摘要:

722 Analyst June 1983 Vol. 108 fip. 722-727 Analytical Errors Associated with Trace Element Determination in Freshwater Particulate Matter by Atomic-a bsorption Spectroscopy* Renato Baudo Gaetano Galanti and Pier Giorgio Varini Consiglio Nazionale della Ricerche Istituto Italian0 di Idrobiologia Largo Vittorio Tonolli 50-52 28048 Verbania Pallanza Italy Nine analytical errors associated with two procedures for determining the trace element content of freshwater particulate matter by atomic-absorption spectroscopy were investigated using both natural samples and suspensions of candidate reference materials. The first method involves an ultrasound treatment of filters to remove collected particles but in a second method the filters are destroyed by ashing. The precision and accuracy of the two procedures have been determined.Keywords Trace element determination ; particulate matter ; precision and accuracy ; reference materials ; atomic-absorption spectroscopy The determination of particulate bound trace elements in aquatic environments has a fundas mental importance in studies aiming to investigate the mechanism of partitioning in variou-physico-chemical species,l the relative roles of food and water for the element transfer along a trophic ~hain,~-4 or in tracing the pathways of metallic elements throughout the aquatic ecosystems. 5-7 However the precision and accuracy of these determinations have seldom been q ~ a n t i f i e d ~ ~ ~ and many workers have complained of the large error associated with the analytical evaluation of the trace element content of particulate matter.1°-12 In this paper we have tried to evaluate the various analytical errors that build up a cumula-tive imprecision on the determination of iron manganese copper chromium and zinc in fresh-water particulate matter.Experimental Apparatus A Perkin-Elmer 460 atomic-absorption spectrophotometer was used for all chemical analyses using an air - acetylene flame (for iron manganese and zinc) or an HGA-76 graphite furnace (for copper and chromium). A Branson B-15P sonifier with a 0.5-in disruptor was used to remove particulate matter from the filters. Weighings were made using a Mettler H20T or a Sartorius electronic microbalance. Reagents Hydrofluoric and nitric acids were Merck Suprapur grade and acetone was Car10 Erba Reagent grade.The solid standards were the candidate reference materials Platihyfinidium riparioides Olea europea Lagarosifikon major and Freshwater plankton (prepared at the Joint Research Centre of Ispra for the Bureau Communitaire de Rbfbrence within the framework of the Measurements, Standards and Reference Techniques programme of the Commission of the European Com-munitiesls). The liquid standards were prepared by diluting known aliquots of BDH Chemicals Ltd. standard solutions for atomic-absorption spectroscopy with the same acid matrix as used in the preparation of solutions of the samples. Procedure Natural water samples or suspensions of the candidate reference materials were filtered under pressure (1.5 atmospheres of nitrogen) on 47-mm Nucleopore 0.4-pm polycarbonate filters held * Paper presented at Euroanalysis IV Helsinki August 23rd-28th 1981 BAUDO GALANTI AND VARINI 723 in a polycarbonate Nucleopore Liquid Holder.Filters and apparatus were washed previously in 1 M nitric acid carefully rinsed with de-ionised water and dried overnight at 65 "C. The dry masses of the filters were then recorded. During the filtration the sample reservoir (a nylon - glass fibre Amicon reservoir) was con-tinuously agitated. When the filter clogged a vacuum was applied to remove the residual water and the filtrate volume recorded. After drying overnight at 65 O C the filters were re-weighed and transferred into the PTFE vessel of a Perkin-Elmer Autoclave-3 with 35 ml of acetone. The tip of the sonifier disruptor was then immersed in the solution for a 3-min pulsed sonification (a duty cycle of 0.8 s ultra-sonic exposure and a 0.2-s rest to avoid excessive heating) and the filter was removed whilst rinsing with acetone.The solution was evaporated at 55 & 1 "C and the vessel placed in the Autoclave-3 with the addition of 5 ml of a mixture of 70% nitric acid and 30% hydrofluoric acid; the particulate matter was rendered completely soluble by heating the bomb a t 160 & 5 "C for 20 min. Another series of filters was destroyed by ashing overnight at 500 5 30 "C (the polycarbon-ate filter is not soluble in the acid mixture) then the ash was introduced into the PTFE vessel of the bomb for the same treatment in order to render it soluble. For both of the methods (the sonifier method and the ashing method) the final solutions for the atomic-absorption spectroscopic determinations were obtained by carefully washing out the PTFE vessel with a volume of de-ionised water of up to 15 ml.The atomic-absorption spectrophotometer was calibrated using the absorbance values of three different standard solutions each containing the five metals and prepared daily by diluting the stock standard solutions with the acid mixture used to dissolve the samples. Standard concentrations were chosen to cover the whole range of samples to be analysed [up to 2.5 p.p.m. of iron 200 p.p.b. (parts per 109) of manganese? 