provided that the inlet tubing, by means of which both sample solutions and nitrogen are introduced into the cell, is run down a junction of two cell faces, the spectra obtained should be independent of whichever face is set in the excitation beam.Ultraviolet absorption spectra are recorded with a Pye Unicam SP8000 spectrometer, the wavelength and absorbance calibrations of which are periodically checked with, respectively, a holmium oxide filter and a potassium dichromate solution in aqueous sulphuric acid.5 A Cahn Gram Electrobalance is used in the preparation of solutions of known concentra- tions of oil. Samples to be weighed are transferred from a needle point to a 1 x 10mm coil of 0.1-mm steel wire hung on one arm of the balance. Glassware Glassware with which samples and solvents come into contact is soaked in Decon 90, rinsed with distilled water and baked out overnight at a temperature of not less than 250 “C.I I I 290 330 370 b 1 g 11,15 290 330 370 Synchronous excitation wavelength/nm I I I 290 330 370 Fig. 1. Examples of synchronous fluorescence spectra obtained a t scanning intervals of (a) 34, (b) 20 and (c) 7 nm from solutions of three different oils (20 p g ml-1) in deoxygenated cyclohexane. The numbers on the spectra identify the featues defined in Table I. The full-line spectra are corrected for variation in the instrumental photon response on the excitation side. The broken-line spectra are both excitation- and emission-corrected. The solvent base line is shown in each instance.February, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 99 Lower temperatures or shorter times are ineffective in the removal of trace amounts of fluorescent contaminants.Materials Nitrogen (BOC Ltd., “oxygen-free”) is passed through a column packed with manganese(I1) oxide on Kieselguhr6 to remove trace amounts of oxygen that may have permeated the supply lines, and then through a pre-saturator containing the relevant solvent. The hydrocarbon solvents are of commercial spectrophotometric grades. If, as is usual, extraneous fluorescence is present at an intensity greater than 20% of the Raman C-H signal synchronously excited at 34 nm,’ the solvent is passed through a column of charcoal (Hopkin and Williams Ltd., “for gas absorption”) activated at 400 “C, and then distilled under nitrogen (not pre-saturated). I am indebted to Mr. C. A. Watson of Hopkin and Williams Ltd. for this technique. The 9,lO-diphenylanthracene (Koch-Light Laboratories Ltd., scintillation grade ; melting- point 250-252 “C, literature value8 245-247 “C) contained (as examined by high-performance liquid chromatography) no fluorescent impurity at a level greater than 1 part per lo4 a t excitation and emission wavelengths of 266 and 425 nm, respectively. An impurity giving a weak absorbance peak at 254nm was eliminated when the sample was evacuated at 0.1 Torr for 30 min at 100 “C. Quantum Yields These are determined essentially according to Parkerg from measurements at 260 nm on deoxygenated cyclohexane solutions, with reference to a value of 0.86 for 9,lO-diphenyl- anthracene in the same solvent.1° Standard Comparison Procedure The weighed sample is dissolved in cyclohexane to give a concentration of 20 pgml-1, and the ultraviolet absorption spectrum recorded at a 1-cm path length.If necessary, the solution is diluted until its absorbance does not exceed 0.10, with reference to a solvent blank, at wavelengths greater than 260nm. For weakly absorbing samples more con- centrated solutions may be needed, but this event is unusual. The solutions are photo- labi1e,l1~l2 and should therefore be stored in darkness and used within about 3 h. Spectral changes occur even in refrigerated samples on prolonged storage. The solution is transferred into the 5-mm spectrofluorimeter cell and purged with solvent pre-saturated nitrogen until there is no further increase in the fluorescence monitored with excitation and emission wavelengths of 286 and 310 nm, respectively. With a nitrogen flow-rate of 10 ml min-l the purging is usually complete within 1-2 min.Three synchronous fluorescence spectra are now recorded, all at the same sensitivity and in the corrected excitation mode , with separations between excitation and emission wavelengths of 34, 20 and 7 nm, and with both excitation and emission half-band widths of 3 nm. The excitation wavelength range of each spectrum is 230430 nm. Numerical Representation of Spectral Features The features, listed in Table I and in most instances indicated in the spectra in Fig. 1, are measured on the ultraviolet absorption and fluorescence spectra of the samples. Those features that are fluorescence intensity values are corrected for any contribution from the solvent blank, and for inner filter effects as described under Discussion.They are also corrected for variations from the true response of the emission detection side of the spectro- fluorimeter by multiplication with the ratio, at the relevant wavelength, of the response obtained when the spectrum of the xenon source lamp is synchronously scanned with the emission corrector accessory in operation to the response obtained without the corrector ; Fig. l(c) shows the effect of the correction. The intensity values (i.e., features 2-6, 9-14 and 16-20) are finally normalised to give a value for their sum of lo4. Discussion Spectrofluorimetric Parameters and Presentation alkylated aromatic hydrocarbons and sulphur heter0cyc1es.l~ The luminescence of petroleum products is due mainly to complex mixtures of heavily In some materials mixtures100 LLOYD: EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, VoZ.105 of unsubstituted or only lightly alkylated polynuclear aromatic hydrocarbons are also present in sufficient amounts to provide the basis of earlier fluorescence fingerprinting techniques,14 but the progress made in the chromatography of this group of compounds has made their characterisation, prior to separation, by luminescence techniques largely redundant. The total fluorescence of petroleums has been represented in three dimensions in the form of spectral contour diagrams by Freegarde et aZ.15 and by a number of other workers more recently. The diagrams are dominated by a common structure,lG however, and usually exhibit only a strong single or occasionally double peak, varying in wavelength between samples to some extent, together with various sub-features, which are largely obscured.An example from a crude oil is shown in Fig. 2. The relatively characterless appearance of the diagrams is reflected in the two-dimensional excitation or emission spectra more commonly used for sample comparisons. Feature 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 TABLE I DEFINITIONS OF SPECTRAL FEATURES Spectrum* uv D34 D34 D34 D34 D34 D34 D34 D20 D20 D20 D20 D20 D20 D20 D7 D7 D7 D7 D7 D34, UV D34, UV Definitiont (Ad, feature 2)/(sample conc., g ml-1) (F, max, before 282 nm) - (F, a t solvent base line) (F, max. 292-302 nm) - (F, min. 282-292 nm) (F, a t 313 nm) - (F, min. 302-308 nm) (F, a t 324 nm) - (F, at solvent base line) (Fo max.343-356 nm) - (F, min. 336-343 nm) Wavelength of feature 2 Wavelength of feature 6 (F, max. 266-277 nm) - (F, a t solvent base line) (Fo a t 278 nm) - (F, min. 278-283 nm) (Fo max. 283-298 nm) - (Fo min. 278-283 nm) (Fo max. 303-312 nm) - (F, min. 298-303 nm) (Fo max. 323-331 nm) - (Fo min. 316-323 nm) (Fo max. 379-385 nm) - (Fo min. 370-379 nm) Wavelength of feature 11 (F, max. 270-284 nm) - (F, a t solvent base line) (F, max. 298-312 nm) - (Fo min. 284-298 nm) (Fo max. 324-334 nm) - (F, min. 312-324 nm) (F, max. 351-359 nm) - (F, a t solvent base line) (F, max. 388-398 nm) - ( F , a t solvent base line) (Feature 2) (Ad, feature 2)-l (D34 solvent Raman)-] (Feature 1) (feature 21) x 10-8 * D34, D20 and D7 are synchronous fluorescence spectra with scanning intervals of 34, 20 and 7 nm.t The A d value used in the calculation of features 2 and 21 is from the inner filter correction of feature 2, as described in the text. Each F, value is a fluorescence intensity, corrected for inner filter and instrumental effects as described in the text, a t a specific point or a t a point of maximum or minimum intensity within the excitation wavelength range indicated. Increased differentiation between samples may be obtained from difference contour diagrams,12Js but the synchronous fluorescence technique1' presented in this form offers greater selectivity. Thus, as in Fig. 3, the contours representing fluorescence intensity can be plotted as a function of excitation wavelength (horizontal axis) and of the interval between emission and excitation wavelengths (vertical axis).This diagram has been con- structed from a set of synchronous fluorescence spectra covering the area in Fig. 2 within the pair of diagonal lines that indicate sections of the diagram corresponding to synchronous spectral emission - excitation intervals of 5 and 45 nm. Clearly, even with allowance for the greater number of spectra taken over the pertinent area, much more detail is apparent in the synchronous diagram. Although Talmi et aZ.18 and WeinerlS have observed that a two- dimensional synchronous spectrum represents a diagonal section through a conventional contour diagram, the increased detail seen in synchronous contour diagrams has not been pointed out before. As Vo-Dinh has noted,20 whether the synchronous spectra are considered as excitation or emission spectra is an arbitrary matter, because when either the excitation or the emissionFebruary, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 101 wavelength is specified, the other is defined by the scanning interval.I t is, however, important to specify as in Fig. 3 whether the spectra are being considered as excitation or emission spectra for the purpose of recording them, otherwise ambiguities arise. In this paper all of the synchronous spectra are plotted with reference to their excitation wavelengths. 430 E -. 5 390 CT, ClI ClI m C - 3 .: 350 E IA .- w 31 0 270 310 3 50 390 Fixed excitation wavelengthhm Fig. 2. Conventional fluorescence contour diagram of a crude oil in deoxygenated cyclohexane.The contours are at 5% intervals of intensity, increasing as the excitation wavelength decreases. The intensities are excitation-corrected only. The pair of diagonal lines indicate the sections represented in synchronous spectra with scanning intervals of 5 (lowermost) and 46 nm. Although rapid techniques are now available18s21s22 that could be developed for the com- puterised production of synchronous fluorescence contour diagrams or related forms of display, it is doubtful whether the construction of complete diagrams could be justified on every occasion when comparisons are to be made. From the considerable variety of samples that have been examined in the present connection, it is apparent that the same groups of fluorescent compounds, which vary only slightly in their peak wavelengths (presumably according to their degree and nature of alkylation), are always present even though their amounts are sometimes very small.Hence, comparisons between samples are essentially quantitative comparisons between the magnitude of, rather than the presence or absence of, specific features. Such comparisons are most easily made between two-dimensional spectra where intensities are displayed continuously rather than incrementally as in the synchronous contour diagrams. Most of the features evident in the diagrams can be collected economically in a small number of two-dimensional spectra obtained with a restricted number of scanning intervals, e.g., in the region of 10, 20 and 30 nm in Fig. 3. The intervals of 7, 20 and 34 nm actually chosen for routine use (see Experimental) are a compromise dictated by the require- ments of resolution, sensitivity and numbers of features found for a variety of samples.Although most of the features used are relative fluorescence intensities, the performance of some fluorescence wavelength and ultraviolet absorbance measurements as discriminators will also be examined elsewhere in Part II.23 All of the features are defined in Table I, and many of them are shown in the spectra in Fig. 1. Some further examples of the spectra, and their ability to discriminate between similar samples, are shown in Fig. 4. These spectra are from three samples of the same brand of a common motor oil, each of which is clearly102 LLOYD : EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, VoZ.105 differentiated from the other two (replicate spectra of each sample are superimposable on one another). When the fluorescent components of a mixture are adequately known, their synchronous fluorescence characteristics can be predicted, which avoids the development of analytical techniques likely to generate spurious data.24 The empirical approach adopted here, which is analagous to Weiner’sf9 except that the basis is a synchronous contour diagram, is necessi- tated by the still poorly known nature of petroleum fluorescence, and inevitably produces some redundant data, as the results obtained show.23 However, the results also show that features of importance are revealed at each of the three scanning intervals. Previous applications of the technique in this field have tended to be restricted to single scanning intervals.925--28 40 E 5 30 2 m *.’ C 0 .- .- : 20 m cn 10 270 310 350 390 Synchronous excitation wavelengthhm Synchronous contour diagram of the area between the diagonal lines in Fig. 2. Otherwise the experimental conditions and the solution are the same as in Fig. 2. The contours are a t 5% intervals of intensity, maximising a t the 280-nm region with intervals of about 40 nm. Fig. 3. Other Techniques Several other luminescence techniques are available by which further detail may be obtained. These include low-temperature luminescence spectroscopy and the presentation of spectra in a derivative form. The increased detail present in the fixed excitation and emission luminescence spectra of petroleums when recorded at 77 K rather than at ambient temperatures was reported by Parker29 and, more recently, by Fortier and Ea~twood.~* Their spectra show contributions due to both fluorescence and phosphorescence. Under these conditions more detail is obtained by the synchronous technique, as first reported for extender oil in tyre prints,31 and subsequently for heavy oils by Eastwood et aL2 However, for samples that are un- differentiated at ambient temperatures by the technique described under Experimental, the increased differentiation obtained at 77 K of samples commonly of interest in this work is usually small.For instance, the synchronous spectra of the three common motor oils of different manufacture shown in Fig. 5, which are not differentiated at ambient temperatures (broken lines spectra; a, b and c), remain undifferentiated at 77 K (full lines) despite the increases in detail both at this scanning interval (20 nm) and at others from which further detail is apparent, as in the 10-nm spectrum shown in Fig.5(4. At the usual synchronous scanning intervals (e.g., less than 100 nm) no phosphorescence is detected from these samples because of the typically large shift between their phosphorescenceFebruary, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 103 excitation and emission spectra. Larger intervals are generally required in phosphorimetric application^.^^ When, under the appropriate conditions, the phosphorescence or both the phosphorescence and the fluorescence are plotted under either fixed or synchronous con- ditions, still no differentiation between the samples in Fig.5 is obtained. Of these various techniques, the excitation spectrum of the total phosphorescence (passed through the emission monochromator set at zero order, and separated from the fluorescence and the scattered excitation by a rotating can phosphoroscope) exhibits the greatest number of features. A spectrum typical of the three samples in Fig. 5 is shown in Fig. 6 (broken line). Even a very different sample, such as that giving the full-line spectra in Fig. 4, is now poorly differentiated, as the spectrum in Fig. 6 (full line) shows. The corresponding phospho- rescence emission spectra compare closely to the spectrum of a motor oil given by Parker,29 and do not vary significantly between any of the samples used to obtain Figs.5 and 6. 270 310 350 390 270 310 350 390 270 310 350 390 Synchronous excitation wavelengthhrn Fig. 4. Synchronous fluorescence spectra a t intervals of (a) 34, (b) 20 and (c) 7 nm of a solution (20 pg ml-1) in de- oxygenated cyclohexane of each of three different samples of the same brand of an unused motor lubricant. The apparent inconsistency between these results and the value placed on low-temperature synchronous luminescence spectroscopy by Eastwood et aL2 must be due in part to differences in the nature of their samples. However, the inconsistency must also be partly due to the increased detail obtained at ambient temperatures in this work by the deoxygenation of the solutions (see below).If 77 K spectra are compared with the spectra of aerated solutions at ambient temperat~res,~O then part of the increased detail at 77 K is due to a reduction in oxygen quenching that is not seen if the point of reference, a t ambient temperatures, is a deoxygenated solution. Examples of the increased detail displayed by the first-derivative presentation of fixed excitation - emission spectra have been given by Green and O’Haver3 and Eastwood et aL2 Increased detail may be similarly obtained from synchronous spectra, as John and Soutar have ~uggested.~’ An example, in the second-derivative form, is given in Fig. 7, where features are now seen that are barely discernible in the original spectrum, shown by the104 LLOYD : EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, VoZ.105 - 280 340 \ \ 280 340 280 340 ‘. - 280 340 Synchronous excitation wavelengthhm Synchronous fluorescence spectra at ambient temperatures (broken line) and 77 K (full line) of three motor oils of different manufacture at a scanning interval of 20 nm [(a), (b) and (c)]. The samples are 20 pg ml-l in methylcyclohexane, which is deoxygenated for the ambient temperature spectra. Spectra (d) are from the solution used for (c) but scanned at 10 nm. The two solvent spectra [(a) and ( d ) ] show narrow peaks originating in the silica Dewar vessel used. Fig. 5. broken line. (Both spectra are uncorrected, to avoid a large contribution from the instru- ment’s spectrum corrector accessory to the background noise that is otherwise present in the derivative spectra.) However, it is again found that the technique does not significantly increase discrimination amongst samples not already discriminated by the standard tech- nique, because the increased detail given by the derivative presentation is not matched by a level of reproducibility appropriate to the close comparison of relative concentrations. In the future, this position might be changed by the application of the techniques recently described by Talsky et ~ 1 .~ ~ for the production of higher order derivatives of absorbance spectra by means of an analogue computer unit. Franks4 has introduced a “fluorescence maxima profile’’ technique, which presents the fluorescence of a sample as an envelope excited at all possible wavelengths. Some examples were also given by TanacredL35 Although the profiles may be of value in the classification of samples,35 the close comparison of samples by the technique is unlikely to afford much discriminat ion.Deoxygenation Effects Deoxygenation of a solution of a petroleum derivative results in a substantial increase in fluorescence inten~ity.~ The effect is illustrated in Fig. 8, where all of the spectra (of a solution of a motor oil) have been plotted at the same sensitivity. Deoxygenation (full-line spectra) not only increases the over-all intensity of the fluorescence, but affects the features present in the spectra differentially. Some features are barely detectable prior to the deoxygenation. The ratio of fluorescence intensity in aerated to deoxygenated cyclohexane solution for a variety of defined features (Table I, Fig.1) is listed in Table 11. The values quoted are from six different samples in whose spectra the features are strongly represented. The results emphasise that the maximum spectral detail cannot be obtained from DetroleumFebruary, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 105 samples at ambient temperatures without deoxygenation. Indeed, one feature (12) that is found to be a valuable di~criminator~~ is almost completely lost in aerated solutions. Although at 77 K the effect of oxygen quenching on the spectra is negligible, precise comparison of emission intensities between samples is less readily made and, as before, there is little increase in differentiation. In certain instances increased differentiation is made available by the selective application of quenching effects, but this is best reserved for specific circumstance^.^ In any event, quenching effects can only be assessed with reference to unquenched spectra.250 280 310 340 Fixed excitation wavelengthhrn Fig. 6. Phosphore- scence excitation spectra (uncorrected) a t 77 K of solutions in methylcyclo- hexane. The full-line spectrum is from the same sample as the full- line spectra in Fig. 4. The broken-line spectrum is from the same sample as the spectra in Fig. 5(a). i i I , , , , 270 310 350 Synchronous excitation wavelengthlnm Fig. 7. Synchronous fluorescence spectra (un- corrected) a t an interval of 20nm of a motor oil (broken line) and the corresponding second de- rivative spectrum (full line).Solvents The spectra are not significantly varied by the use of different saturated hydrocarbon solvents (hexane, 2,2,4-trimethylpentane, cyclohexane, methylcyclohexane and the petroleum fraction boiling at 40-60 "C) . Methyl~yclohexane~~ and the petroleum fraction form glasses at 77 K and are useful, therefore, for low-temperature spectroscopy. Cyclohexane is used in most of the present work, at ambient temperatures, mainly because of its availability at a reasonable cost in a relatively pure state. Its strong Raman peak does not interfere at the sensitivities generally used, but interfering fluorescent impurities are normally present, both TABLE I1 FLUORESCENCE INTENSITIES OF PETROLEUM PRODUCTS IN AERATED RELATIVE TO DEOXYGENATED CYCLOHEXANE Feature* 2 3 5 6 9 11 12 13 16 17 18 19 20 Meant .. 0.67 0.11 0.62 <0.05 0.53 0.84 0.09 0.55 0.51 0.73 0.67 0.66 0.67 Standard deviation . . 0.02 0.06 0.04 - 0.04 0.11 0.11 0.08 0.06 0.02 0.06 0.08 0.06 * Defined in Table I. t Six samples of different oils.106 LLOYD: EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, vd. 105 in this and in the other solvents, whatever their specified purity. The usual techniques of purification over silica gel, aluminium oxide and aluminium oxide loaded with silver nitrates have proved inferior to the activated charcoal technique (see Experimental). Unless quenching effects are to be deliberately employed, the presence of halogenated (except fluorinated) solvents is avoided, either for extraction of samples from substrates or for spectroscopy: the presence of 1% by volume of carbon tetrachloride in cyclohexane, for instance, reduces the emission intensity of petroleum products in the region of 300nm to approximately 10% of the unquenched value.Quenching and spectral modification are also induced by chloroform and, to some extent, by di~hloromethane,~ although Keizer and Gordon3' have reported otherwise. 270 310 350 390 430 270 310 350 390 A ' 0 310 350 390 Synchronous excitation wavelengttdnm Fig. 8. Synchronous fluorescence spectra (excitation-corrected) at intervals of (a) 34, (b) 20 and (c) 7 nm of a 20 pg ml-l solution of a motor oil in cyclohexane at ambient temperature. The dotted-line spectra are from an aerated solution; the full-line spectra are from a deoxygenated solution.All of the spectra were run at the same instrumental sensitivity. Concentration and Inner Filter Effects The ultraviolet absorbance spectra of petroleum products, of which some examples are shown in Fig. 9, generally rise gradually from the visible region to a plateau or weakly resolved peak at 250-265 nm. Thereafter, absorbances rise steeply, sometimes to give a peak at about 230 nm. The EiZm values at 260 nm found for the range of samples en- countered to date in forensic work vary between 5 and 500, and are useful discriminators. (They are, incidentally, strongly correlated with the refractive indices of the samples, which have also been used as discriminators.= Thus, for seven samples in which at 260 nm varies from 10 to 222, and 6'" varies from 1.474 to 1.513, the correlation coefficient is Many of the ultraviolet-absorbing compounds in these complex mixtures are not signifi- cantly fluorescent-evidently, from their phosphorescence, many undergo inter-system crossing to triplet states after excitation. Thus, the quantum yield of fluorescence emission (see Experimental) in the region of the excitation maximum at 260nm of a sample of a crude oil, an extender oil, a motor oil and a light lubricating oil, all in deoxygenated cyclo- hexane, are found to be 0.16, 0.19, 0.13 and 0.13, respectively.Further, the wavelength range of the fluorescence emission is widely dispersed with a width at half peak height corresponding to about 5000cm-l, e.g., 340410nm. For single compounds a value as great as 4000 cm-l is unusual. Hence, relatively strongly absorbing solutions must be used if a sample signal strength is to be obtained that is sufficiently in excess of solvent blanks and the spectrometer noise level for the reproducible measurement of spectral features.0.90.)February, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 107 At the concentrations used, commonly 20 pg ml-l (see Experimental), the absorbances of the solutions in the region of the shortest wavelength features may be as great as, e.g., 0.09 and 0.05 per centimetre of path length, at excitation and emission wavelengths of 270 and 290 nm, respectively. If the provisional assumption is made that the observed fluorescence intensity, F , at a given wavelength can be corrected for inner filter effects39 by the factor loAd, i.e., F, = 1OAdF, where Fo is the corrected intensity, A is the sum of the absorbances per centimetre of path length at the excitation and emission wavelengths and d is the cell depth at which the fluorescence is monitored; then, in the present example (0.5-cm cell, d = 0.25 cm; A = 0.09 + 0.05), F, = 1.084F.0.1 5 0.05 Wavelengthhm Fig. 9. Ultraviolet absorption spectra (with reference to a solvent blank), (1-cm path length) of solutions (20 pg ml-1) in cyclohexane of three different motor oils, a crude oil and a grease, in order of increasing absorbance at 275 nm. However, this factor is an underestimate. For non-polar molecules in weakly interacting solvents, and if excited-state intermolecular processes may be neglected, the corrected fluorescence intensity should vary linearly with the concentration (C) of the sample: F, = KC = 1OAdF, where K is a constant.In agreement with this, experimental plots of log(F/C) against A are linear over the range up to A = 1.0, but with slopes more negative than the nominal value of - d . An empirical value of d, derived from a variety of features and samples, is therefore used in the calculation of the correction factors. If the fluorescence intensities from solutions of a sample at two concentrations, indicated by subscripts 1 and 2, are compared: Hence, log (F/C) = logK - Ad. F,,,/F,,, = c,/c, = (F,/F,) ( 10d~d/lOd~d) When the ratio of the concentrations is 10, and if the Beer - Lambert law applies (this has been confirmed), A , - A , = 0.9A,, and then d = [l - log(F,/F,)] (0.9A1)-' In this way, ten features of four typical samples at concentrations of 100 and 10 pg ml-I gave 33 values of d (in seven instances the fluorescence intensity and the absorbances were too weak to be measured reliably), distributed with a mean and standard deviation of 0.353 and 0.060 cm, respectively.There is no significant variation of d with either feature or sample, and therefore the mean value is adopted for the calculation of the correction factors. For the above example, the factor is now 1.12. Other correction procedures are available that make allowance, for example, for the finite vanation in the depth %vithin the cell from which the measured fluorescence is ~ollected,~,108 LLOYD : EXAMINATION OF PETROLEUM PRODUCTS FOR FORENSIC Analyst, VoZ.105 and for the spectral distribution of the light emerging from a m0n0~hr0mator,4~ but these require that the optical path lengths in the cell be defined. Although the initial assumption in the present treatment may be poorly justified theoretically, the errors involved are written off into the empirical d , for which an absolute value is unnecessary. In earlier studies of petroleum luminescence, consideration of the strong inner filter effects sometimes evident has often been avoided by the use of standard concentrations. Conse- quently, whet her the variations between samples in apparent fluorescence intensities are due to variations in the ultraviolet absorbances of the samples or in the concentrations of their fluorescent components cannot be determined.Hence, one set of discriminating features, from the ultraviolet absorbance spectrum, is lost ; and another, the fluorescence intensities relative to a given amount of sample, is conditional on the equivalence of the inner filter effects in the samples concerned. Petroleums have been discriminated from one another partly on the basis of the changes caused to their fluorescence spectra by increases in c~ncentration,~~ but these changes are due to inner filter effects, which yield information more readily obtained by ultraviolet spectroscopy. Conclusion In Part I1 of this work23 the standard fluorescence comparison procedure (see Experi- mental) will be applied to a collection of samples, principally from case work, chosen to represent the range of high relative molecular mass petroleum products likely to be encountered as contact traces, with the object of assessing the efficiency of the various defined features as discriminators.Given these results, the development of computerised data collection and processing should facilitate the subsequent work on the establishment of evidential significance in specific circumstances. As in an earlier paper,14 it is strongly emphasised that for the characterisation of small amounts of petroleum derivatives there are other techniques available, such as infrared spectroscopy and those based on chromatography, that give results uncorrelated with the fluorescence results. Therefore, where the sample size permits, all of these techniques must be applied if evidential significance is, in general, to be optimised, even though on certain occasions fluorescence techniques have yielded evidence inaccessible by any other means.I very gratefully acknowledge the help, at the points already indicated, of Mr. D. A. Collins and Mr. C. A. Watson. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. References Parker, C. A., and Barnes, W. J., Analyst, 1960, 85, 3 . Eastwood, D., Fortier, S. H., and Hendrick, M. S., Int. Lab., 1978, July/August, 51. Green, G. L., and O’Haver, T. C., Anal. Chem., 1974, 46, 2191. Lloyd, J . B. F., Analyst, 1974, 99, 729. Milazzo, G., Caroli, S., Palumbo-Doretti, M., and Violante, N., Anal. Chem., 1977, 49, 711. McIlwrick, C. R., and Phillips, C. S. G., J . Phys. E. Scient. Instrum., 1973, 6, 1208.Lloyd, J . B. F., Analyst, 1977, 102, 782. Bradsher, C. K., and Smith, E. S., J. Am. Chem. Soc., 1943, 65, 451. Parker, C. A., “Photoluminescence of Solutions.” Elsevier, Amsterdam, 1968, p. 262. Moms, J. V., Mahaney, M. A., and Huber, J. R., J . Phys. Chem., 1976, 80, 969. Coakley, W. A., in “Proceedings of the Joint Conference on the Prevention and Control of Oil Hargrave, B. T., and Phillips, G. A., Envir. Pollut., 1975, 8, 193. Lloyd, J. B. F., J. Forens. Sci. Soc., 1971, 11, 235. Lloyd, J . B. F., J . Forens. Sci. Soc., 1971, 11, 153. Freegarde, M., Hatchard, C . G., and Parker, C. A., Lab. Pract., 1971, 20, 35. Chisholm, B. R., Eldering, H. G., Giering, L. P., and Hornig, A. W., Report No. BERC/RI-76/16, Lloyd, J. B. F., Nature Phys. Sci., 1971, 231, 64. Talmi, Y., Baker, D. C., Jadamec, J. R., and Saner, W. A., Anal. Chem., 1978, 50, 9368. Weiner, E. R., Anal. Chem., 1978, 50, 1583. Vo-Dinh, T., Anal. Chem., 1978, 50, 396. Warner, I. M., Callis, J. B., Davidson, E. R., and Christian, G. D., Clin. Chem., 1976, 22, 1483. Rho, J. H., and Stuart, J. L., Anal. Chem., 1978, 50, 620. Lloyd, J. B. F., Evett, I. W., and Dubery, J . M., to be published. Lloyd, J. B. F., and Evett, I. W., Anal. Chem., 1977, 49, 1710. Bentz, A. P., Editor, Report No. CG-D-52-77, US Coast Guard Research and Development Centre, Spills,” American Petroleum Institute, Washington, D.C., 1973, p. 215. Energy Research and Development Administration, Bartlesville, USA, 1976. Groton, Conn., 1977, Appendix E.Febrzcary, 1980 PURPOSES BY SYNCHRONOUS FLUORESCENCE SPECTROSCOPY 109 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. Gordon, D. C., Keizer P. D. Hardstaff, W. R., and Aldous, D. G., Envir. Sci. Technol., 1976, 10, John, P., and Soutar, I., Anal. Chem., 1976, 48, 520. Wakeham, S. G., Envir. Sci. Technol., 1977, 11,272. Parker, C. A., “Photoluminescence of Solutions, Elsevier, Amsterdam, 1968, p. 464. Fortier, S. H., and Eastwood, D., Anal. Chem., 1978, 50, 334. Lloyd, J . B. F., Analyst, 1975, 100, 82. Vo-Dinh, T., and Gammage, R. B., Anal. Chem., 1978, 50, 2054. Talsky, G., Mayring, L., and Kreuzer, H., Angew. Chem., Int. Edn Engl., 1978, 17, 785. Frank, U., in “Proceedings of the Joint Conference on the Prevention and Control of Oil Pollution,” American Petroleum Institute, Washington, D.C., 1975, p. 87. Tanacredi, J . T., J. Wat. Pollut. Control Fed., 1977, 216. Murray, E. C., and Keller, R. N., J. Org. Chem., 1969, 34, 2234. Keizer, P. D., and Gordon, D. C., J. Fish. Res. Bd Can., 1973, 30, 1039. Yip, H. L., J. Forens. Sci., 1973, 18, 263. Parker, C. A., and Rees, W. T., Analyst, 1962, 87, 83. Parker, C. A., and Barnes, W. J., Analyst, 1957, 82, 606. Leese, R. A., and Wehry, E. L., Anal. Chem., 1978, 50, 1193. Thruston, A. D., and Knight, R. W., Envir. Sci. Technol., 1971, 5, 64. 580. Received April 26th, 1979 Accepted September 1 lth, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500097

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