100 p.p.b. of copper and chromium and 1.5 p.p.m. of zinc] ; the standard S3 had an absorbance slightly higher than the more concentrated sample the standard S2 was half the concentration of S3 and the standard S1 was one third of the concentration of S2.The microprocessor incorporated into the atomic-absorption instrument then used the algorithm K3.A2 + K1.A C = K2.A - 1 to determine the concentration C of a sample from its absorbance A and the absorbances K1-K3 of the standards S1-S3.14 The detection limits calculated as Standard concentration x 2 (standard deviation) average of 20 independent measures of S1 were iron 14 p.p.b. ; manganese 12 p.p.b. copper 1.2 p.p.b. ; chromium 1.9 p.p.b. ; and zinc 16 p.p.b. Results and Discussion The metal (M) concentration in particulate matter is calculated from the measured concentra-tion of the sample solution using the equation ([Mlsample - [M]blank) X solution Volume [Mlperticulate = sample mass (or sample volume) Each atomic-absorption measurement is biased by two instrumental errors the first (A,) refers to the variability of consecutive measurements of the same sample whilst the second (A,) is an estimate of the calibration imprecision introduced by the operators setting up the instrument for the analysis on different days.The quantification of these two errors was obtained by measuring each of three samples (at different concentration levels) ten consecutive times on each of five consecutive days (resulting in 15 estimates of error A, five for each concentration level); because on different days th 724 BAUDO et al. ANALYTICAL ERRORS IN TRACE ELEMENT Autalyst Vol. 108 calibration of the instrument was carried out by different operators we estimated the error A , as the variability of the five mean values of each sample on the various days (one estimate of A for each concentration level).Table I shows the ranges of the A and A errors thus obtained in the concentration range considered; as expected A is usually higher than A, but the imprecision of each measure-ment seems to be small enough to allow the comparison of data collected on different days. TABLE I INSTRUMENTAL ERRORS Results are reproducibilities of the measure (n = 10) during an analytical run (Al) for samples of various concentrations and variability of these mean values on 5 consecutive days after separate calibration of the atomic-absorption instrument (A2). Sample concentration, p.p.b.- Element Minimum Maximum Fe . . 1200 2 100 Mn . . 40 110 c u . . 10 60 Cr . . 10 80 Zn . . 500 1200 Relative standard deviation of absorbance r 1 A A1 A -v 0.00 0.08 4.8 5.3 0.65 0.73 0.74 0.78 3.0 3.2 3.3 3.4 1.6 1.7 1 .o 1.1 0.00 0.07 0.58 0.62 The same exercise could be carried out for the blank values as these two errors also affect this measure but in routine practice the blank concentration used to correct sample data is the average value of more than just one blank. In this instance without denying that A and A , are involved it would seem more realistic to assume that the blank error (B) is due to the variability among the different blank solutions the reproducibility obtained by measuring several blanks within a single batch is reported in Table 11.When acid only is placed in the PTFE bomb the variability of the analytical determination is high as the metal content is low; if unused filters are treated according to the procedures described for the ashing and sonifier methods the variability falls but the metal content of the blank solutions increases noticeably. According to our estimates the amount of metals that filters plus reagents add to the blank solution could be as high as 30% for iron 75% for manganese 90% for chromium, 85% for copper and 20% for zinc of the sample signal (for Lake Maggiore particulate matter). Another error shared by both sample and blank solutions is due to the variability of the measurement of the volume of the final solution; this error (C) does not contribute very much to the cumulative imprecision as we found when we weighed 20 sample solutions after filling a plastic cylinder to the 15-ml mark estimating a relative standard deviation of not higher than 0.99%.In addition to these four errors there is another that affects the sample analysis the error associated with the sample mass (0,) or volume (0,) measurement. The former occurs when the final results refer to the particulate mass the latter if the results refer to the filtrate volume. As the error D cannot be directly evaluated (the particulate net mass is obtained as the difference between gross mass and filter tare and cannot be measured repeatedly) we assumed for this error the value of 16 times the relative standard deviation of the balance weighing (N = 20) of a new filter that is d2 x 17% = O.24y0.TABLE I1 BLANK REPRODUCIBILITY Fe* Mn* cu* Cr* A A c Method k p.p.b. RSD 2 5,p.p.b. RSD 2 k,p.p.b. RSD 3 %:p,p.b. RSD 3 qcid only (n = 20) 50 98 10 89 7 74 3 132 Ashing method (n = l 4 j . . . . 