110 Analyst, February, 1980, Vol. 105, PP. 110-118 Determination of Tetramethyl- and Tetraethyllead Vapours in Air Following Collection on a Glass-fibre = lodised Carbon Filter Disc S. E. Birnie and F. G. Noden Associated Octel Company Limited, Research and Development Department, Ellesmere Port, South Wirral, L66 4HF Methods are described for the determination of tetraalkyllead vapour in air. The related sampling procedure operates efficiently over a wide range of lead- in-air concentrations, flow-rates and sampling periods. It is also effective under extreme conditions of humidity, can be applied in the presence of gasoline vapour and is especially convenient for personal monitoring or for use in the field. The air sample is passed through a glass-fibre - iodised carbon filter on which tetraalkyllead compounds are collected.Lead is then extracted from the filter and converted into the inorganic state by treatment with iodine solution, followed by colorimetric determination as lead dithizonate using a commercially available test kit. In an alternative procedure involving the use of nitric acid - bromine reagent for extraction, the lead is determined by atomic-absorption spectrophotometry with electrothermal atomisation. The efficiency of the methods has been confirmed using laboratory prepared air samples containing 5-400 pg m-3 of lead. Keywords : Tetramethyllead determination ; tetraethyllead determination ; air analysis Tetraethyllead (TEL) and tetramethyllead (TML) are manufactured on a large scale for use as antiknock additives to gasoline.Owing to the toxicity of organolead compounds it is necessary to monitor exposure to tetraalkyllead during manufacturing operations, distri- bution and use. The current threshold limit value - time-weighted average (TLV - TWA) is 100 pg m-3 for TEL and 150 pg m-3 for TML. Several of the published methods for the determination of tetraalkyllead compounds in air have been based on dry-sampling procedures. Snyder and Henderson1 described a simpli- fied field method in which iodine crystals were used as a collector. In this method a standard volume of sample was taken at a fixed sampling rate. A later method by Snyder2 involved collection of tetraalkyllead on active carbon over a period of up to 4 d but it was necessary to carry out a lengthy extraction procedure with nitric acid - perchloric acid prior to the lead determination.Coker3 also used a dry-sampling system, with an 8-h sampling period, and completed the determination by atomic-absorption spectrophotometry, as did Harrison et aZ.4 as the final stage of their method in which the sample was collected over a period of 30 min on a gas - liquid chromatographic column cooled by dry-ice. Laveskog5 established a method involving the use of gas chromatography linked with mass spectrometry, the sampling time being 15 min. This paper describes a sampling procedure that has a wide range of applications and is suitable for use in monitoring concentrations of tetraalkyllead in air both below and above the TLV. Sampling can take place over periods from 10min to 24h with appropriate sampling rates from 1 to 40lmin-l.The tetraalkyllead compounds are collected on a glass-fibre-iodised carbon filter. Only a simple treatment of the filter is necessary to remove the collected lead and the determination can be completed by a variety of analytical techniques. When laboratory facilities are not available the tetraalkyllead is extracted from the glass-fibre - iodised carbon filter with iodine solution. The concentration of lead in the resulting solution is determined spectrophotometrically using dithizone (diphenylthio- carbazone) . When laboratory facilities are available an alternative procedure can be used. The tetraalkyllead is extracted from the glass-fibre - iodised carbon filter with nitric acid - bromine solution and the con- The reagents are commercially available as a lead-in-air kit.BIRNIE AND NODEN 111 centration of lead in the resulting solution is determined by atomic-absorption spectrophoto- metry using electrothermal atomisation.The determination can also be completed by anodic-stripping voltammetry or by a spectrophotometric method. Experimental Instruments Surrey. atomiser accessory. Pitman lead-in-air analyser, Model 705. Supplied by D. A. Pitman Ltd., Weybridge, Perkin-Elmer 300 S atomic-absorption spectrophotometer with a n HGA-76 carbon furnace Reagents For use with the Pitman lead-in-air analyser lead-in-air analyser. The following ampouled reagents (D. A. Pitman Ltd.) are supplied for use with the Iodine, 0.2 N in methanol. Iodine, 2 N aqueous solution in potassium iodide.Solution A. A lead-free aqueous solution of potassium cyanide, sodium sulphite, Dithizone in chloroform, 40 mg 1-l. Details of the preparation of lead-free reagent solutions were given by Snyder and ammonium citrate and ammonium hydroxide contained in two ampoules. Caution-Chloroform is a suspected carcinogen and inhalation of the vapour should be avoided. Henders0n.l For use in alternative methods of analysis Potassium iodide solution, 25% m/V. Dissolve 500 g of potassium iodide in about 1 1 of distilled water. Make the solution slightly alkaline by the dropwise addition of ammonia solution (sp. gr. 0.880) and then de-lead it by shaking it with successive 50-ml portions of 20 mg 1-1 of dithizone in chloroform solution until the green colour of the dithizone solution remains unchanged. After separation of the organic phase make the solution slightly acidic by the dropwise addition of dilute nitric acid and then wash it with chloroform in order to remove any dissolved dithizone.Separate the chloroform layer and make up the volume of the remaining solution to 2 1 with distilled water. Mix 445 ml of de-leaded 25% m/V potassium iodide solution with 445 ml of analytical-reagent grade concentrated hydrochloric acid. Slowly, with cooling, add 7 5 g of analytical-reagent grade potassium iodate, stirring the solution until all of the free iodine has re-dissolved to give a clear orange- red solution; dilute the solution to 1 1 with distilled water. Dilute 1 volume of iodine monochloride stock solution with 9 volumes of distilled water.Iodine monochloride stock solution, 1.0 M. Iodine monochloride solution, 0.1 M. Caution- 1. 2. Rubber bungs must not be used to stopper vessels containing iodine monochloride solution. Iodine monochloride will react with ammonium ions under certain conditions to yield nitrogen triiodide, which has violently explosive properties. It is important, therefore, that ammonia and ammonium salts should be excluded from solutions containing iodine monochloride, except when an excess of reducing agent is also present. Nitric acid solution, 50% V/V. Dilute 1 volume of Aristar-grade concentrated nitric acid with 1 volume of distilled water. Nitric acid solution, 1% V/V. Dilute 1 volume of Aristar-grade concentrated nitric acid with 99 volumes of distilled water.Bromine, Aristar grade. Standard inorganic lead solution, 100 pg ml-l. Dissolve 0.160 g of analytical-reagent grade lead nitrate in distilled water, add 10 ml of Aristar-grade concentrated nitric acid and make up the volume to 1 1 with distilled water. Standard inorganic lead solution, 0.2 pg ml-l. Take 2 ml of 100 pg ml-l standard inorganic lead solution, add 10 ml of Aristar-grade concentrated nitric acid and dilute to 11 with distilled water.112 BIRNIE AND NODEN: DETERMINATION OF TML AND TEL VAPOURS IN AIR Analyst, Vol. 105 Pre-treatment of the glass-jibre - carbon jilters Soak Whatman glass-fibre - carbon filters, grade ACG/B, 25 mm diameter, in 0.2 N methanolic iodine in a sealed container for 15min. Remove the filters carefully using ivory-tipped forceps and allow them to dry in a desiccator for a minimum period of 2 h.This procedure may be carried out in the laboratory prior to use. Blank determination and limit of detection It is necessary to carry out several blank determinations as the lead content may vary, but normally it is less than 1 pg of lead per filter. The determination is carried out by the methods described later in the paper. At the lower end of the disc a scale reading of 0.5 can be estimated with reasonable confidence. Blank levels using this method rarely exceed a disc reading of 0.5 (equivalent to 1 pg of lead). It was therefore considered acceptable to base the limit of detection on a disc reading of 0.5, which when expressed as a concentration of tetraalkyllead in air the detection limit is 8 pg m--3 for a 1-h sampling period at a sampling rate of 2 1 min-l.When using the atomic-absorption method the limit of detection was calculated in the following way. Ten blank glass-fibre - iodised carbon filters were extracted and the solutions were diluted to 100 ml. Ten more blank glass-fibre - iodised carbon filters were spiked with 1 pg of lead, extracted and the solutions diluted to 100 ml so that the lead content was slightly above the blank value. Volumes of 2 0 4 of each of the sets of solutions were injected alternately into the furnace. The total lead content of each solution was calculated by reference to a calibration graph. The mean value obtained was 0.89 pg of lead. The lead found was calculated by subtracting the mean of two adjacent blanks from the blank + standard value.The mean value obtained for the lead found was 1.05 pg and the standard deviation was 0.2357 pg. The detection limit was considered to be twice the standard deviation. When expressed as a concentration of tetraalkyllead in air the detection limit is 4 pg m-3 for a 1-h sampling period at a sampling rate of 2 1 min-1. The method based on the Pitman analyser involved the use of a comparator disc. Collection of the sample Dismantle a Gelman 25 mm diameter, in-line filter holder and, using ivory-tipped forceps, place one of the treated carbon filters centrally on the mesh support. Re-assemble the holder and, using the minimum length of plastic tubing, connect the holder to the pump. Connect the other end of the holder to a tube flow meter, switch on the pump in a lead- free atmosphere and adjust the flow to the desired rate.Turn off the pump and disconnect the tube flow meter. Attach the pump and holder in the required position and sample for the required period. If the sampling period is greater than 30min it is essential to check the sample volume using a dry-air pollution meter. Analysis of the sample The concentration of tetraalkyllead compounds collected on the treated glass-fibre - carbon filter can be determined by one of the following two methods. Using the Pitman, Model 705, lead-in-air analysey and reagents. After completion of the sampling, dismantle the filter holder and, using ivory-tipped forceps, transfer the carbon filter into a midget impinger tube. Add 15 ml of 0.2 N iodine in methanol, stopper the tube and shake the contents for 1 min.Add 10 ml of 2.0 N iodine in aqueous potassium iodide solution and swirl the tube to mix the solution. Warm the solution to at least 27 "C and hold it at this temperature for a minimum of 5 min. Filter as much as possible of the iodine solution through a fine glass-wool plug into the comparator flask. Wash the flask and glass-wool with two 10-ml volumes of distilled water followed by three 5-ml volumes of distilled water. Add 35ml of solution A to the comparator flask. Care should be taken not to allow any free iodine to remain around the neck of the flask. Stopper the flask and shake it until the iodine is completely decolorised. Pour 5ml of chloroform into the vial containing 0.2 mg of dry dithizone and add this solution to the comparator flask.Vigorously shake the comparator flask for 30s and allow the two layers of liquid to separate. If the lower layer is colourless or has a slight greenish tint, there is no lead present. If a red colour is obtained in the lower layer and the upper layer is orange - yellow, place the com-February, 1980 AFTER COLLECTION ON GLASS-FIBRE - IODISED CARBON FILTER DISCS 113 parator flask in the Lovibond comparator and rotate the colour disc until a colour match is obtained. NOTES- 1. If an orange or red - orange colour is obtained in the chloroform layer an incomplete reaction of organic lead with the iodine is indicated. The analysis must be repeated, making sure that the solution is heated to and held at a minimum temperature of 27 "C for at least 5 min to permit complete reaction of the lead compounds with the iodine.If the upper layer does not appear yellow and the colour of the chloroform layer is darker than the deepest shade on the disc, insufficient dithizone is present and another portion of dithizone should be added. If the colour is still too deep for matching then it will be necessary to repeat the test but running for a shorter length of time. Carry out a blank determination using a treated glass-fibre - carbon filter. Using atomic-absorption spectro$hotometry. After completion of sampling, dismantle the filter holder and, using forceps, transfer the carbon filter into a 150-ml beaker. Add 10 ml of 50% V/V nitric acid and 1 ml of bromine. Cover the beaker with a watch-glass and warm the solution gently until all of the bromine has evaporated.Digest the solution on a hot-plate for 1 h, then evaporate the solution almost to dryness. Cool, add 10 ml of 1% V/V nitric acid to the residue and warm. Filter the solution through a Whatman No. 54 paper to remove any insoluble matter and dilute to 100 ml with 1% V/V nitric acid or to a suitable volume depending on the lead-in-air concentration, sampling rate or sampling time. Inject 20 pl of the solution into the carbon furnace using the following conditions: drying, 10 s at 150 "C; ashing, 10 s at 490 "C; and atomisation, 4 s at 2100 "C. Prepare a calibration graph using 0.2 pg ml-l standard inorganic lead solution under the conditions specified above. The linear range of the calibration is 0-2ng of lead injected but it is acceptable to use a range of 0 4 ng of lead injected. Carry out a blank determination using a treated glass-fibre - carbon filter.2. Shake the comparator flask for a further 30 s and obtain a colour match. Comparison of results obtained by various analytical techniques The method using the Pitman analyser and that applying atomic-absorption spectro- photometry with electrothermal atomisation are described in detail in the paper. However, at various times during the course of our investigations it was found convenient to complete the determination by anodic-stripping voltammetry or by a dithizone spectrophotometric method. A bulk stock of the solution used for extracting the lead from the carbon filter was prepared and spiked with amounts of lead consistent with the range of lead-in-air con- centrations studied.This stock solution was analysed by the four methods used in order to check their comparability. The results are shown in Table I. It was concluded that valid results are obtained from the various techniques described in this paper. TABLE I COMPARISON OF RESULTS OBTAINED BY VARIOUS ANALYTICAL TECHNIQUES Lead found/pg Lead added/pg 6 6 12 12 48 48 &man Electrothermal analyser atomic absorption 6 6.5 6 6.8 10 11.4 12 11.7 48 49.6 44 49.6 Dithizone spectrophotometry 6.0 6.3 11.6 11.8 49.3 48.6 Anodic-stripping voltammetry 6.9 6.0 11.1 11.8 48.1 48.3 Removal of particulate matter When particulate lead compounds are present in an air sample, part or all of this lead will be included with the tetraalkyllead determined.The extent to which the determina- tion of the tetraalkyllead is affected will depend on the nature of the lead compounds present114 BIRNIE AND NODEN: DETERMINATION OF TML AND TEL VAPOURS IN AIR Analyst, Vol. 105 and in particular on their solubility in the reagent solutions. When it is necessary to take account of particulate lead compounds the sampling procedure should be modified as follows. Dismantle a Gelman 25mm diameter open filter holder and, using forceps, place one 25 mm diameter, 0.8 pm, Millipore Type AA membrane filter centrally on the mesh support. Re-assemble the holder and, using PTFE tubing as a sleeve, make a direct connection to the inlet of the holder containing the treated carbon paper.Loss of tetraalkyllead dzle to adsorption on particulate matter It was considered necessary to check whether tetraalkyllead is adsorbed on particulate matter in situations where a pre-filter is used. Under sampling conditions relevant to our studies the volume of air taken is less than 1.5 m3 and the amount of particulate matter collected from this size of sample is rarely sufficient to give a visible stain on the pre-filter, even in an industrial environment. Therefore, in order to obtain a suitable amount of particulate matter for the investigation a much larger volume of air (32 m3) was taken. Several weighed samples of particulate matter were collected on a Millipore membrane filter and these were used to check for adsorption of tetraalkyllead as follows.Air samples containing a range of concentrations of tetraalkyllead were prepared by a diffusion method.6 The sample was passed through a previously prepared particulate filter followed by a glass-fibre - iodised carbon filter at a rate of 2 1 min-l for a period of 1 h. A duplicate air sample was passed through a glass-fibre - iodised carbon filter only. The tetraalkyllead collected on both glass-fibre - carbon filters was determined using atomic- absorption spectrophotometry as previously described. The results indicated that the tetraalkylIead concentration was unchanged after passing the air sample through the pre-filter and it was concluded that there is no significant adsorption of tetraalkyllead on particulate matter. The results are shown in Table 11.TABLE I1 LOSS OF TETRAALKYLLEAD DUE TO ADSORPTION ON PARTICULATE MATTER Lead-in-air concentration determined from glass-fibre - iodised carbon filter connected in series with a Lead-in-air concentration determined from glass-fibre - iodised carbon collected on pre-filterlmg filter alonelpg m-s pre-filterlpg m-* Mass of solid matter 1.1 0.9 1.3 0.8 0.7 0.8 9 33 72 81 178 236 8 31 87 80 188 22 7 Collection eficiency of zcntreated glass-fibre - carbon filter Air containing a range of concentrations of tetraalkyllead was prepared by a diffusion method.6 For the TML in air stream, a range of dilute solutions of pure TML in toluene were used in the diffusion cell. With TEL, pure TEL was used and the temperature of the diffusion cell was varied in order to cover the required range of lead-in-air concentrations.The sample stream was passed through a glass-fibre - carbon filter at a rate of 2 1 min-f for a given period of time. A duplicate air sample was passed through two scrubbers containing 15 ml of 0.1 M iodine monochloride at the same time and at the same flow-rate as for the test sample. It had previously been shown' that the collection efficiency of this system was satisfactory. The tetraalkyllead collected in the iodine monochloride was