208 13 22 13 78 24 12 22 Sonifier method (n = 18) . . . . 190 19 30 22 60 37 9 8 * = mean; RSD = relative standard deviation June 1983 DETERMINATION IN FRESHWATER PARTICULATE MATTER BY AAS 725 The error D can be estimated by weighing n lakewater samples collected in a 1-1 plastic cylinder marked off in 10-ml divisions filled to the 1000-ml mark; on 20 independent measures we obtained a relative standard deviation of 0.15y0. The errors from A to D combine to give the final imprecision of each sample but the true metal content of the particulate matter suffers from another two errors which refer to dis-solution efficiency (E) and filtration reproducibility ( F ) .Taking n sub-samples from a homogeneous sample of the candidate reference materials and analysing them independently both as dried-only samples (as with the sonifier method) and after ashing them (as with the ashing method) we obtain the results reported in Table 111 A and B. TABLE I11 REPRODUCIBILITY OF SAMPLE DISSOLUTION FOR DIFFERENT MATERIALS AND DIFFERENT SAMPLE SIZES A. Ashing method-Platihypnidium riparioides 2000 mg 8 p.p.m. RSD % 10 mg 8 p.p.m. RSD % 2000 mg 8 p.p.m. RSD % 2000 mg 8 p.p.m. RSD % 10 mg 8 p.p.m. RSD %I Olea europea Lagarosiphon major Freshwater plankton B.Sonifier method-Platihynidium riparioides 200 mg 8 p.p.m. RSD. O/ - ,- Olea europea Lagarosiphon major Freshwater plankton 200 mg R p.p.m. RSD % 200 mg X p.p.m. RSD % 10 mg 8 p.p.m. RSD % Fe Mn 9 480 5 7 080 7 340 1 2 300 3 3 800 9 9 720 3 340 2 2 430 1 3 700 5 3 900 1 3 470 4 57 1 1800 3 250 6 4 170 2 60 3 1880 1 230 16 cu 660 11 640 2 50 1 50 2 40 5 690 1 50 3 50 1 40 19 Cr 550 4 520 9 --40 2 70 7 630 1 --40 1 60 25 Zn 640 1 700 16 20 16 410 2 - -640 1 20 7 360 3 --* R = mean; RSD = relative standard deviation. These errors (which contain all the previous ones) are generally higher for samples with a lower metal content ; moreover the variability is strongly enhanced for the samples of smaller size.The comparison of the two methods demonstrates that the ashing of the samples intro-duces some additional scattering and the mean values are generally lower with this method than for the dried-only samples (with the exception of Freshwater plankton samples). As the mass of particulate matter samples generally is lower than those used in this test it is very likely that the variability of the sample dissolution step introduces an even greater error than those estimated using the reference materials. As stated previously for real samples the filtration reproducibility has to be considered too; this was estimated by filtering sub-samples of candidate reference material suspensions and Lake Maggiore water; from the data shown in Table IV it is evident that the filtration yield (on a mass basis) is more variable when the particles are larger.However the removal of the larger particles by a pre-filtration step significantly reduces the mass of particulate matter collected on the filters and the subsequent analyses have a proportionally higher bias as the sample signal decreases 726 BAUDO et aZ. ANALYTICAL ERRORS IN TRACE ELEMENT AIzaZyst VoZ. 108 TABLE IV FILTRATION REPRODUCIBILITY (ON A MASS BASIS) Sample Meanmasslmg RSD % n Candidate reference material suspensions-Platihypnidium riparioides . . 6.82 19 10 Freshwater plankton . . 10.12 9 6 23 8 11.06 10 8 Lagarosiphon major .. 7.89 Freshwater particulate matter-Lake Maggiore (pre-filtered on a 60-pm screen) . . * . 4.41 30 12 Lake Maggiore (pre-filtered on 10.0-pm glass-fibre pre-filters) . . 1.84 8 8 New filters (0.4 pm) 14.81 7 20 Acid washed and dried filters . . 15.35 11 20 Filters-Moreover the variability of the mass of the filter after drying is higher than for new filters; that means that a part of the filtration variability is not due to the sample heterogeneity but is introduced in calculating the net mass by the allowance made for filter tare. The bias due to this error could be as great as the true filtration error or even greater as indicated by a comparison with the much better reproducibility obtained by calculating the filtration error on the filtrate volume (Table V).TABLE V FILTRATION REPRODUCIBILITY (ON A VOLUME BASIS) Sample Mean filtrate volume/ml RSD % n Candidate reference material suspension-Freshwater particulate matter-Platihypnidium riparioides 490 6 10 Lake Maggiore 1.0-pm filters (pre-filtered on a 60-pm screen) . . 720 7 14 Lake Maggiore 0.4-pm filters (pre-filtered on 10.0-pm glass-fibre Lake Maggiore 0.4-pm filters (pre-filtered on a 60-pm screen) . . 600 6 12 pre-filters) 1380 6 8 The sonifier method is affected by another error namely the efficiency of the sonification treatment in removing the particulate matter from the filter this error was quantified by weighing ten filters before and after sonification. Because at the end of the treatment only a tiny increase of under 1% of the initial filter net mass was observed even assuming that this was all due to particles left on the filter itself about 91% of the 1.39 mg collected has clearly been removed.This is very likely to be an undervaluation of the efficiency of the method because as previously seen after wetting and drying (Table IV) the filter mass is somewhat higher than for new filters. Moreover the microscopical observation of the filters after the sonification treatment indicates that practically all of the particles were removed. In Table VI the cumulative precision for both the ashing method and the sonifier method in estimating the metal content of freshwater particulate water is reported (determined by run-TABLE VI PRECISION (FRESHWATER PARTICULATE MATTER) Parameter Fe Mn Cu Cr Zn Concentrationlpg g-1 .. . . 6200 300 36 136 860 Ashing method RSD % . . 21 11 20 26 17 Sonifier method RSD % . . 10 7 17 33 16 Concentrationlpg 1-1 of filtrate . . 12 3.6 0.4 1.6 10 Sonifier method RSD % . . 11 10 16 19 16 On a mass basis-On a volume basis-Ashing method RSD yo . . 22 16 20 16 1 Jzcne 1983 DETERMINATION IN FRESHWATER PARTICULATE MATTER BY AAS 727 ning 20 sub-samples of Lake Maggiore water through each of the treatment steps of the two procedures) ; with the exception of chromium these precisions range between 10 and 20% both referring the final results to the particulate mass and to the filtrate volume. The comparison of these data with the previously reported errors demonstrates that the cumulative precision is controlled mainly by the efficiency of filtration and sample dissolution, even if the error owing to the use of the blank mean value must be included (especially for samples with low metal content).Some improvements could probably be obtained by using a constant acid mixture - sample ratio instead of a fixed volume (5 ml of acid mixture) for all samples irrespective of their size. As regards the choice between the two analytical methods another parameter has to be taken into consideration that is their respective accuracy. No definitive figures for the candidate reference materials we used are at present available but from data available to us15 we estimate that recovery is satisfactory for both the methods (Table VII) only for chromium does a systematic over-valuation common to the two methods seem likely.TABLE VII INDICATIVE DATA ON THE ACCURACIES OF THE ASHING AND SONIFIER METHODS Method and sample Platihypnidium riparioides (n Lagarosiphon major (n = 6) Freshwater plankton (n = 8) Platihypnidium riparioides (n Freshwater plankton (n = 8) Ashing method-Sonifier method-Recovery % f standard deviation Fe Mn cu Cr Zn I A 7 = 10) 95 f 11 107 f 9 104 f 11 126 f 35 108 f 8 88 f 10 106 f 9 105 f 10 139 f 45 103 f 5 96 f 11 96 f 7 100 f 10 116 f 20 105 f 6 = 10) 88 f 8 105 f 9 104 f 11 138 f 24 97 f 9 . . 104 f 10 102 f 10 103 f 12 144 f 29 101 f 9 As even the precisions of the two procedures are not statistically different we believe that the sonifier method might be preferable because it applies to a wider range of elements includ-ing the volatile ones such as cadmium and lead which could be lost with the ashing method.The final results should by preference be referred to the filtrate volume rather than to the particulate mass for the first instance requires the filter to be handled less and the method is also suitable for field conditions. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. References Bourg A. J . Fr. Hydrol. 1979 10 159. Brooks R. R. and Rumsby H. G. Limnol. Oceanogr. 1965 10 521. George S. G. Pirie B. J. S. and Coombs T. L. J . Ex#. Mar. Biol. EcoE. 1976 23 71. Weigel H. P. Mar. Biol 1977 44 217. Gibbs R. J. Science 1973 180 71. Copin-Montegut C. and Copin-Montegut G. Deep-sea Res. 1978 25 911. Cossa. D. and Poulet S. A. J . Fish. Res. Board Can. 1978 35 338. Baker E. T. and Piper D. Z. Deep-sea Res. 1976 23 181. Duinker J. C. Nolting R. F. and Van Der Sloot H. A. Neth. J . Sea Res. 1979 13 282. Duinker J. C. and Nolting R. F. Neth. J . Sea Res. 1976 10 71. Harris J. E. and Fabris G. J. Mar. Chem. 1979 8 163. Guglielmucci G. Baudo R. Galanti G. and Varini P. G. Mem. 1st. Ital. Idrobiol. 1981 39 243. Muntau H. paper presented at the International Symposium of Use and Production of Reference Perkin-Elmer “Analytical Methods for Atomic Absorption Spectrophotometry,” Perkin-Elmer, Muntau H. personal communication. Materials Berlin November 1979. Norwalk CT 1982. Received December 15th 198 1 Accepted January 14th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800722