determined by an atomic-absorption technique in conjunction with a classical separation using dithizone.8 The tetraalkyllead collected on a glass-fibre - carbon filter was compared with that collected by iodine monochloride solution. A series of trials were carried out using dry air and water-saturated air streams.The tetraalkyllead collected on the glass-fibre - carbon filter was extracted and determined using the Pitman lead-in-air analyser as previously described. The tetraalkyllead collected by iodine monochloride solution was determined as lead dithizonate using a spectrophotometric single-extraction mono-colour method after reduction of the excess of iodine monochloride with sodium sulphite solution. The results obtained are shown in Table 111.Febr~ary, 1980 AFTER COLLECTION ON GLASS-FIBRE - IODISED CARBON FILTER DISCS 115 It was concluded that the collection efficiency for dry air streams was satisfactory but the presence of water vapour seriously impairs the lead collection efficiency of the glass-fibre - carbon filter. TABLE I11 COLLECTION EFFICIENCY OF GLASS-FIBRE - CARBON FILTER AT A SAMPLING RATE OF 2 1 min-l FOR A PERIOD OF 1 h Alkyllead Condition of present air stream TML Dry Dry Dry Dry Water saturated Water saturated Water saturated Water saturated TEL Dry Dry Dry Dry Water saturated Water saturated Water saturated Water saturated Lead-in-air concentration determined from glass- fibre - carbon filter/ 200 150 67 67 33 25 17 17 Pg 233 150 67 50 133 133 42 33 Lead-in-air concentration determined from iodine monochloride/ pg m-a 220 170 70 74 247 168 116 116 251 163 69 64 205 190 55 44 Lead collected by glass-fibre - carbon filter, % 91 88 96 91 13 15 15 16 93 92 97 93 65 70 76 76 Application to water-saturated air streams containing tetraalkyllead Attempts to remove water vapour selectively from sample streams by means of desiccating agents were unsuccessful, and means were therefore sought of minimising the effect of water vapour by modifying the collection system.The following treatment gave good results. The glass-fibre - carbon filters were treated with 0.2 N methanolic iodine solution and then allowed to dry in a lead-free atmosphere. The tetraalkyllead collected on a pre-treated glass-fibre - carbon filter was compared with that collected by iodine monochloride solution. A series of trials were carried out using dry air and water-saturated air streams. The range of the method was extended to sampling periods of 7 h. The tetraalkyllead collected on the glass-fibre - carbon filter was extracted and determined using the Pitman lead-in-air analyser as previously described.The tetra- alkyllead collected by iodine monochloride solution was determined as lead dithizonate using a spectrophotometric single-extraction mono-colour method after reduction of the excess of iodine monochloride with sodium sulphite solution. When the sampling period was 7 h, the tetraalkyllead collected by both the glass-fibre - carbon filter and the iodine monochloride solution was determined by anodic-stripping voltammetry after conversion into the inorganic state. It was concluded that the collection efficiency for both dry air and water-saturated air streams was satisfactory at tetraalkyllead in air concentrations up to 250 pg mW3 of lead. The results obtained are shown in Table IV. Collection of tetraalkyllead from air over 24-h sampling periods using $re-treated glass-- bre - carbon @ter It was decided to check that the sampling procedure was efficient for monitoring tetra- alkyllead (TML and TEL) over a period of 24 h with a sampling rate of 1 1 min-1.The tetraalkyllead collected on a pre-treated glass-fibre - carbon filter was compared with that collected by iodine monochloride solution connected in series behind the filter. The tetra- alkyllead collected on the glass-fibre - carbon filter was extracted and determined as lead dithizonate using a spectrophotometric single-extraction mono-colour method. The tetra- alkyllead collected by iodine monochloride solution was also determined as lead dithizonate using a spectrophotometric single-extraction mono-colour method after reduction of the excess of iodine monochloride with sodium sulphite solution.A comparison of the results obtained is shown in Table V.116 BIRNIE AND NODEN: DETERMINATION OF TML AND TEL VAPOURS IN AIR Analyst, VoZ. 105 TABLE IV COLLECTION EFFICIENCY OF GLASS-FIBRE - IODISED CARBON FILTER AT A SAMPLING RATE OF 2 1 min-l Alkyllead Condition of present air stream TML Dry Dry Dry Dry Dry Water saturated Water saturated Water saturated Water saturated Water saturated TEL Dry Dry Dry Dry Dry Water saturated Water saturated Water saturated Water Saturated Water saturated Period of samplinglh 1 1 1 1 7 1 1 1 1 7 1 1 1 1 7 1 1 1 1 7 Lead-in-air concentration determination from glass-fibre - carbon filterlpg m-8 200 200 58 67 173 160 160 26 60 218 133 200 67 67 218 200 200 33 60 202 Lead-in-air concentration determination from iodine monochloride/ 203 218 63 63 183 168 168 27 67 239 137 193 71 71 240 206 200 34 66 199 Lead collected by glass-fibre - carbon filter, % 99 92 92 106 96 96 89 93 88 91 97 104 94 94 91 97 97 97 89 102 It was concluded that the collection efficiency of the pre-treated glass-fibre - carbon filter was satisfactory for monitoring tetraalkyllead in air at concentrations up to 200 pg m-3 of lead for periods up to 24 h.TABLE V COLLECTION EFFICIENCY OF GLASS FIBRE - IODISED CARBON FILTER AT A SAMPLING RATE OF 1 1 min-l FOR A PERIOD OF 24 h Lead-in-air concentration determined from glass-fibre - iodised carbon filterlpg m-* 6 9 29 37 68 61 70 78 92 106 126 131 187 Lead collected in iodine monochloride connected in series behind the carbon Glterlpg m-s 0.1 0.1 1 2 2 4 6 2 6 2 4 2 6 Lead collected by glass-fibre - carbon filter, % 98 99 97 96 97 94 03 98 96 98 97 98 97 Collection of tetraalkyllead from air streams at high flow-rates At sampling rates of 40lmin-l it is not possible to pass a duplicate air stream through iodine monochloride solution as the collection efficiency of tetraalkyllead is lo+ at sampling rates above 4lmin-1.To overcome this problem, a single air stream was prepared and passed through two pre-treated glass-fibre - carbon Nters placed in series. The lead collected on the second glass-fibre - carbon filter was a measure of the collection efficiency of the first glass-fibre - carbon filter. The tetraalkyllead collected on both the glass-fibre - carbon filters was determined using atomic-absorption spectrophotometry as previously described.A comparison of the results obtained is shown in Table VI.February, 1980 AFTER COLLECTION ON GLASS-FIBRE - IODISED CARBON FILTER DISCS 117 It was concluded that the collection efficiency of pre-treated glass-fibre - carbon filters was satisfactory at sampling rates up to 40 1 min-l for periods up to 2 h when the tetra- alkyllead in air concentration did not exceed 400 pg m-3 of lead. TABLE VI COLLECTION EFFICIENCY OF GLASS-FIBRE - IODISED CARBON FILTER AT A SAMPLING RATE OF 40 1 min-l FOR PERIODS UP TO 2 h Lead-in-air concentration determined from 1st glass- Alkyl lead Period of fibre - iodised carbon filter/ present sampling/h P?3 TML 0.6 29 0.6 14 2 23 2 19 2 116 2 2 172 333 Lead-in-air concentration determined from 2nd glass- fibre - iodised carbon filter/ 0.4 0.2 0.1 0.1 0.7 5 24 Pg Lead collected by glass-fibre - carbon filter, % 99 99 100 99 99 97 93 2 404 40 91 TEL 0.6 0.6 2 2 2 2 2 2 128 196 18 23 127 176 414 626 0.2 0.1 0.1 0.1 0.4 1 40 81 100 100 99 100 100 99 91 87 Application to air streams containing gasoline vapour An obvious application of the method would be the determination of tetraalkyllead in the atmosphere at oil-refinery sites.It was therefore necessary to check that the collection efficiency of the glass-fibre - iodised carbon filter is not impaired by the passage of gasoline vapour. Air containing a range of concentrations of tetraalkyllead produced from four-star grade gasoline was prepared by a diffusion method as previously described.The air was passed through a glass-fibre - iodised carbon filter at a rate of 1-2 1 min-l for periods of up to 24 h. The tetraalkyllead collected on the glass-fibre - iodised carbon filter was compared with that collected by iodine monochloride connected in series behind the filter. The tetraalkyllead collected on the glass-fibre - iodised carbon filter was extracted and determined by anodic- stripping voltammetry. The tetraalkyllead collected in iodine monochloride solution was also determined by anodic-stripping voltammetry after conversion into the inorganic state. A comparison of the results obtained is shown in Table VII. It was found that the collection efficiency of pre-treated glass-fibre - carbon filter was greater than 98% in the presence of gasoline vapour for tetraalkyllead in air concentrations up to 200 pg m--3 of lead.TABLE VII COLLECTION EFFICIENCY OF GLASS-FIBRE - IODISED CARBON FILTER FROM AIR CONTAINING GASOLINE VAPOUR Sampling rate/l min-l 2 2 1 1 40 40 Period of sampling/h 1 1 24 24 2 2 Lead-in-air concentration determined from glass-fibre - iodised carbon filter/pg m-3 207 200 106 204 127 136 Lead collected in iodine monochloride connected in series behind the carbon filterlpg m-3 2 <1 (1 (1 2 2 Lead collected by glass-fibre - carbon filter, yo 99 > 99 > 99 > 99 98 99I18 BIRNIE AND NODEN Conclusion The method developed can be used for the determination of tetraalkyllead in air over a wide range of concentrations, flow-rates and sampling periods. Collection efficiencies greater than 90% can be achieved for samples containing 5400 pg m-3 of lead by suitable choice of sampling period and flow-rate. The sampling equipment is especially convenient for personal monitoring or application in the field for periods of up to 24 h. Acknowledgements are made to colleagues for their advice during the course of this work, and to the Associated Octel Company for permission to publish this paper. 1. 2. 3. 4. 6. 6. 7. 8. References Snyder, L. J., and Henderson, S. R., Anal. Chem., 1961, 33, 1175. Snyder, L. J., Anal. Chem., 1967, 39, 591. Coker, D. T., Ann. Occup. Hyg., 1978, 21, 33. Harrison. R. M., Perry, R., and Slater, D. G., Atrnosph. Envir., 1974, 8, 1187. Laveskog, A., “Second International Clean Air Congress,” International Union of Air Pollution Associates, Washington, D.C., 1970, pp. 549-557. Altshuller, A. P., and Cohen, I. R., Anal. Chem., 1960, 32, 802. Moss, R., and Browett. E. V., Analyst, 1966, 91, 428. Hancock, S., and Slater, A., Analyst, 1975, 100, 422. Received May 301h, 1979 Accepted August 22nd, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500110

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, February, 1980, Vol. 105, $@. 119-124 119 Determination of Copper, Lead, Cadmium, Nickel and Cobalt in EDTA Extracts of Soil by Solvent Extraction and Graphite Furnace Atomic-absorption Spectrophotometry Birgitte Pedersen, Marta Willems and S. Storgaard Jsrgensen Chemistry Department, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Copen- hagen V , Denmark A procedure is described for the determination of copper, lead, cadmium, nickel and cobalt in EDTA extracts of soil and similar material. Diethyl- ammonium diethyldithiocarbamate or ammonium tetramethylene dithio- carbamate metal complexes are extracted into xylene from the EDTA extracts, and metals are determined in the xylene phase by atomic-absorption spectrophotometry using a graphite furnace atomiser.The detection limits (concentrations in soil) are approximately copper 0.8, lead 0.3, cadmium 0.07, nickel 2.5 and cobalt 0.8 p g g - l . These detection limits might be improved by a t least a factor of 10. Iron, manganese, aluminium, calcium and zinc do not interfere in amounts likely to be found in extracts of natural or con- taminated soils. Keywords : Copper, lead, cadmium, cobalt and nickel determination ; solvent extraction ; atomic-absorption spectrophotometry ; graphite furnace atomisa- tion; soil analysis Of several aqueous solutions prepared for extracting the “plant available” portion of a number of metallic trace elements, a buffered EDTA solution is one of the most commonly used. Various concentrations of EDTA, pH ranges, soil to solution ratios and extraction times have been employed in order to optimise the reproducibility and effectiveness of the extraction.Trace metal concentrations in EDTA extracts of soil etc. are often in the nanograms per millilitre range, which makes atomic-absorption spectrophotometry with electrothermal atomisation a convenient method of determination. By introducing a solvent extraction step before the atomic-absorption measurement metal concentrations can be increased and matrix interferences decreased. Several trace elements form dithiocarbamate complexes, which can be quantitatively extracted into an organic solvent over a wide pH range.2 In earlier work carbon tetrachloride or chloroform was often used as the solvent and the metal in question was determined by molecular spectrophotometry.2 For the determination of several trace metals by flame atomic-absorption spectrophotometry extraction of dithio- carbamate complexes into 4-methylpentan-2-one has been empl~yed.~,~ Tjell and co- w o r k e r ~ ~ ~ ~ used extraction with diethylammonium diethyldithiocarbamate (DDDC) in xylene from nitric acid solution prior to atomic-absorption spectrophotometry with electro- thermal atomisation.Xylene is a convenient solvent as it is lighter than water, nearly insoluble in water, has good separation characteristics and is halogen free. It has been demonstrated that hydrocarbon halides, such as chloroform, may cause considerable loss of several elements during elect rot hermal at omisation .’ The purpose of this work was to demonstrate the usefulness of graphite furnace atomic- absorption spectrophotometry after solvent extraction with dithiocarbamates in xylene for the determination of copper, cadmium, lead, nickel and cobalt in EDTA extracts of soil and similar materials, and further to investigate the tolerance of the proposed method towards pH variations and interfering species.The use of EDTA solutions in soil analysis has been reviewed by B0rggaard.l * Presented at Euroanalysis 111, Dublin, August, 1978.120 Reagents All chemicals were of analytical-reagent grade, if not otherwise stated. Doubly de- ionised water direct from the de-ioniser proved satisfactory. All plastic apparatus and glass- ware are kept in nitric acid (1 + 1) when not being used. PEDERSEN et al. : DETERMINATION OF Cu, Pb, Cd, Ni Experimental Analyst, Vol.105 Xylene. Diethylammonium diethyldithiocarbamate (DDDC) , 1 .O% ml V solution in xylene. Ammonium tetramethylene dithiocarbamate (ammonium pyrrolidine dithiocarbamate, APDC) , 5.0% m/V solution in water. This solution is prepared daily and filtered before use. Ethylenediaminetetruacetic acid, disodium salt (EDTA ) , 0.20 M soladion. Acetate bufer, p H 4.6. Acetic acid (1 mol) plus sodium acetate (1 mol) are dissolved in water and diluted to 1 1 with water. Before use, 1 1 of this solution is purified by extraction five times with 25 ml of 1% DDDC in xylene solution and five times with 25 ml of xylene. Nitric acid, 1 + 3. Sodium acetate solution, 2 M. Standard solutions. Stock solutions of copper, lead, cadmium, nickel and cobalt, each containing 1 OOO pg ml-l of one of these elements, are prepared from CuSO,.SH,O, Pb(NO,),, 3CdS0,.8H20, NiSO,, (NH4),S04.6H,0 and CoS0,.7H20, respectively, and diluted to volume with 0.2 M nitric acid.Stock solutions of iron, manganese, aluminium, calcium and zinc, each containing lo00 pg ml-l of one of these elements, are prepared from metallic iron, MnSO,.H,O, metallic aluminium, CaCO, and ZnS0,.7H20, respectively, and diluted to volume with hydrochloric acid so that the final solutions are 0.1 M in hydrochloric acid. BDH ; laboratory reagent, sulphur free. Apparatus The atomic-absorption spectrophotometer used was a Perkin-Elmer, Model 303, instru- ment (without background corrector) equipped with Perkin-Elmer HGA-74 graphite cell, HGA-2100 controller and Model 56 recorder. Single-element hollow-cathode lamps were used as radiation sources.Slits and wavelengths were set according to the instrument manual. The argon flow-rate through the graphite cell was 50 ml min-l. When lead was determined the argon flow was interrupted during the atomisation stage. Samples were injected with a Finnpipette (5-5Opl) or by means of the Perkin-Elmer AS-1 Auto-Sampling system (20 pl) usihg polyethylene sample cups. The atomisation conditions utilised are shown in Table I. For cadmium a relatively high charring tempera- ture is necessary and for all of the elements determined a relatively long charring time is required in order to destroy the metal complexes. TABLE I ATOMISATION CONDITIONS Metal I 8 Condition cu Pb cd Ni co Drying temperaturel'c .. .. 175 175 175 175 176 Drying time/s . . .. .. .. 20 20 20 20 20130 Charring time/s . . .. .. 40 60 60 60 60 Wavelength/nm . . .. .. 324.7 283.3 228.8 232.0 240.7 Charring temperaturel'c . . . . 1000 500 500 1 000 1100 Atomisation temperaturel'c . . . . 2500 2 200 1800 2 700 2 700 Atomisation timels . . .. .. 