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Analyst June 1983 Vol. 108 pp. 728-732 Effects of Temperature Variation on the Zero, Second and Fourth Derivative Ultraviolet Absorption Spectra of Benzenoid Drugs Alexander G. Davidson Department of Pharmacy University of Strathclyde Royal College Building 204 George Street Glasgow, G1 1XW The effect of temperature in the range 0-40 "C on the absorbance at three wavelengths of maximum absorption and two wavelengths of minimum absorption in the zero-order ultraviolet absorption spectra and on the second-and fourth-derivative amplitudes of eight benzenoid drugs displaying fine vibrational structure in the region 250-270nm was studied. All of the drugs show a linear increase in absorbance a t the Amin. between the bands of the fine structure with an increase in temperature with temperature co-efficients ranging from $0.06 to +0.325y0 per degree.A concomitant linear reduction occurs in the second- and fourth-derivative amplitudes with temperature coefficients in the range from -0.60 to -1.12% and -0.7 to - 1.2 yo respectively. The implications of these high temperature coefficients on the accuracy and precision of derivative spectrophotometric assay pro-cedures for benzenoid drugs are discussed. Keywords Benzevaoid drugs ; ultraviolet absorbance ; derivative spectrophoto-metry ; temperature effects Derivative measurement in ultraviolet - visible spectrophotometry has proved to be a powerful analytical procedure capable of improving the resolution of closely related absorption bands of a normal (zero order) absorption spectrum and of reducing or eliminating inter-ference from other absorbing components of the sample.The majority of applications of the technique have been concerned with the improvement of specificity of spectrophotometric procedures and satisfactory methods of assay of many substances in a wide variety of samples have been published.1-6 The ordinate value in a derivative spectrum at any wavelength depends not on the absorbance but on the slope of the fundamental spectrum at that wavelength. The technique therefore discriminates in favour of substances with narrow spectral band widths against those with broad bands. In favourable situations complete elimination of serious inter-ference in a normal absorbance measurement may be achieved. The technique is particularly suitable for the assay of weakly absorbing benzenoid substances that show the fine structure of narrow spectral band widths in the region 250-270 nm.The instrumental parameters that provide optimum resolution sensitivity and specificity in the derivative measurements have been investigated thoroughly.7~8 One experimental variable that has not been examined for its effect on derivative amplitudes is temperature, despite the o b s e r v a t i ~ n ~ ~ ~ ~ over 50 years ago that certain substances show a sharpening of their ultraviolet - visible absorption bands at low temperature and a broadening at high temperature. The effect is particularly evident with those substances that have fine structure in their electronic absorption spectrum owing to its associated vibrational and rotational transitions e.g.benzenoid compounds10J1 and certain dyes.12 The increased resolution at low temperature is attributed to a reduction of the vibrational and rotational energies that cause band broadening and also to a decrease in solute - solvent interaction. The objective of this work was to investigate the effect of temperature on the second- and fourth-derivative amplitudes of benzenoid drugs with the aim of assessing the importance of temperature control in the assay of these drugs in pharmaceutical formulations by derivative ultraviolet spec tropho t ome try DAVIDSON Experimental Drug Substances Ephedrine hydrochloride. Diphenhydramine hydrochloride. Atropine szclphate. BDH Chemicals Ltd. Bethanidine sulphate. Cyclizine hydrochloride.Diphenylpyraline hydrochloride. Hydroxyxine hydrochloride. Pfizer Ltd. Dextropropoxyphene hydrochloride. Dista Products Ltd. All drug substances were gifts from the manufacturers. Sigma London Chemical Co. Ltd. Sigma London Chemical Co. Ltd. Burroughs Welcome and Co. Burroughs Welcome and Co. Smith Kline and French Laboratories Ltd. 729 Spectrophotometer Zero- and second-derivative spectra of the benzenoid drugs were recorded from 290 to 220 nm using a Perkin-Elmer double-beam ultraviolet - visible recording spectrophotometer with a keyboard expansion accessory. The scan speed was 60 nm min-l the spectral slit width 1 nm the response 0.5 s and the recorder range was 1 V. The minimum and maximum ordinate settings were selected to record the maximum absorbance or amplitude at not less