6 6 6 8 10 Procedure EDTA extraction Before extraction the material is air dried at room temperature (20-24 "C), passed through a 2-mm sieve and further homogenised by hand grinding in an agate mortar. Then a 0.54-g amount is shaken for 24 h in an end-over-end shaker (3040 rotations per minute) with 15.0ml of acetate buffer, 15.0ml of EDTA solution and 60.0ml of water, and then filtered. The pH of this solution is about 4.6.Febrzlary, 1980 AND CO IN EDTA EXTRACTS BY GRAPHITE FURNACE AAS 121 Solvent extraction and atomic-absorption spectroplhotometric determination Extraction is performed in 25-ml borosilicate test-tubes with standard ground-glass stoppers.All volumes are dispensed by means of Finnpipettes with disposable polyethylene tips. A 5-7-ml aliquot of aqueous phase and 1.00 ml of xylene phase are used throughout and the stoppered test-tube is shaken vigorously for 2-3min by hand or by means of a reciprocating shaker. The metal content of the xylene phase is determined by graphite furnace atomic-absorption spectrophotometry on the same day (for cadmium, within 2 h). Samples should not be left in the polyethylene cups of the Auto-Sampler for more than 2 h and unknowns should be interspaced with standards in order to compensate for evaporation losses.It was found that 4% of the xylene would evaporate in 2 h. For the determination of 15-150 ng of copper, 1-10 ng of cadmium or 5-100 ng of lead, 5.0 ml of EDTA extract (if less than 5 ml has to be used, acetate buffer is added until the total volume of the aqueous phase is 5 ml) plus 1.00 ml of 1% DDDC in xylene are shaken for 2 min. For the determination of 50-800 ng of nickel, 5.0 ml of EDTA extract (or EDTA extract plus acetate buffer), 0.23 ml of nitric acid (1 + 3) decreasing the pH to about. 3.8, 1.00 ml of 5% APDC in water and 1.00 ml of xylene are shaken for 3 min. For the determination of 10-200ng of cobalt the same procedure as for nickel is used, except that nitric acid is omitted.The two phases separate very rapidly. Calibration procedure In 25-ml test-tubes are prepared 5-ml volumes containing suitable amounts of metals within the ranges given above and acetate buffer, EDTA and nitric acid corresponding to unknowns. To these volumes are added 1.00 ml of 1% DDDC in xylene (copper, cadmium and lead) or 1.00 ml of 5% APDC in water plus 1.00 ml of xylene (cobalt and nickel). After shaking, the determination is carried out under the same conditions as for unknowns. Standard-additions tests To a suitable amount of an EDTA extract of soil (copper and lead 1 ml; cadmium 2 ml; cobalt 5 ml) were added 50 ng of copper, 25 ng of lead, 4 ng of cadmium or 10 ng of cobalt before solvent extraction. After measurement, the percentage recovery was determined.Influence of pH on Metal Extraction Copper, cadmium and lead Amounts of nitric acid to give a range of pH values, acetate buffer, sodium acetate, EDTA and water to a volume of about 5ml were added to amounts of diluted stock solutions containing 100ng of copper, 50ng of lead or 4ng of cadmium. These solutions were prepared in duplicate. In one set of solutions pH was measured (pH 2-6) and to the other set of solutions was added 1.00 ml of 1% DDDC in xylene. Extraction and metal deter- mination were performed as described above. Nickel and cobalt Two sets of solutions containing 500 ng of nickel or 200 ng of cobalt were prepared as above and 1 ml of 5% APDC was added (pH 3-6). The pH was measured as above. After addition of 1 ml of xylene and extraction, the metals were determined as described.Interferences Various amounts of the stock solutions of iron, manganese, aluminium, calcium and zinc were added to known amounts of the metals to be determined (copper 100 ng, lead 50 ng cadmium 4 ng, nickel 200 ng and cobalt 100 ng). Acetate buffer and EDTA solution corresponding to approximately 5 ml of EDTA extract were added. Then the determina- tion was completed as described under Procedure.122 PEDERSEN et al. : DETERMINATION OF Cu, Pb, Cd, Ni Results and Discussion Analyst, VoZ. 105 It was found in preliminary experiments that copper, lead and cadmium could be extracted into the xylene phase by DDDC when the pH of the aqueous phase was about 4.6. The absorbances obtained were the same whether EDTA was present or not.With nickel and cobalt no absorbance was obtained when DDDC was used as extractant and the aqueous phase was 0.033 M EDTA solution and had a pH of 34.6. Using APDC as extractant cobalt gave the same absorbance with and without EDTA at pH 4.6. Using APDC nickel was extracted from an EDTA solution only when the pH was below 4. Using accepted values of metal - EDTA complexation constants and side-reaction co- efficients* and published metal - DDDC extraction constants?-12 it can be calculated that more than 99% of the copper, cadmium and lead present would be extracted into carbon tetrachloride or chloroform at pH 4.6 and the concentrations and volume ratios of EDTA and other reagents employed in the present study. Reported values of Ni-DDDC and Co - DDDC extraction constants vary ~onsiderably.~@~~3 Employing the lowest values of the extraction constants, neither cobaltg nor nickeP3 should be extracted at pK 4.6. Calculations based on published constants for the extraction of Ni - APDC13 and Co - APDP indicate that it should not be possible to extract the two metals into chloroform in the presence of EDTA at pH 4-4.6.The distribution of metal - DDDC or metal - APDC complexes between an aqueous phase and xylene does not seem to have been investigated. Extraction constants for the extraction of Cu -, Pb -, Cd -, Ni - and Co - DDDC complexes into benzene have been found to be very similar to those for extraction into carbon tetrachloride or chloroform.10 The results obtained in this study agree with the results predicted by the above calcula- tions for copper, cadmium and lead, but not for cobalt and nickel.It is not clear whether the reason for this is that there exist equilibria that have not been taken into account, or that there is a difference between xylene and the solvents that have been studied. Injluence of pH on extraction within the pH range 1.5-6 using DDDC. at pH < 5 using APDC. It can be seen from Fig. 1 that the extraction of copper, cadmium and lead is complete Nickel is extracted only at pH (4 and cobalt cu Pb Cd *O 0 i...... 1' 2 3 4 5 6 PH Fig. 1. Relative amounts of metals extracted from aqueous EDTA solution into xylene phase by DDDC (Cu. Pb, Cd) or APDC (Co, Ni) as a function of pH in the aqueous phase.February, 1980 123 Interferences It should be noted that only in a few instances are the maximum amounts that can be tolerated shown. How- ever, larger amounts than those tested are rarely encountered in EDTA extracts of soils, etc.When iron is present the xylene phase turns brown after extraction and the colour intensity increases with iron content. When about 1 mg of iron is present a precipitate is formed in the xylene phase, but even this amount did not interfere in the determination of cadmium. However, when determining nickel and cobalt the iron content may become close to the interference limit. With cobalt, up to 150pg of iron may be present without interfering if the pH is decreased to 3.5. AND Co IN EDTA EXTRACTS BY GRAPHITE FURNACE AAS The results of the interference tests are demonstrated in Table 11.TABLE I1 INTERFERENCE TESTS Amounts of common elements in the aqueous phase that did not interfere in the presence of the stated amounts of elements to be determined in 1 ml of xylene phase. Amount, of metal f A > Amount of metal tolerated/pg to be determined/pg Fe Mn Al Ca Zn cu, 0.100 100 100 100 100 100 Pb, 0.050 100 100 100 100 100 Cd, 0.004 1000 100 100 1000 100 Nil 0.200 100 100 1000 2 000 100 co. 0.100 50 200 1000 2 000 50 Reproducibility and determination limits When repeating the determination on the same xylene extract a deviation of 2% absorption or less between duplicates was considered satisfactory. It should be noted that the cali- bration graphs for copper, lead and cadmium intercepted the absorption axis at about 2-3% absorption.Calibration graphs for cobalt and nickel intercepted the absorption axis at a negative value. Representative results from duplicate determinations on sewage sludge - soil mixtures, fly ash and soil are shown in Table 111. A relative deviation of about 10% or less between duplicate EDTA extracts of solid sub-samples was considered satisfactory. Standard- additions tests demonstrated the absence of matrix interferences. By using the procedure described, the following or larger amounts dissolvable in an EDTA solution from soil or similar samples can be determined: copper 0.8, lead 0.3, cadmium 0.07, TABLE I11 REPRESENTATIVE RESULTS Deter- Metal mination* Cu .. A B Pb .. A B Cd .. A B Ni .. A B Co .. A B Sludge - soil mixture/ 156, 157 166, 169 227, 216, 227 257, 244, 251 10.17, 9.10, 9.25, 9.85 7.29, 6.90, 7.29, 6.90 15.74, 14.75, 14.63 15.67.15.75 Pg g-' Fly ash/ 22.7 22.2 22.3 22.0 0.96, 1.12 1.00, 1.18 18.9 19.7 Pg g-l soil/ Pg g-l 9.79, 9.12 8.21, 8.70 8.44, 8.78 8.73, 8.21 0.122, 0.098 0.093, 0.086 0.32, 0.19 0.27 Standard-additions tests r * Amount added to EDTA soil Recovery, 50 101 A extracting % 25 104 4 94 10 99 * For each metal, A represents replicate determinations on the same EDTA extracts and B represents replicate determinations on EDTA extracts of an identical sub-sample.124 PEDERSEN, WILLEMS AND STORGAARD JgRGENSEN nickel 2.5 and cobalt 0.8pgg-l. It should be noted that the ratio of sample to EDTA solution (1 : 167) is not typical, normally a higher ratio (1 : 10) being employed. Therefore, it should be simple to lower the determination limits, although the possibility of interference, especially from iron, should not be overlooked.The determination limits might also be lowered by increasing the aqueous to organic volume ratio and/or by injecting a larger volume of xylene phase into the graphite furnace atomiser. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Borggaard, 0. K., Acta Agric. Scand., 1976, 26, 144. De, A. K., Khopkar, S. M., and Chalmers, R. A., “Solvent Extraction of Metals,” Van Nostrand Hannaker, P., and Hughes, T. C., Anal. Chem., 1977, 49, 1486. Evans, W. H., Read, J. I., and Lucas, B. E.. Analyst, 1978, 103, 680. Tjell, J. C., and Palfeldt, K., “Report on NJF Seminar on Heavy Metal Circulation in Agriculture,” (in Danish), Ladelund, Denmark, October 1975, p. 188 (quoted in reference 6). Tjell, J. C., and Hovmand, M. F., Acta Agric. Scand., 1978, 28, 81. Volland, G., Kolblin, G., Tschopel, P., and Tolg, G., 2. Anal. Chem., 1977, 284, 1. Fritz, J. S., and Schenk, G. H., “Quantitative Analytical Chemistry,” Third Edition, Allyn and Star$, J., and Kratzer, K., Anal. Chim. Ada, 1968, 40, 93. Usatenko, Y. I., Barkalov, V. S., and Tulyupa, F. M., Zh. Anal. Khim., 1970, 25, 1458. Ooms, P. C. A., Brinkman, U. A. T., and Das, H. A., Radiochem. Radioand. Lett., 1977, 31, 317. Bajo, S., and Wyttenbach, A., Anal. Chem., 1979, 51, 376. Sharfe, R. R., Sastri, V. S., and Chakrabarti, C. L., Anal. Chem., 1973, 45, 413. Likussar, W., and Boltz, D. F., Anal. Chem., 1971, 43, 1273. Reinhold, London, 1970, Chapter 8. Bacon, Boston, 1974, p. 226. Received June 26th, 1979 Accepted August 30th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500119

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, February, 1980, Vot. 105, pp. 125-130 Simultaneous Acid Extraction of Six Trace Metals from Fish Tissue by Hot-block Digestion and Determination by Atomic-absorption Spectrometry 125 Haig Agemian, D. P. Sturtevant and K. D. Austen Canada Centre for Inland Waters, Water Quality Branch, 867 Lakeshore Road, P.0 Box 5050, Burlington, Ontario, Canada, L7R 4A 6 A simple and rapid digestion method is reported for the simultaneous acid extraction of chromium, copper, zinc, cadmium, nickel and lead from high-fat fish tissue. Samples are digested with nitric and sulphuric acids at 150 "C in a modified aluminium hot-block. The method is specially set up for fish sample sizes of up to 5 g, for low level detection of these elements. After digestion, acid extracts of the sample are analysed by direct flame atomic- absorption spectrometry for copper, zinc and chromium.The other three elements, cadmium, nickel and lead, are concentrated by chelation with ammonium tetramethylene dithiocarbamate followed by solvent extraction with isobutyl methyl ketone and determined by flame atomic-absorption spectrometry. The ease, rapidity and safety by which samples can be pro- cessed by this method make it suitable for the routine preparation of a large number of samples simultaneously. Keywords : Trace metal determination ; fish analysis ; hot-block digestion; atomic-absorption spectrometry In environmental pollution studies there is often the need for routine monitoring of toxic constituents in aquatic substrates. In large-scale studies, several trace parameters are often of interest in each sample obtained.Recently, our laboratory has been engaged in such a project, requiring methods for six trace elements (chromium, copper, zinc, cadmium, nickel and lead) in whole fish tissue. In the interest of speed, economy and simplicity, it would be a great advantage to be able to utilise a method that would simultaneously acid extract the six trace metals from the fish tissue, as well as making the digestion of a large number of samples possible. Evans et aZ.1 described a method for the simultaneous acid extraction of cadmium, nickel and lead from foodstuffs using Method (1)C of the Analytical Methods Committee.2 It calls for digestion with nitric and sulphuric acids in Kjeldahl flasks. As specified2 the method could be used with or without the addition of perchloric acid or hydrogen peroxide.As a simple and safe method is required for routine use, it is preferable to obtain a procedure that avoids the use of troublesome and possibly hazardous reagents, such as perchloric acid. The method indicated above has several limitations that make it unsatisfactory for the requirements of this study. The apparatus used does not lend itself to the simultaneous digestion of a large number of samples. In addition it was found that if high-fat samples were to be satisfactorily digested without the use of perchloric acid then the method had to be substantially modified. Aluminium hot-blocks3~* have been used for the digestion of fish tissue at temperatures of 180 "C and over for the extraction of even a volatile element such as mercury.The possibility of using long test-tubes in the block produces a refluxing action during digestion and together with the ease of temperature control in the block makes it a very efficient and reproducible sample processing technique. The aluminium hot-block used by these workers is well suited to the simultaneous digestion of a large number of samples. Under the specified conditions3~* only 0.1-0.5 g of fish could be used. In this study, in order to achieve the required detection limits, it was necessary to use sample sizes of about 5 g, much larger than those used in previous hot-block digestions. It was therefore necessary to modify the aluminium hot-block as well as the digestion mixture used, in order to digest large samples satisfactorily and also to oxidise all the fat tissue.126 AGEMIAN et aZ.: SIMULTANEOUS ACID EXTRACTION OF TRACE Analyst, VoZ. 105 In order that a parameter could be confidently determined in a sample, the analytical technique used must provide good sensitivity, and detection limits substantially lower than the level of the analyte in that substrate. The natural levels of chromium, copper and zinc are high enough that direct flame analysis of acid extracts, based on a 5-g sample size, provides satisfactory detection limits. For cadmium, nickel and lead, however, natural levels are so low that a further concentration step is required. In this study, it is shown that the digestion method adopted gives rise to solutions that are useful for chromium, copper and zinc deter- minations, and are also adaptable to the conventional ammonium tetramethylene dithio- carbamate (ammonium pyrrolidine dithiocarbamate) - isobutyl methyl ketone (APDC - MIBK) chelation - solvent extraction system of concentration.Under the scheme used in this study, all six trace elements could be determined with satisfactory detection limits from a single acid extract. The method outlined in this paper is simple, economical and safe and provides an effective means of multi-element determinations in the routine monitoring of a large number of fish samples. Experimental Apparatus Calibrated digestion test- tubes are used in the block to digest the samples. The block is placed on a hot-plate capable of heating the block to a constant temperature of 150 "C.The temperature is monitored with a thermometer suspended in mineral oil placed in a tube in the block. The aluminium block used in this study (Fig. 1) is a modification of that used by Bishop et aZ.4 As large samples and large volumes of acid are used, it is essential to have as much of the digestion tube in the hot-block as possible. The depth of the holes was therefore changed from 1.5 to 2.5in. This change allows the sample - acid mixture to be heated more uniformly . The aluminium hot-block is constructed as shown in Fig. 1. 