than 70% of the recorder full-scale deflection (250 mm) .Fourth-derivative spectra recorded simultaneously with the second-derivative spectra using a second recorder on range 10 V, were generated using a Hitachi derivative accessory operating in the second-derivative mode (mode 5) in series with the spectrophotometer in the second-derivative mode. Temperature Variation The spectra of the drug solutions were recorded in a 1-cm jacketed quartz cell lightly stoppered to prevent evaporation using a conventional 1-cm quartz cell containing only the solvent as the reference. Variation of temperature of the drug solutions was achieved by circulating water at a controlled temperature from a 20-1 water-bath through the jacket of the cell by means of a thermostatically controlled heating - pumping unit.Ice was used to reduce the temperature of the circulating water to 0-0.5 "C and the heating unit was used to increase the temperature in 2-3 "C increments to 40 "C. The temperature of the solution during the spectral measure-ment was recorded as the mean of the temperature of the water-bath read to 0.1 "C and that of the water returning to the bath. The temperature difference did not exceed 0.6 "C at the extremes of the temperature range studied. Condensation on the faces of the sample cell at low temperatures was avoided by flushing the cell compartment with oxygen-free nitrogen. Treatment of the Results The measurements of absorbance in the zero-order spectrum and amplitudes in the second-and fourth-derivative spectra of each solution were made in triplicate at each temperature.The mean values of amplitudes (in millimetres) or absorbance ( y ) and temperature in degrees centigrade (x) were analysed by regression analysis using the method of least squares. If a significant linear correlation was found to exist at the 95%.probability level between the measured y value and temperature the linear regression equation was calculated as y = a + bx. The temperature coefficients as a percentage per degree at 20 "C were calculated from 100b/y where yZ0 is the amplitude or absorbance at 20 "C calculated using the regression equation. Results and Discussion Benzenoid drugs displaying fine structure in their ultraviolet absorption spectra have been classified according to their chemical structure and absorption characteristics.13 The selec-tion of benzenoid drugs in this study was made on the basis of providing representatives of each class as follows.(a) Simple monoalkylbenzene derivatives with saturated a and 18 carbon atoms that hav 730 DAVIDSON EFFECTS OF TEMPERATURE VARIATION ON THE Analyst VoZ. 108 Amax. at 252.5 & 1 258 & 1 and 263 & 1 nm and Amin. at 254 & 1 and 262 & 1 nm: bethanidine ephedrine and dextropropoxyphene. ( b ) Simple monoalkylbenzene derivatives with a carbonyl group showing spectral properties similar to those drugs in group (a) atropine. (c) Biphenylmethyl derivatives having Amax. and Amin similar to those in group ( a ) : diphenhydramine and diphenylpyraline which show a loss of resolution of the maximum at 263 nm.( d ) Biphenylmethyl or substituted benzene derivatives showing a shift in Amax. and Amin, of 5-6 nm to higher wavelengths and differences in the general shape of the spectrum from those drugs in groups (a) cyclizine and hydroxyzine. The pH values of the solutions of the drugs were adjusted if necessary with hydrochloric acid to at least 2 pH units less than the pK of the drug (pKa1 for cyclizine) to ensure that the drugs were substantially in the protonated form. Certain drugs e.g. bethanidine and ephedrine which show only very small differences in the zero-order spectra of their pro-tonated and non-protonated forms display considerably larger differences in their derivative spectra.I4 Effect of Temperature on the Fundamental Spectra The normal zero-order spectrum of bethanidine at 20 "C as an example of a drug showing typical benzenoid absorption bands in the region 250-270 nm is shown in Fig.l ( a ) . The maxima at 251.5 257.0 and 263.0 nm are designated C E and H respectively the minima at 253.5 and 261.5nm are designated D and G and shoulders at 241.5 246.5 259.5 and 266.0 nm are designated A B F and I respectively. The effect of temperature variation on the absorbance of the eight drugs at the three wavelengths of maximum absorption (C E and H) and two wavelengths of minimum absorp-(a) 0.4 -0.3 -0 m f! 0.2 -2 2 0.1 -0 -1 1 I I 220 240 260 280 240 260 280 240 260 280 Wavelengthlnm Fig. 1. (a) Zero-order absorption spectrum ; (b) second-derivative spectrum; and (c) fourth-derivative spectrum of bethanidine sulphate (500 pg d-l) in water.tion (D and G ) is shown in Table I. The only drugs that display a significant change of absorbance at the central Amax. (Amax.2 E) over the 0 4 0 "C range are diphenhydramine and diphenylpyraline which have a small positive dependence