0.5 in Fig. 1. Aluminium hot block. All the metals were analysed using a Perkin-Elmer, Model 603, atomic-absorption spectro- meter equipped with a triple-slot burner and deuterium background correction. An acid- resistant nebuliser was used for the direct analysis of chromium, copper and zinc from the acid extracts.A conventional nebuliser was used for the analyses of cadmium, nickel and lead after chelation and solvent extraction with the APDC - MIBK system. Deuterium background correction was used for the direct flame analyses.February, 1980 METALS FROM FISH TISSUE BY HOT-BLOCK DIGESTION AND AAS Reagents 127 High-purity certified reagents were used for all analyses. Nitric acid, 16 N. Su~phuric acid, 36 N. Ammonium tetramethylene dithiocarbamate (ammonium pyrrolidine dithiocarbamate, A PDC) Isobutyl methyl ketone (MIBK). Hydrogen peroxide, 30% VlV. solution, 1% m/V. Procedure Digestiort Weigh about 5 g of fish into a calibrated digestion tube. Add 5 ml of nitric acid (16 N) and then 5 ml of sulphuric acid (36 N) to the sample.Allow the reaction to proceed, taking care not to allow any overflow of the sample. When the reaction slows, place the digestion tubes in the hot-block, heated to 60 "C, for 30 min. Remove the tubes from the hot-block and allow to cool for 5 min and then add another 10 ml of nitric acid. Return the tubes to the hot-block and increase the temperature, in steps, to 120 "C (the contents of the tubes should be boiling) until the liquid is about level with the top of the block. Increase the temperature to 150°C. Remove the tubes when the samples go black, allow to cool for 5 min and add 1 ml of hydrogen peroxide (a vigorous reaction may occur). Return the tubes to the block. Repeat the hydrogen peroxide additions until the samples are clear.Remove the tubes and make up to 50 ml with de-ionised water when cool. Analysis Zinc, copper and chromium are analysed directly by flame atomic-absorption spectro- metry. Standards must be treated in the same way as samples and carried through the whole procedure. It is very important that the standards contain the same amount of acid as the samples, especially sulphuric acid as the final solution contains 10% of this acid and its effect on the viscosity of the solution results in a significant suppression of sensitivity. This is strictly a physical effect and results in suppressions of sensitivity of 15, 10 and 25% for copper, zinc and chromium, respectively. Nickel, lead and cadmium are concentrated by chelation - solvent extraction as follows.A volume of the sample digest (40 ml) is made up to 100 ml; 5 ml of the APDC solution and 5 ml of MIBK are added to the sample. Nickel, lead and cadmium are analysed in the MIBK phase by flame atomic-absorption spectro- metry. Standards containing 1-100 pg 1-1 of each metal and 4% V/V sulphuric acid are also run. The mixture is shaken vigorously for 5 min. Results and Discussion Heavy metals are usually acid extracted from biological tissue by digestion with several different mixtures of nitric, sulphuric and perchloric acids. The most popular combinations are nitric acid - perchloric acid, nitric acid - sulphuric acid and nitric acid - perchloric acid - sulphuric acid.5 Perchloric acid, as is well known, requires special precautions to be takens and is undesirable for use in a simple routine method.Therefore, the mixture of choice would be nitric acid - sulphuric acid. Previously reported hot-block digestion methods have made use of a 1 + 4 nitric acid - sulphuric acid mixture to extract mercury from 0.1-0.5 g of fish tissue. As indicated earlier, in view of the low detection limits required in this study, 5-g sample sizes are required. The increased organic matter in such sample sizes makes the above acid mixture inadequate, Because most of the oxidative action of this mixture is due to the nitric acid, it was found that a modification of the mixture to 3 + 1 nitric acid - sulphuric acid effectively dissolved the whole sample under the conditions specified under Experimental. It is important to point out that the ease and rapidity of sample dissolution are a function of fat content.For example, fish fillet samples, which contain mainly protein tissue, dissolve completely at room temperature when left in contact with the acid mixture overnight. However, whole fish samples, which contain varying amounts of fat tissue, are only partially solubilised. For the dissolution of fat, heating at 150 "C was required. This128 AGEMIAN et al. : SIMULTANEOUS ACID EXTRACTION OF TRACE Analyst, VoZ. 105 temperature together with the improved hot-block design described under Experimental, made the method satisfactory for the complete dissolution of 5 g of resistant high-fat tissue in the hot-block (Fig. 1). As large volumes of acid are used for the dissolution of the large sample sizes used, the heating process in the aluminium block had to be controlled. It is important that the initial temperature be low to prevent the violent boiling-over of nitric acid.A ramp-heating step was used at 60 "C for 30 min, followed by 150 "C until dissolution was complete, in order to achieve the controlled heating. This arrangement proved very practical as a large number of fish samples could be treated with a minimum of attention. Under these con- ditions the nitric acid will boil continuously until it is completely evolved and only sulphuric acid is left. At this point, charring will definitely occur, owing to the large amount of carbon in these samples. While this is undesirable for volatile metals, such as mercury and selenium, because of the reducing environment set up, it h'as no detrimental effect on the relatively non-volatile elements of interest.Indeed, charring is advantageous in this method because it provides a means of breaking down resistant fat tissue. Sulphuric acid, being a strong dehydrating agent, effectively reduces the organic matter to carbon black, which can easily be cleared by the subsequent use of hydrogen peroxide. If charring is not allowed, by main- taining the nitric acid content in the digestion tube many more additions of acid would be required before all the organic matter is oxidised. By using a charring step no organic matter is left behind, which may interfere with the subsequent chelation - solvent extraction step. It should be noted that while charring, which occurs in dry-ashing methods, may cause the loss of some of the volatile elements such as cadmium and lead, charring in wet digestions has no such effect owing to the strong acid environment.As indicated earlier, the natural levels of cadmium, nickel and lead are low enough that direct flame analysis of aqueous extracts does not provide satisfactory detection limits. The APDC - MIBK system, as outlined in the Water Quality Branch "Analytical Methods Manual,"' recommends the use of a buffer to adjust the pH of water samples to 3.5 for optimum simultaneous extraction of the six trace metals. It was found, however, that APDC complexes of cadmium, nickel and lead were distributed into MIBK with apparently little or no loss in solvent extraction efficiency when the buffering step was eliminated. This is fortunate as the sample extracts are highly acidic and buffering is not easily obtained, making it impractical for the desired routine method. Furthermore, much lower reagent blanks were obtained by eliminating the large amount of buffer needed.TABLE I DETECTION LIMITS TO BE EXPECTED FROM PROPOSED METHOD Chromium, copper and zinc were analysed by direct aspiration and cadmium, nickel and lead were analysed by solvent extraction. Detection limitslpg g-1 Metal A* Bt c: limitlpg g-l f A 'I Adopted detection Cadmium .. .. . . 0.01 - 0.02 0.02 Nickel . . .. . . 0.01 0.05 0.04 0.05 Lead .. .. . . 0.01 - 0.08 0.1 Chromium .. . . 0.1 0.2 0.09 0.2 Zinc .. .. . . 0.1 - 0.2 0.2 Copper .. .. . . 0.1 - 0.2 0.2 Normal levels found in fishlpg g-l 0.02-0.05 0.05-0.19 0.1-0.2 0.2-0.3 0.57-1.3 0.9-1.5 * Lowest detectable signal.t Concentration that gives a coefficient of variation of over 20%. $ Twice the standard deviation at the lowest detectable level. Table I shows the practical detection limits that can be expected from the proposed method for the six trace metals of interest. As there is no universal agreement on a definition for detection limit, three commonly used methods of calculating it were used, as shown in Table I, and the highest of the three was used to obtain the recommended value. In this way the adopted value can more easily be reproduced by any laboratory. As shown in Table I these values are well below the required detection levels of the six trace elements found in whole fish tissue and thus the recommended technique becomes an effective tech-February, I980 METALS FROM FISH TISSUE BY HOT-BLOCK DIGESTION AND AAS TABLE I1 STATISTICAL DATA FOR TISSUE FROM CONTAMINATED FISH 129 All data are based on 10 replicate determinations.increasing levels in fish tanks until mortality. Fish were exposed to cadmium, nickel and lead at Copper and zinc levels were naturally found levels while chromium was spiked into the rainbow trout homogenate (4) sample. Cd Ni Pb c u Zn Cr Mean/ R.S.D.,* Mean/ R.S.D.,* Mean/ R.S.D.,* Mean/ R.S.D.,* Mean/ R.S.D.,* Mean/ R.S.D., -c----------* Tissue type wgg-' % w g - ' % wgg-' % wgg-l % w g - ' % wgg-' % Rainbow trout Rainbow trout Rainbow trout homogenate (3) .. . . 0.11 9.1 0.20 17 0.94 8.3 0.53 6.4 11.8 9.3 - - Rainbow trout homogenate (4) ... . 0.06 20 0.21 8.9 0.83 10 0.58 7.1 12.6 3.9 2.2 7.9 Fish liver .. .. .. 2.2 14 0.92 11 8.0 17 - - - - - - Fish kidney .. .. .. 7.1 9.2 1.9 5 36 11 - - - - - - - - - - 1.06 5.4 24.6 6.7 0.21 15 Coho jack .. .. .. - - homogenate (1) .. . . 0.10 12.1 0.16 15 0.92 4.3 0.6 13 10.9 4.8 - - homogenate (2) .. . . 0.10 8.2 0.15 13 0.98 6.0 0.80 8.8 10.9 3.8 - - R.S.D. = relative standard deviation. nique in determining trace metal levels for routine monitoring purposes. Table I1 shows some precision data for several types of fish tissue for the metals under study. The precision data for cadmium, nickel and lead are for unspiked fish tissue, but the fish had been subjected to metal-contaminated fish tanks for toxicity studies. Thus, the metal content was that which was ingested by the fish during their stay in these tanks.Therefore, these data give the true precision of recovery of the metals from the fish tissue, as compared with the reproducibility of spike experiments. The data for copper and zinc were natural levels found in fish and were therefore also incorporated into the fish tissue. Rainbow trout homogenate, sample 4, was spiked with chromium as no high levels of chromium were found in most of the fish analysed. The data in Table I1 show that the precision of the method depends not only on the level of metal but also on the nature of the sample. TABLE I11 RECOVERY AND PRECISION DATA FOR STANDARD REFERENCE MATERIALS Data are based on 10 replicate determinations. Concentrationlwg g-' A r % Material Value Cd Ni Pb Cr c u Zn NBS* orchard leaves .. Found 0.15 * 0.05 Certified 0.11 f 0.1 NBS* bovine liver . . Found 0.30 f 0.07 Certified 0.27 & 0.04 IAEAt fish flesh MA-A-2 .. .. .. Found 0.13 & 0.04 Tentative by IAEA 0.16 -& 0.04 * NBS = National Bureau of Standards. t IAEA = International Atomic Energy Agency. 1.18 f 0.08 1.3 & 0.2 - - 1.34 f 1.47 1.2 -f 0.2 42 & 1.7 45 f 3 0.28 f 0.04 0.34 & 0.08 0.40 & 0.04 0.7 f 0.2 2.7 f 0.17 2.6 f 0.3 0.088 f 0.012 2.92 f 0.13 2.9 f 0.6 - 11.2 + 0.18 12 z 1 187 f 2.3 193 -f 10 3.7 f 0.46 4.6 f 0.4 25.3 f 0.5 25 f- 3 131 f 1.4 130 f 13 32.4 f 0.6 36 f 3 Recovery experiments were made by spiking fish samples with the six trace elements before digestion and then processing them in the same manner as the unspiked samples.Recoveries ranged from 90 to 110% in all instances and were classed as acceptable. To check further on the recovery of the method some certified standard reference materials were used to test the method. Table I11 shows the data for the mean of 10 replicate analyses for these reference materials. It can be seen that the data are satisfactory in relation to certified levels. The proposed method is a simple, rapid, multi-element technique with satisfactory precision, accuracy and detection limits. As such, it provides an effective means for the routine monitoring of large numbers of samples.130 AGEMAIN, STURTEVANT AND AUSTEN The authors thank D. M. Whittle and P. V. Hodson for supplying the fish tissue. References 1. Evans, W. H., Read, J. I., and Lucas, B. E., Analyst, 1978, 103, 580. 2. Analytical Methods Committee, Analyst, 1960, 85, 643. 3. Hendzel, M. R., and Jamieson, D. M., Anal. Chem., 1976, 48, 926. 4. Bishop, J. N., Taylor, L. A., and Diosady, P. L., “High Temperature Acid Digestion for the Deter- mination of Mercury in Environmental Samples,” Ministry of the Environment, Laboratory Services Branch, Ontario, Canada, March 1975. Christian, G. D., and Feldman, F. J ., “Atomic Absorption Spectroscopy,” Wiley-Interscience, New York, 1970. “Properties and Essential Information for Safe Handling and Use of Perchloric Acid Solution, ” Chemical Safety Data Sheet SD-11, Manufacturing Chemists Association, Washington, D.C., 1965. “Analytical Methods Manual,” Inland Waters Directorate, Water Quality Branch, Ottawa, 1974. 6. 6. 7. Received August 20th, 1979 Accepted September 24th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500125

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, February 1980, "01. 105, pp. 131-138 131 Rapid Spectrophotometric Method for the Determination of Monofluorophosphate John L. Stuart and Edward J. Duff Department of Child Dental Health, Turner Dental School, The University, Bridgeford Street, Manchester, MI5 6FH A method is described for the rapid spectrophotometric determination of orthophosphate in the presence of monofluorophosphate. Subsequent acid hydrolysis of monofluorophosphate enables this ion to be determined in the same sample. The technique is based on a single-reagent modification of the molybdophosphovanadate procedure. Using this method as described it is possible to determine up to 0.5 mg of phosphorus but this can be extended to 1.8 mg of phosphorus using an alternative procedure. Free phosphorus can be determined in the presence of up to a 50-fold excess of monofluoro- phosphate.The method is especially well suited for following hydrolysis reactions of monofluorophosphate. Keywords ; Monojluorophosphate determination ; orthophosphate ; spectro- photometry ; molybdophosphovanadate reagent The analytical determination of monofluorophosphate (MFP) or its ion (P03F2-) presents a number of difficu1ties.l The ion hydrolyses rapidly in strongly acidic or strongly basic solutions, especially at elevated temperature~.~-5 The P03F2- anion lacks characteristic properties of analytical importance and a method for its direct determination has yet to be rep~rted.~ Established methods for the deter- mination of MFP generally adopt a separation step to remove contaminating species followed by hydrolysis of MFP to F- and PO,3-, one or both of which is then determined.Methods for determining MFP based on the spectrophotometric determination of ortho- phosphate have been developed. Ericsson7y8 suggested that orthophosphate could con- veniently be determined in the presence of MFP by an acid - molybdate reduction technique, provided that the hydrolysis reaction was controlled by rapid and accurately timed addition of reagents. Ingramg utilised the ascorbic acid reduction modification of the molybdenum blue procedurelo to follow the acid hydrolysis of MFP. The method had the disadvantage of requiring frequent absorbance measurements. Also, the limit of determination for the method was 40 ,ug of phosphorus and, like all molybdenum blue methods, it was subject to a number of interferences.In a later work6 Ingram reverted to a silver orthophosphate precipitation technique. In the course of our work on the acid hydrolysis of MFPl it became necessary to determine the PO,3- in the presence of P03F2- in a large number of samples. This paper describes a simple spectrophotometric procedure and is based on the single-reagent modification11*12 of the molybdophosphovanadate technique.13 Experimental Reagents All reagents were of analytical-reagent grade except where stated. De-ionised distilled water was used throughout for the preparation of solutions. Standard orthophosphate solution. Dissolve 2.197 g of potassium &hydrogen ortho- phosphate in water and dilute to 1 1 (1 ml = 0.5 mg of phosphorus).Acetate bufer solution. Adjust the pH to 5.2 using 10 M sodium hydroxide solution and dilute the solution to 1 1 with water. Molybdophosphovanadate spectrophotometric reagent. Dissolve 1.17 g of ammonium meta- vanadate in a mixture of water (400 ml) and 8 M perchloric acid (25 ml); then dilute the solution to 500 ml with water (solution A). Dissolve 35 g of ammonium molybdate tetra- hydrate in water and dilute to 1 1 (solution B). Prepare sodium sulphate solution (0.75 M) by dissolving 53 g of anhydrous sodium sulphate in water and diluting to 500 ml (solution Add glacial acetic acid (57 ml) to water (443 ml). c>.132 STUART AND DUFF: SPECTROPHOTOMETRIC METHOD Analyst, VoZ. 105 Mix 70-72 yo perchloric acid, solution A, solution B and solution C (1 + 2 + 4 + 2) and store in a polythene bottle.Final spectrophotometric reagent. Apparatus pH meter. Pye Dynacap. Spectrophotometers. Perkin-Elmer, Model 124, ultraviolet - visible double-beam grating spectrophotometer connected to a Perkin-Elmer, Model 56, self-balancing multi-range recorder and fitted with a Hitachi 124-0319 10-mm thermostated cell holder. Also, a Varian, Model 654, ultraviolet - visible spectrophotometer with digital read-out calibrated to register concentrations directly and fitted with a thermostatic cell holder. Procedure Dilute to about 20 ml with water. Cool in ice-water to about 4 "C. Add the sample and dilute to 25ml. Measure the absorbance at 407 nm (or at the observed Amax.). Prepare a calibration graph using known amounts of standard phosphorus solution containing between 0 and 0.5 mg of phosphorus (as KH,PO,).Treat unknown samples similarly and calculate the phosphorus concentration from the calibration graph. Pipette the prepared reagent (10 ml) into a 25-ml calibrated flask. Alternative Procedure sensitivity if all volumes are doubled and 50-ml calibrated flasks are used. Samples containing up to 1.8 mg of phosphorus may be analysed with a resultant loss of Results and Discussion Sensitivity and Range The Perkin-Elmer spectrophotometer used in much of this work was fitted with a pen recorder incorporating a variable voltage control. The range and sensitivity of the method using this equipment were, to a large extent, dependent on the recorder setting. When the alternative analytical procedure was used in conjunction with a recorder-voltage span of 50 mV, the calibration graph was linear for up to 1.8 mg of phosphorus and results were reproducible for up to at least 2.0 mg of phosphorus. 0.5 1 .o 1.5 2.0 2.5 Phosphorus concentration/mg Fig.1. Calibration graphs for the alternative molybdophospho- vanadate procedure: 20 ml of reagent, 60 ml total volume, 1 cm path-length cells, absorbance measured a t 407 nm. Recorder voltage range: A, 0-1 absorbance scale, 100 units = 20 mV; B, 0-2 absorbance scale, 100 units = 20 mV; and C, 0-2 absorbance scale, 100 units = 60 mV.February, 1980 FOR THE DETERMINATION OF MONOFLUOROPHOSPHATE 133 Under these conditions, however, the sensitivity was poor; a change in concentration of 0.5 mg of phosphorus produced a deflection of only 10% on the recorder.Reduction of the recorder-voltage span to 20 mV produced a four-fold increase in sensi- tivity, but reduced the upper limit to 1.2 mg of phosphorus. Fig. 1 shows typical calibra- tion graphs produced under these conditions. When the recommended procedure was in use the linear range was reduced to 0-0.5 mg of phosphorus, but the sensitivity was greatly improved. Fig. 2 shows that under these conditions 0.5 mg of phosphorus represented a deflection of 86% and 0.05 mg of phosphorus a deflection of 9 & 1%. 125 - v) CI .- = l o o - t c 2 7 5 - + .- lu oi 5 5 0 - e 2 2 2 5 - I I I I I I 0.25 0.5 0.75 1.0 1.25 1.5 Phosphorus concentration/mg Fig. 2. Sensitivity of the molybdophosphovanadate procedure. Absorbance scales : A (recommended pro- cedure), 100 units = 20 mV; B (alternative procedure), 100 units = 20 mV; and C (alternative procedure), 100 units = 60 mV.Reproducibility The reproducibility of the method was examined using 10 aliquots, each containing 0.25 mg of phosphorus as KH,PO,. The mean recovery was 0.2455 mg and the standard deviation was 0.0083 mg. This represents a recovery of 98.2% with a standard deviation of 3.33%. Effect of pH The measured pH of the final spectrophotometric solution was 0.4. The effect of increasing the pH was examined by preparing a series of molybdovanadate reagents identical with the standard reagents except that the perchloric acid concentration used in this preparation was adjusted to give reagents of varying pH. These were then used to prepare calibration graphs in the normal way using 20 ml of reagent in 50-ml calibrated flasks.The standard reagent with a pH of 0.4 gave a linear graph for up to 1.5 mg of phosphorus, which was the highest phosphorus concentration used. As the pH was increased, the range over which linear calibration graphs were obtained was progressively reduced until at pH 5.0 it was linear only up to 0.6 mg of phosphorus. However, the sensitivity of the method remained essentially constant over the linear range. Effect of Fluoride and Temperature on Colour Stability Interference by fluoride ions would seriously restrict the value of the method for the determination of MFP in the presence of its hydrolysis products. Possible interference was examined by adding aliquots of sodium fluoride solution to the reagents: (i) before the addition of phosphorus as KH,PO,; (ii) in admixture with phosphorus; and (iii) to the developed, molybdophosphovanadate solution.134 STUART AND DUFF : SPECTROPHOTOMETRIC METHOD Analyst, Vd.105 When the experiment was carried out at room temperature (18 "C) there was no interference from fluoride ion at concentrations of up to 0.02 M in the final solution (9.5 mg of fluoride), whatever the stage at which the fluoride was added. However, if the colour was developed at elevated temperatures (60 "C) fluoride ion could only be tolerated up to a final concentration of 0 . 0 0 4 ~ . If added at a concentration of 0.02 M, fluoride exhibited a pronounced positive effect. This was a relatively constant interference equivalent to about 0.08mg of phosphorus and bore no relationship to the amount of phosphorus present.It is thought that this effect may be due to attack by hydrogen fluoride on the walls of the glass calibrated flasks used in the experiment when the acidic reagent is heated at 60 "C in the presence of fluoride. Time Required for Full Colour Development Kitson and Mellon14 suggested that the presence of fluoride might increase the time required for full colour development. The time needed for development of maximum colour intensity in the presence or absence of fluoride ion at a concentration of 0.004 M was noted for samples where the colour was developed inside or outside the spectrophotometer. In all instances this was found to be between 1.5 and 3 min. After 2.0 min 98% of the final absorbance was attained in five of the six samples studied. Thus, fluoride present at concentrations similar to those produced during hydrolysis of MFP did not affect the method.Determination of Total Phosphorus in MFP Weighed samples of MFP (0.1-3 g) were added to 50 ml of 1.0 M hydrochloric acid and heated at 60 "C for 30min. Aliquots were analysed by the described method. Table I shows that the recovery of phosphorus was 98.25% of the theoretical value and the standard deviation was 1 .36y0. TABLE I RECOVERY OF PHOSPHORUS AFTER HYDROLYSIS OF MFP AT 60 "c Volume of Phosphorus found Phosphorus Purity of MFP Mass of 1.0 M HCl/ Sample Number of in aliquot (mean)/ found in MFP, (calculated), 0.360 50 0.20 4 0.300 20.83 96.76 0.720 60 0.10 4 0.305 21.18 98.38 1.440 60 0.05 4 0.304 21.11 98.06 2.880 50 0.05 4 0.620 21.53 100.0 MFP/g ml volume/ml of aliquots mg % % Mean 98.25 Standard deviation 1.36 Comparison with Alternative Method Four samples of the same batch of MFP were analysed by a method involving potentio- metric titration after precipitation of silver orthophosphate.12 This gave a recovery of phosphorus of 98.65% with a standard deviation of O.Slyo (Table 11).TABLE I1 RECOVERY OF PHOSPHORUS FROM HYDROLYSED MFP BY POTENTIOMETRIC TITRATION Sample Titre of Total phosphorus/mg MFP Mass of MFP volume/ 0.0998 M NaCl ,---*-, purity, per 100ml/g ml solution/ml Found Calculated yo 0.225 1 20 9.2 47.44 48.46 97.9 0.294 1 20 12.2 62.91 63.31 99.4 0.276 0 20 11.3 58.26 69.42 98.0 0.274 2 20 9.3 47.95 48.27 99.3 Mean 98.66 Standard deviation 0.81February, 1980 FOR THE DETERMINATION OF MONOFLUOROPHOSPHATE 135 Determination of Total Phosphorus in Mixed Solutions A number of samples containing mixed MFP - KH,PO, solutions were analysed by the method described above.Table I11 shows that the mean recovery of phosphorus from MFP for nine samples was 98.71% with a standard deviation of 2.40%. TABLE I11 RECOVERY OF PHOSPHORUS FROM MIXED MFP - ORTHOPHOSPHATE SAMPLES Phosphorus added as- Phosphorus found/mg MFP MFP/mg KH,PO,/mg Total From MFP % r purity, 0.315 0.630 0.945 0.315 0.630 0.945 0.316 0.630 0 0 0 0 0.315 0.315 0.315 0.630 0.315 0.315 0.300 0.625 0.910 0.640 0.910 1.255 0.935 0.950 0.312 0.300 0.625 0.910 0.325 0.620 0.950 0.310 0.620 95.24 99.21 96.30 98.41 98.41 98.41 103.2 100.5 Mean 98.71 Standard deviation 2.40 Purity of MFP When results from all methods were considered, the mean content of the MFP used, expressed as a percentage of the theoretical value, was 98.6% with a standard deviation, determined over 16 samples, of 1.8%.Determination of Free Orthophosphate in the Presence of MFP, Acetate Buffer Solution and Added Orthophosphate When studying the rate of hydrolysis of MFP it is desirable to determine both liberated F- and liberated Po,3- on the same sample. For reasons discussed elsewherel it was found that the optimum buffer solution for this purpose was acetate buffer, pH 5.2. The effect of 70 E D B C A / I - 1 0 5 10 15 Time/min Fig. 3. Absorbance of PO,*- - MFP mixtures containing acetate buffer solution, over a period of time.A, 0.25 mg of P as KH,PO,; B, 0.25 mg of P as KH,PO, + 1.44 mg of MFP; C, 0.25 mg of P as KH,PO, + 4.32 mg of MFP; D, 0.25 mg of P as KH,PO, + 7.2 mg of MFP; E, 0.25 mg of P as KH,PO, + 14.4 mg of MFP.136 Phosphorus added as mg 0 0 0 0 0 0.25 0.25 0.25 0.25 0.25 0.25 K H ,PO d STUART AND DUFF: SPECTROPHOTOMETRIC METHOD TABLE IV RECOVERY OF ADDED ORTHOPHOSPHATE WITH TIME IN PRESENCE OF AnaZyst, VoE. 105 MFP AND ACETATE BUFFER MFP added/ mg 1.44 2.88 4.32 7.20 14.40 0 1.44 2.88 4.32 7.20 14.40 Ratio of Phosphorus found/mg MFPto 7 A phosphorus 2 min 4 min 5 min 8 min 10 min - 0.006 1 - 0.012 2 - 0.0146 - 0.021 3 - 0.042 6 - 0.252 5.76 0.257 8.64 0.257 17.28 0.266 28.80 0.273 57.60 0.297 0.006 1 0.0122 0.0159 0.0240 0.048 2 0.250 0.255 0.257 0.266 0.274 0.299 0.006 1 0.0122 0.017 1 0.0244 0.0488 0.250 0.255 0.258 0.266 0.274 0.302 0.006 1 0.0122 0.0183 0.029 8 0.061 0 0.250 0.255 0.258 0.266 0.276 0.307 0.006 1 0.013 7 0.0198 0.0300 0.067 1 0.250 0.255 0.258 0.266 0.278 0.313 TABLE V RECOVERY OF ADDED MFP WITH TIME : EFFECT OF ADDED ORTHOPHOSPHATE The results are for phosphorus (mg) as orthophosphate recovered from MFP.7 15 rnin 0.0070 0.017 1 0.024 3 0.039 6 0.085 0 0.250 0.255 0.258 0.266 0.280 0.323 Time/ Orthophosphate min addition* 2 A B 5 A B 10 A B 15 A B Sodium monofluorophosphate added/mg 1 14.4 0.042 6 0.0470 0.048 8 0.0520 0.067 1 0.063 0 0.085 0 0.073 0 0 <0.005 <0.005 <0.005 <0.005 (0.005 <0.005 <0.005 <0.005 1.44 0.006 1 0.005 0 0.006 1 0.0050 0.006 1 0.005 0 0.0070 0.005 0 2.88 0.0122 0.007 0.0122 0.008 0.013 7 0.008 0.017 1 0.008 4.32 0.0146 0.0160 0.017 1 0.0160 0.0198 0.0160 0.024 3 0.0160 ~ ~~ 7.20 0.021 3 0.0230 0.024 4 0.024 0 0.0300 0.028 0 0.039 6 0.0300 * A, No added orthophosphate; B, 0.25 mg of phosphorus added as KH,PO, (this figure is given as Pfound - 0.25 mg).TABLE VI RELEASE OF PHOSPHORUS FROM MFP WITH TIME : EFFECT OF ADDED ACETATE BUFFER The results are for phosphorus (mg) found as orthophosphate. Time/ Acetate buffer rnin addition* 2 A 0 5 A 0 10 A 0 15 A 0 r 0 0.006 <0.005 0.006 <0.005 0.006 <0.005 0.006 <0.005 Sodium monofluorophosphate added/mg 1.44 2.88 4.32 7.20 0.006 1 0.0122 0.0146 0.021 3 0.000 0.0061 0.0110 0.0183 0.0061 0.0122 0.0171 0.0244 0.0030 0.0091 0.0171 0.0305 0.0060 0.0137 0.0198 0.0300 0.0043 0.0122 0.0238 0.0457 0.0070 0.0171 0.0243 0.0396 0.0061 0.0183 0.0305 0.0598 A 1 14.4 0.0426 0.0426 0.0488 0.061 0 0.067 1 0.0976 0.085 0 0.0134 * A, Acetate buffer (4 ml) added; 0, no acetate buffer.February, 1980 FOR THE DETERMINATION OF MONOFLUOROPHOSPHATE I37 this buffer on the recovery of phosphorus from KH,PO, and/or MFP solutions was examined by adding known amounts of phosphorus (as KH,PO,) and MFP to 4 ml of prepared acetate buffer solution and analysing aliquots of the resultant mixture.Table IV shows the recovery of added phosphorus from the reagent, as a function of time when different amounts of MFP were added. Table V, similarly, shows the recovery of orthophosphate from MFP when additional amounts of phosphorus (as KH,P04 solution) were added, and Table VI illustrates the effect of added acetate buffer on the release of phosphorus from MFP as a function of time.These tables show that the addition of large amounts of acetate buffer solution inhibits the hydrolysis of MFP by the reagent at room temperature. The evidence from Tables IV and VI indicates that the acetate buffer used in this study contained about 0.001 5 mg ml-l of free phosphorus. In the absence of MFP, full colour development occurred at the same rate whether or not acetate buffer was present. This suggests that acetate buffer does not inhibit the rate of colour development. The observed effect must' therefore be due to inhibition of hydrolysis. From Table V it can be seen that at high concentrations of MFP relative to phosphorus (above 4.32 mg of MFP in the presence of 0.25 mg of phosphorus) the earlier recoveries are higher than expected and also higher than those for the hydrolysis of MFP alone.However, even at an MFP to phosphorus ratio of 51.6: 1, the hydrolysis reaction is inhibited by the addition of free orthophosphate. Fig. 3 shows the absorbance of a number of - MFP mixtures containing acetate buffer solution recorded over a period of 15 min. There is virtually no hydrolysis over this period unless the MFP to phosphorus ratio exceeds 17: 1 and it is only at an MFP to phosphorus ratio of over 50: 1 that reagent hydrolysis is a serious problem. This experiment was conducted at room temperature (18 "C). TABLE VII REPRODUCIBILITY OF THE METHOD : FIRST BLIND TRIAL Phosphorus added as- - MFP/mg KH,PO,/mg 0.158 0 0.158 0.158 0.210 0 0.052 5 0 0.079 0 0.210 0 0.1575 0 0.210 0.158 0.105 0.105 0.263 0.052 5 0.236 0 0.105 0.210 0 0.315 0.157 5 0.1575 0.210 0.105 0.263 0.052 5 0.236 0.079 Phosphorus found as orthophosphate/mg Free Total Phosphorus phosphorus phosphorus from MFP 0 0.165 0.165 0.158 0.315 0.150 0 0.215 0.215 0.01 0.055 0.054 0 0.075 0.075 0 0.215 0.215 0 0.150 0.150 0.158 0.360 0.205 0.105 0.205 0.100 0.052 0.312 0.260 0.01 0.241 0.240 0.215 0.315 0.100 0.315 0.315 0 0.165 0.315 0.150 0.110 0.310 0.200 0.055 0.315 0.260 0.075 0.315 0.240 I A $ Orthophosphate, yo MFP, % MFP recovered, 104.4 94.9 102.4 102.9 94.9 102.4 95.2 97.6 95.2 98.9 101.7 95.2 95.2 95.2 98.9 101.7 Y O - Mean recovery .. .. 100.98 98.54 Standard deviation .. 3.27 3.53 Reproducibility of the Method in MFP Determinations Seven samples, each containing 0.1575 mg of phosphorus as MFP, were analysed for free and total phosphorus by the methods described. The mean free phosphorus found was 0.007 mg and the mean total phosphorus was 0.155 7 -J= 0.002 8 mg (which was 98.82% of the calculated result, standard deviation 1.82%).138 STUART AND DUFF Blind Trial To establish the suitability of the method under actual experimental conditions two series of samples were analysed for free and total phosphorus. These samples contained unknown amounts of MFP and orthophosphate. Hydrolysis was carried out at 60 "C in polythene bottles. Tables VII and VIII show that there is good reproducibility between samples containing different proportions of phosphorus and MFP.The mean recoveries from these trials compared well with those where the amounts of MFP and phosphorus were known, but the standard deviation was higher [3.75y0 compared with 1.82% (see above) or 2.4% (see Table III)]. TABLE VIII REPRODUCIBILITY OF THE METHOD : SECOND BLIND TRIAL Phosphorus added as- A I Phosphorus found as orthophosphate/mg Total r A \ KH,PO,/ phosphorus/ Free Total Phosphorus MFP/mg mg mg phosphorus phosphorus from MFP 0.248 0.062 0.310 0.056 0.300 0.244 0 0.310 0.310 0.310 0.310 0 0.186 0.124 0.310 0.120 0.306 0.186 0.280 0.030 0.310 0.027 5 0.300 0.273 0.062 0.248 0.310 0.250 0.310 0.060 0.156 0.165 0.310 0.145 0.305; 0.160 0.093 0.217 0.310 0.210 0.310 0.100 0.031 0.279 0.310 0.265 0.310 0.045 0.217 0.093 0.310 0.080 0.300 0.220 Orthophosphate, yo MFP, yo Mean recovery .. .. 98.4 100.7 Standard deviation . . 1.61 3.76 Conclusion The described method offers a simple and rapid procedure for the determination of free orthophosphate in the presence of monofluorophosphate. Subsequent rapid acid hydrolysis allows MFP to be determined. It appears to offer many advantages over the molybdenum blue reduction technique, particularly in its relative freedom from interference by fluoride ion. The method is of particular value for following the acid hydrolysis of monofluorophosphate.1 The help and advice of Dr. Kjell Bjorvatn, Dental School, University of Bergen, Dr. Philos. Sverre Hauge, University of Bergen and Professor Dr. Philos. Tormod Mgrch, Department of Pedodontics, University of Bergen, are greatly appreciated. Part of this work was undertaken at the University of Bergen. One of us (J.S.) expresses appreciation to the University of Bergen for the research facilities and financial aid provided. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Stuart, J. L., PhD Thesis, University of Manchester, 1978. Devonshire, L. N., PhD Thesis, University of Oklahoma, USA, 1954. Devonshire, L. N., and Rowley, M. M., Inorg. Chem., 1962, 1, 680. Clark, H. R., and Jones, M. M., Inorg. Chem., 1971, 10, 28. Thompson, K. M., PhD Thesis, University of Bath, 1974. Ingram, G. S., Caries Res., 1977, 11, 30. Ericsson, Y., Caries Res., 1967, 1, 144. Ericsson, Y., Acta Odont. Scand., 1949, 8, suppl. 3. Ingram, G. S., PhD Thesis, University of London, 1967. Murphy, J., and Riley, J. P., Anal. CAim. Acta, 1962, 27, 31. Barton, C. J., Anal. Chem., 1948, 20, 1068. Stuart, J. L., MSc Tliesis, University of Manchester, 1974. Misson, G., Chem. Z., 1908, 32, 633. Kitson, R. E., and Mellon, M. G., Ind. Eng. Chem. Anal. Ed., 1944, 16, 379. Received May 22nd, 1979 Accepted August Sth, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500131