and cyclizine which has a small inverse dependence on temperature. The absorbances of most of the drugs at their Amax.1 (C) and Amax.3 (H) show a small positive dependence on temperature the exception being cyclizine which shows no significant change at its Amax. and a reduction in absorbance at its Amax.) with an increase in temperature. More consistently all of the drugs display a larger positive temperature dependence at their Amin.1 (D) and Amin.2 (G) as high as +0.325% per degree at 20 "C for atropine at its Amin-p.These results are consistent with the earlier observation^^^^^ that spectral band widths are increased at higher temperatures resulting in a loss of resolution of closely related bands as a broadening of two adjacent bands will produce a relatively greater increase in absorbance at the trough between them than at th J m e 1983 DERIVATIVE UV ABSORPTION SPECTRA OF BENZENOID DRUGS 731 maximum of either. An interesting example of the improvement in resolution at low temperatures in addition to those shown in Table I is the resolution of a small but clearly defined band a t 266 nm in the spectrum of bethanidine at 0 "C which is seen as a barely discernible shoulder at 40 "C. TABLE I EFFECT OF TEMPERATURE ON THE ABSORBANCE OF BENZENOID DRUGS Bethanidinesulphate .. . . 500 Water 7.00 +25 6.57 +125 8.64 NS? 6.12 +187 6.48 +57 Ephedrine hydrochloride . . . . 500 Water 7.10 +25 6.29 +196 9.11 NSt 6.10 +253 7.01 +32 Dextropropoxyphene hydrochloride . . 500 0.01 M HCI 8.43 +18 7.80 +168 10.71) NSt 7.27 +240 8.20 $28 Atropine sulphate . . 1000 Water 4.42 +46 3.91 +221 5.56 NST 3.45 +325 4.13 $18 Shoulder Diphenhydramine hydrochloride . . 350 Water 13.31 +43 12.99 + 103 15.60 + 61 Diphenylpyraline hydrochloride . . 350 Water 12.36 + 51 12.08 + 101 14.47 + 22 Shoulder Cyclizine hydrochloride . . . . 250 1 M HCI 23.20 NSt 21.40 +80 24.61 -21 14.02 +207 18.20 -90 Hydroxyzine hydrochloride . . . . 250 1 M HCI 16.60 +20 16.20 +61 17.51 NSt 13.49 +94 14.08 NSt * Temperature coefficient as a percentage of the A$!! at 20 "C per degree x 10'.The values are uncorrected for the coefficient of cubic t NS = no significant linear correlation. expansion of water which increases with increasing temperature and which at 20 O C is 0.000 21 per degree." Effect of Temperature on the Second- and Fourth-derivative Spectra The second- and fourth-derivative spectra of bethanidine are shown in Fig. l ( b ) and (c), respectively. The minima at C E and H in the second-derivative spectrum and maxima at C E and H in the fourth-derivative spectrum correspond after correction for scan-speed effects,15s16 with the wavelengths of maximum absorbance (C E and H) in the zero-order spectrum. Additional minima in the second-derivative spectrum and maxima in the fourth-derivative spectrum are generated owing to the improved resolution of the shoulders A B, F and I in the zero-order spectrum.Examination of the second- and fourth-derivative spectra obtained for the eight drugs at various temperatures in the range 0 4 0 "C shows that the amplitudes of all the bands are reduced as the temperature increases. Moreover the temperature dependence varies con-siderably for the different derivative bands of each drug. This is particularly evident with atropine whose second-derivative band measured from the minimum H to its shorter wave-length satellite is the largest band at 0 "C whereas above 15 "C the largest band is that measured from the minimum E to its longer wavelength satellite. The reduction in ampli-tude of all the bands is linearly related to the temperature with correlation coefficients of not less than 0.992.Table I1 records the temperature coefficients as a percentage of the ampli-tude a t 20 "C per degree of the largest amplitude in the second- and fourth-derivative spectra selected because they would be the amplitudes of choice in quantitative derivative spectrophotometric assays of these drugs in the absence of interference by other components of the sample. The small increase in spectral band width and consequent reduction of the slope at wavelengths adjacent to the Amax. values that result in temperature coefficients of absorbance at the Amin. in the range +0.06 to +0.32y0. per degree (Table I) also result in significantly greater dependences of the derivative amplitudes from -0.6 to - 1 .12y0 per degree in the second-derivative spectra and from -0.71 to -1.2% per degree in the fourth-derivative spectra .Temperature coefficients of the magnitude of those in Table I1 are high and have an obvious relevance to the accurate and precise measurement of derivative amplitudes for quantitative purposes. A temperature difference of 5 "C between the sample solution and the external standard solution used for reference would introduce a systematic error of 3.0-5.6% in the second-derivative assay and 3.5-6.0% in the fourth-derivative