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, February 1980, Vol. 105, $9,139-146 139 Automatic Determination of Boron (0.10-10.0 mg I-’) in Raw and Waste Waters R. A. Edwards Forth River PurQication Board, Colinton, Edinburgh, EH13 OPH An automated method, employing azomethine-H in aqueous medium, for the determination of boron in raw waters and effluents is described. The method is capable of a limit of detection of 0.1 mg 1-1 and the response is essentially linear to 4.0 mg 1-l, although a calibration graph is required above this level. A wide range of possible interferences was tested, but none proved of practical importance except sample colour, for which adequate correction is provided. Recovery after spiking a typical range of raw and waste waters was adequate. Keywords ; Boron determination ; A utoAnalyzer ; water analysis The level of boron in raw waters has tended to increase in recent years, resulting from its greater use in cleansing materials and in industrial pr0cesseq.l As sewage treatment does not significantly reduce this level the increase is passed to surface waters, which may then be utilised for crop irrigation or potable water supply.Many varieties of fruit can tolerate no more than 0.5 mg 1-1 of boron2 in irrigation water. The US Environmental Protection Agency3 recommends limits of 0.75 mgl-1 for most fruits, 1.0mg1-1 for most cereals, potatoes, peas and tomatoes and 2.0 mg 1-1 for tolerant species including sugar beet, turnips and cabbage. The Anglian Water Authority4 apply criteria such that the “maximum desirable” limit for crop irrigation is 0.5 mg 1-1 and the “maximum permissible” limit is 1.0 mg 1-l.They also apply criteria of 0.8 and 1.0-1.2 mg l-l, respectively, to potable water abstractions. Several reagents have been developed for the routine determination of boron, but most require a final stage in concentrated sulphuric acid. Apart from the hazards involved in pumping the concentrated acid the procedures are difficult to automate, principally because of density effects resulting from small changes in acid concentration. These methods include the carminic acid procedure, which has been automated by the Water Pollution Research Laboratory. It has not been found possible to obtain the precision we require using this method. The curcumin procedure does not involve these dangers, but includes an evaporation stage unsuitable for automation.A method employing ferroin has been developed6 involving extraction into chloroform, but this method suffers from interference by anionic detergents. The principle adopted employs azomethine-H (see Fig. l), a readily prepared condensation product of H-acid (8-aminonaphth-l-ol-3,6-disulphonic acid) and salicylaldehyde. This reagent is very sensitive to borates, forming, in aqueous medium, a yellow ion-associated compound by reversible reaction. The method was developed by Shanina et aL7 A manual procedure is employed by the Yorkshire Water Authority8 and an automated method for the determination of boron in plant tissue has been developed by Basson et aL9 This laboratory has further developed the procedure for use with the Technicon AutoAnalyzer 2 and to meet its own requirements of precision and sensitivity. SO, H I SO, H Fig.1. Azomethine-H.140 EDWARDS : AUTOMATIC DETERMINATION Analyst, Vol. 105 Experimental Apparatus A sample cam of 30 h-l is employed, with a 2: 1 sample to wash water ratio; 410-nm wave- length filters are fitted. A Technicon AutoAnalyzer 2 system is fitted with the manifold shown in Fig. 2. Sampler ;I min-1 Q cv 4 .- Ln 0 157-0248 170.01 03 8 z 9 0 cv .- In 0 % F 10- mm flow cell 410 nm Wash Sample Air Azomethine reagent (or background correction reagent) Fig. 2. Manifold arrangement for AutoAnalyzer 2. Sample Preservation The samples are fdtered through Whatman grade C glass-fibre filter-papers immediately after collection and stored in completely filled polythene bottles at about 4 "C until required for analysis.Reagents was used in all solutions. All reagents were of analytical-reagent grade unless stated otherwise. De-ionised water Boric acid. Dried to constant mass at 105 "C. H-acid (8-aminonaphth-l-ol-3,6-disul~honic acid, monosodium salt). Salicylaldehyde. Concentrated hydrochloric acid. Ethunol. Industrial methylated spirit. ascorbic acid. EDTA (ethylenediaminetetraacetic acid, disodium salt dihydrate). Acetic acid. Ammonium acetate. Brij concentrate. Dissolve 1 g of Brij 35 (polyoxyethylene lauryl ether) in water and dilute to 100ml. Potassium hydroxide solution, 10% m/V. Dissolve 10 g in water and dilute the solution to 100ml. Azomethine-H. Dissolve 18 g of H-acid in 1 1 of water with gentle heating, neutralise to pH 7 0.5 with 10% potassium hydroxide solution and filter if necessary.Add con- centrated hydrochloric acid, to pH 1.5 & 0.1, while still warm. Add 20 ml of salicylaldehyde and stir vigorously while heating gently (not exceeding 40 "C) for 1 h. Allow the azo- methine-H to settle for 16 h. Centrifuge and discard the supernatant liquid, slurry the residue with ethanol and filter through a Whatman grade C glass-fibre filter-paper. Dry at 105 "C for 3 h and store in a desiccator. The yield of azomethine-H should be approximately 18 g. SynchemicA grade (Hopkin & Williams, Chadwell Heath, Essex). 36% m/m HCI.February, 1980 OF BORON IN RAW AND WASTE WATERS 141 Add 1O.Og of EDTA and 300 g of ammonium acetate and warm to dissolve. Adjust to pH 5.2 & 0.1 with acetic acid or ammonia solution if necessary.Dissolve 900 & 10 mg of azomethine-H and 2 g of ascorbic acid in 70 ml of water while heating gently (not exceeding 70 “ C ) , and dilute to 100 ml. Add 100ml of buffer solution and mix thoroughly. The reagent is stable for 2 d only. The solution without buffer will keep for 14 d provided that it is filtered prior to use. Background correction reagent. Dilute 100 ml of buffer solution with an equal volume of water. Standard solutions. A stock working standard solution of 80 mgl-1 of boron and an independent standard stock solution are prepared as follows. Dissolve 0.458 g of boric acid in water and dilute to 1 1 in a calibrated flask. Store in a polythene bottle. Prepare the independent standard stock solution in an identical manner. Prepare a 4.0 mg 1-1 of boron working standard solution by diluting 5.0 ml of the stock working standard solution to 100 ml with water.Prepare an independent standard solution containing 4.0 mg 1-1 of boron similarly by diluting the independent standard stock solution. Bufer solution. Azomethine reagent. Dilute 500ml of glacial acetic acid to 750ml with water. Add 1.0 ml of Brij concentrate. Procedure Adjust the “standard calibration” to the usual setting for 80% full-scale deflection of 4.0 mg 1-1 (or as required). Adjust the reference filter aperture to fully open and then close down three turns. Pump azomethine reagent through the system for 15min and establish a flat base line by use of the sample filter aperture control with final adjustment by the base-line control. Set 4.0mg1-1 to 80% full-scale deflection with the standard calibration control.An independent standard of 4.0mg1-1 should be included in the run of samples to allow for quality control. Following completion of the sample run, pump background correction reagent instead of the azomethine reagent. After 15 min, adjust the base line by opening the reference filter aperture and, if necessary, by adjusting the base-line control. Do not adjust the standard calibration or sample aperture control. Re-run the samples. The background is subtracted from the gross figure to give the actual boron concentration. Air blips may occur at the beginning and end of each peak due to increased reagent concentrations in the segments containing sampler-introduced air.These should not interfere with peak measurement provided that the stated sample cam is used. Set the base-line control nearly fully clockwise. 0.4 0.3 I 1 I 400 410 420 430 440 450 Wavelengthhm Fig. 3. Absorbance graphs for standard response (1 mg 1-11 A, With boron; and B, reagent relative to reagent blank. blank.142 EDWARDS : AUTOMATIC DETERMINATION Analyst, "01. 105 Discussion Absorbance occurs over a wide peak at 410420 nm, as illustrated in Fig. 3. The effect of azomethine-H concentration on response was not tested, but Basson et found that response increased with increased concentration, as illustrated in Fig. 4. 0.5 0.4 0.3 0, C m -e 0.2 51 P Q 0.1 0 2 4 6 8 10 12 Azomethine-H concentration/g I-' Fig. 4.Effect of azomethine-H concentration (plus L-ascorbic acid at 20g 1-l) on absorbance (after Basson et d.). A, 5 mg 1-l of boron; and B, 1 mg I-' of boron. They also found that at concentrations greater than 12 g 1-1 difficulty was experienced in obtaining a clear solution and they therefore recommended an azomethine-H concentration The effect of time on colour development was tested and the results are shown in Fig. 5. of 9.og1-? Time/min Fig. 5. Effect of time on colour development of 1 mg 1-1 of boron (relative to reagent blank).February, 1980 OF BORON IN RAW AND WASTE WATERS 143 The time dependence graph indicated that a development time of at least 20min was necessary for optimum response but difficulty was experienced in controlling the flow pattern with such long development times.A delay of 8 min was finally chosen, as this gives adequate precision and sensitivity. With a fixed development time of 8 min the effect of development temperature was investi- gated. The results are illustrated in Fig. 6. Also, long setting-up times are incurred. 15 20 30 40 50 60 70 Development temperature/OC Fig. 6. Effect of development temperature on response of 1 mg 1-l of boron. It can be seen that the response can be improved by over 10% by incorporating a heating bath at about 40 "C into the manifold, but this was not considered necessary for the require- ments of this laboratory. The response was found not to be linear over the whole range. Deviation from linearity was found to be significant at 4.0mgl-l, although not of practical importance, approxi- mately 99% of linear response being obtained.However, at 5.0 mg 1-1 a reduction to 96% of linear response was observed. All precision testing at this laboratory has assumed linearity over the range 04.0 mg 1-l. Precision and Accuracy Determination of boron in a range of raw waters and sewage effluents was found to be satisfactory relative to the standard carmine method,1° which involves destruction of organic matter at 500 "C. The river waters tested gave a mean of 0.2 mg 1-1 with a recovery of TABLE I RECOVERY OF BORON FROM REAL SAMPLES SPIKED AT 2.0 mg 1-1 Sample description* Clean river water 1 . . .. .. .. Clean river water 2 . . .. .. .. Clean peaty stream water . . .. .. Good sewage effluent . . . . .. .. Poor sewage effluent .. . . .. .. Septic tank effluent . . .. .. .. Industrial maltings effluent . . .. .. Detergent manufacture effluent . . .. Surface water from detergent manufacture Poor petrochemical treatment plant effluent Colour/Hazen units 200 70 500 40 60 40 500 50 30 50 Boron concentration in sample/ mg 1-1 <0.1 (0.1 <0.1 0.89 0.74 1.45 0.16 0.64 0.26 (0.1 Recovery, Significance, t % 100 f 5% 106.1 N.S. 103.7 N.S. 101.5 N.S. 103.3 N.S. 99.6 N.S. 99.4 N.S. 102.1 N.S. 104.4 N.S. 99.7 N.S. 109.8 S. See Table 11. t N.S. = not significant at 95% confidence level; S. = significant at 95% confidence level.144 EDWARDS : AUTOMATIC DETERMINATION Analyst, Vol. 106 TABLE I1 ANALYSIS OF SAMPLES USED FOR SPIKING RECOVERY TESTS Total Suspended oxidised Anionic solids a t Ammonia nitrogen Electrical detergent BOD/ lOS0C/ (asN)/ (asN)/ conductivity/ (Manoxol Sample mgl-1 mgl-' mgl-l mgl-l pH $3 cm-l OT)/mgl-1 Clean river water 1 .. . . 1.5 5 0.14 0.35 6.6 90 - Clean river water 2 . . . . 0.3 1 0.41 1.8 7.2 475 - Peaty stream water . . . . 2.1 7 0.29 0.8 6.6 100 - Good sewage effluent . . . . 9.6 17 2.6 28.0 6.8 580 - Poor sewage effluent . . . . 30.0 31 14.8 17.1 7.0 560 - Septic tank effluent . . .. - 56 17.2 0.2 7.4 800 - Maltings effluent . . .. 234 80 1.8 0.1 7.2 1200 - Detergent manufacture Surface water from Petrochemical treatment effluent . . .. . . 8.0 13 0.2 0.6 7.5 720 0.16 detergent manufacture . . 1.2 2 0.8 17.0 7.5 3000 0.37 plant effluent . . .. 270 84 28.8 0.1 6.5 760 - 105.3% relative to ,the standard method and the effluents gave a mean of 1.3 mg 1-1 with a relative recovery of 102.9%.Sea water gave a recovery of lOOyo of added boron and indicated a boron content itself of 4.7 mg 1-l. This compares with a calculated figure of 4.64 mg 1-1 for a salinity of 34 parts per thousand as quoted by Barnes.ll Spiking recovery from real samples was tested and the results are displayed in Table I (results corrected for colour). There was no evidence of bias in any of the effluents or raw waters tested except for the effluent containing substantial petrochemical industry waste, an effluent that would not normally be analysed for boron in this laboratory. Results of analyses of the samples used in spiking recovery tests are given in Table IT. TABLE I11 SPECIES PRODUCING INTERFERENCE Level producing interference/ Interfering species mg 1-l Interfering species Level producing interference/ mg 1-l Chromium(V1) .. .. 10 Hydrogen carbonate. . .. 1000 Iron(II1) . . .. 25 Carbonate . . .. .. 1000 Nitrite (as N) . . .. 50 Calcium . . .. .. 3 000 Aluminium . . .. 100 Colour .. .. . . According to absorbance at Iron(I1) .. .. 300 410 nm. Adequately corrected for by re-run. TABLE IV SPECIES NOT INDICATING INTERFERENCE Species Cadmium . . .. .. .. .. Lead(I1) . . .. .. .. .. Sulphide . . .. .. .. .. Ammonia (as N) . . .. .. .. lauryl sulphate) . . .. .. .. Chromium(II1) . . .. . . . . Copper(I1) . . .. .. .. .. Fluoride . . .. .. .. .. Manganese . . .. Non-ionic detergent'(lissa& NX)' . . Anionic detergent (Manoxol OT, sodium Highest concentration tested/mg 1-l 100 100 100 1 000 1000 1000 1000 1000 1000 1000 Species Nickel .. .. Nitrate (as N) Orthophosphate (as'P) Silica (SiO,) . . Zinc .. .. Chloride . . .. Magnesium .. Potassium . . .. Sodium . . .. Sulphate . . .. .. .. .. .. .. .. .. .. .. .. Highest concentration tested/mg 1-l 1000 1000 1000 1000 1000 10000 10000 10000 10 000 10000February 1980 OF BORON IN RAW AND WASTE WATERS 145 The method was tested for interferences by specific factors and Table I11 illustrates the level at which interferences became possibly significant for a recovery of boron of 100 & 5% at a 95% confidence level. Table IV shows the highest concentrations tested for those factors that do not produce interference. Interference by sample colour is significant and (for the river waters found within the Forth catchment) linearly related (Fig.7); 100 Hazen units gave a background reading of 0.16 mg 1-1 of boron. Most river waters and effluents however give a background reading below the limit of detection. c 1.0 E n .c 0 0.8 - I en - 5 0.6 8 0.4 L 3 - U - C 3 g 0.2 0.10 0.08 0.06 0.04 0.02 0 loo 200 300 400 500 0 20 40 60 80 100 Sample colour/Hazen units Fig. 7. Correlation between background absorbance at 410 nm and sample colour. A, Best fit for x < 500 Hazen units with correlation coefficient 0.999 for the whole range and 0.930 for x < 70 Hazen units; and B on the expanded diagram, on the right, is best fit for x < 70 Hazen units. x , River water; 0, sewage effluents; and 0, industrial effluents, Sample pH did not cause interference between pH 2 and 13.The performance characteristics of the method were determined for standard solutions, river waters and sewage works effluents and the results are illustrated in Table V. Freshly prepared standards were analysed in triplicate on each of ten days. River water and sewage works effluents were obtained from different localities on each of ten days and analysed in triplicate. Data concerning the analysis of raw waters and effluents used are given in Table VI. Similar analysis of blank solutions indicated a limit of detection capability (defined as 4.65SBJ where S, is the within-batch standard deviation of a single blank determination) TABLE V PERFORMANCE CHARACTERISTICS All figures in mg 1-' of boron. Sample description* r 0.5 mg 1-l 4.0 mg 1-' Clean Dirty Good sewage Poor sewage A 1 Parameter standard standard river water river water effluent effluent deviation .. . . 0.0170 0.0267 0.0148 0.0150 0.015 6 0.0168 Within-batch standard Between-batch standard Total standard deviation . . 0.0220 0.0382 - Mean . . .. .. . . 0.533 4.010 0.101 0.148 0.871 0.631 deviation . . .. . . N.S.? N.S. - - - - - - - * See Table VI. t N.S. = not significant.146 EDWARDS TABLE VI ANALYSIS OF SAMPLES USED FOR PRECISION TESTING Sample type Clean river water- Mean . . .. .. Range .. .. Dirty river water- Mean . . * . .. Range .. .. Good sewage effluent- Mean . . .. .. Range . . .. Poor sewage effluent- Mean.. .. .. Range .. .. BOD/ mg 1-1 1.9 0.8-2.5 12.8 5.4-24.0 20.9* 2.5-120 207 78-435 Suspended solids a t Ammonia mg 1-1 mg 1-1 105 “C/ (as N)/ 9.6 0.30 2-29 0.06-0.95 39.4 2.06 11-82 0.03-5.0 22.9 13.2 6-52 0.70-2 1 .O 223 21.3 32-910 12.6-31.4 Total oxidised nitrogen mg 1-l 3.15 1.3-6.4 (as N)/ 2.27 1.4-3.1 6.47 0.8-23.2 1.56t 0.1-14.0 PH 7.5 6.6-7.9 7.4 6.9-8.4 7.2 6.9-7.3 7.0 5.9-7.4 Electrical conductivity/ pS cm-l 495 100-1 070 603 440-960 639 300-900 856 650-1 020 of 0.10 mg 1-1 and this is also indicated by the analysis of samples.to &O.lO mg 1-1 or &loyo, whichever is the greater (at a 95% confidence level). Results may be defined Conclusion The automated method employing azomethine-H is capable of a limit of detection of 0.1 mg 1-1 with a precision of hO.1 mg 1-1 or &lo%, whichever is the greater. The response is virtually linear up to a level of 4.0 mg 1-1 but decreases to about 96% at 5.0 mg 1-1.No important interferences have been identified except sample colour, for which adequate correction is provided. Spiking recovery from a typical range of raw waters and sanitary and industrial effluents was adequate. The permission of the Forth River Purification Board to publish this paper is gratefully acknowledged. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. References Waggott. A., Wat. Res., 1969, 3, 749. Wilcox, L. V., Agric.,Inf. Bull., 1958, 197, 1. Environmental Protection Agency, “Water Quality Criteria, ’’ Ecological Research Series, EPA R373033, Washington, D.C., 1972, p. 341. Price, D. R. H., and Pearson, M. J.. Wat. Pollut. Control, 1979, 78, 118. Water Pollution Research Laboratory, “Automatic Determination of Boron,” Laboratory Pro- cedure No. 18, Water Pollution Research Laboratory, Stevenage (now Water Research Centre, Stevenage Laboratory), 1969. Bassett, J., and Matthews, P. J., Analyst, 1974, 99, 1. Shanina, T. M., Gel’man, N. E., and Mikhailovskaya, V. S., J. Anal. Chem. USSR (Translated Edition), 1967, 22, 663. “Methods of Chemical Analysis Manual,” 1979-80 Edition, Yorkshire Water Authority, Leeds, 1979. Basson, W. D., Bohmer, R. G., and Stanton, D. A., Analyst, 1969, 94, 1135. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Standard Methods for the Examination of Water and Wastewater,” Thirteenth Edition, American Public Health Association, New York, 1971, pp. 72-73. Barnes, H., “Apparatus and Methods of Oceanography, Part One, Chemical,” Allen and Unwin, London, 1969, p. 313. Received July 9th, 1979 Accepted July 27th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500139

出版商:RSC    

年代:1980

数据来源: RSC