assay of benzenoid compounds. While adjustment of the temperature of the solutions to a pre-selected value does not appear to be necessary it is however essential that the temperature of the standard and sample solutions at the time of measurement be identical.The practice in this laboratory now is to equilibrate all solutions to room temperature and to record the derivative spectrum without delay after filling the cuvette. If replicate measurements ar 732 DAVIDSON TABLE I1 EFFECT OF TEMPERATURE ON SECOND- AND FOURTH-DERIVATIVE AMPLITUDES OF BENZENOID DRUGS Drug Bethanidine sulphate . . Ephedrine hydrochloride . . Dextropropoxyphene hydrochloride . . Atropine sulphate . . Diphenhydramine hydrochloride . . Diphenylpyraline hydrochloride . Cyclizine hydrochloride . . Hydroxyzine hydrochloride . . Second derivative Amphtude* TCt - E S - 1.12 E S - 0.63 Es -0.70 E L -0.61 I L -0.60 Is -0.99 Hs -0.76 HL -0.74 Fourth derivative Amplitude* TCt ’ HL - 1.16 HL - 1.12 HL - 1.20 HL - 1.10 IS -0.71 I S - 1.03 HS - 0.95 HS - 1.00 * For identification of the bands see Fig.1. t Temperature coefficient as a percentage of the amplitude a t 20 “C per degree. The subscripts S and L indicate measurement to The values the shorter or longer wavelength satellite respectively. are uncorrected for the coefficient of the cubical expansion of water. made for example for evaluation of precision the cell is re-filled with fresh solution to avoid an increase in temperature of the solution in the cell compartment due to the proximity of the light source. The effect of temperature on the derivative amplitudes of other non-benzenoid com-pounds is currently being investigated and will be the subject of another paper.The technical assistance of Mrs. Pauline Constable is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Fell A. F. Proc. Anal. Div. Chem. Soc. 1978 15 260. O’Haver T. C. Clin. Chem. 1979 25 1548. Traveset J. Such V. Gonzalo R. and Gelpi E. J . Pharm. Sci. 1980 69 629. Fell A. F. and Allan J. G. Anal. Proc. 1981 18 291. Fell A. F. Jarvie D. R. and Stewart M. J. Clin. Chem. 1981 27 286. Davidson. A. G. and Elsheik H. Analyst 1982 107 879. Cahill J. E. Am. Lab. 1979 11 79. Cahill J. E. and Padera F. G. Am. Lab. 1980 12 101. Conant J . B. and Crawford F. H. Proc. Nut. Acad. Sci. USA 1930 16 552. Arnold L. B. and Kistiakowsky G. B. J. Am. Chem. Soc. 1932 54 1713. Yarborough V. A. Haskin. J. F. and Lambdin W. J. Anal. Chem. 1954 26 1576. Sheppard S. E. and Brigham H. R. J . Am. Chem. Soc. 1944 66 380. Siek T. J. in Sunshine I. Editor “CRC Handbook of Spectrophotometric Data of Drugs,” CRC Press Boca Raton Florida 1981 p. 83. Davidson A. G. unpublished observations. Talsky G. Mayring L. and Kreuzer H. Angew. Chem. Int. Ed. Engl. 1978 17 785. “Perkin-Elmer Model 552 UV-VIS Spectrophotometer Operator’s Manual,’’ Perkin-Elmer Corpora-Kaye G. W. C. and Laby T. H. “Tables of Physical and Chemical Constants,” 14th Edition, Received November 15th 1982 Accepted January 28th 1983 tion Nonvalk CT 1979. Longmans London 1973 p. 52

ISSN:0003-2654    

DOI:10.1039/AN9830800728

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Analyst June 1983 Vol. 108 pp. 733-737 Fluorimetric Determination of Trace Amounts of Gold as an Ion-association Complex with 2-Phenyl benzo[8,9]quinolizino[4,5,6,7-fed]phenan-thridinylium Perchlorate 733 Tomas Perez- Ruiz Concepcion Sanchez-Pedreiio and Joaquh A. Ortuiio Department of Analytical Chemistry University of Murcia Murcia Spain and Pedro Molina-Buendia Department of Organic Chemistry University of Murcia Murcia Spain The synthesis characterisation and applications of 2-phenylbenzo[8,9]-quinolizino[4,5,6,7-fed]phenanthridinylium perchlorate (PQPP) as a reagent for the formation of ion-association complexes is described. This reagent reacts with AuC1,- to produce a 1 1 complex in 0.5 M hydrochloric acid, which is slightly soluble in water and can be extracted into isoamyl acetate with an extraction efficiency of 90.7o/b.The PQPP shows an intense fluorescence and is used for the fluorimetric determination of trace amounts of gold in the range 0.2-1.75 pg per 5 ml of organic layer. The interferences of many metallic ions have been examined and under appropriate working conditions the method is applicable to the determination of gold in lead. Keywords 2-Phenylbenzo[8,9]quinolizino[4,5,6,7-fed]phenanthridinylium per-chlorate ; spectrojhorimetry ; gold determination ; lead Extraction methods for the determination of gold are numerous. The more useful photo-metric reagents are Brilliant green1 and Crystal violet . 2 9 3 The most common fluorimetric reagents are rhodamine~,~-~ rhodanines,1° kojic acidll and bipyridylglyoxal diphenylhydra-zone.12 In this work we have introduced a new reagent 2-phenylbenzo[8,9]quinolizino[4,5,6,7-fedlphenanthridinylium perchlorate (PQPP) which is simple to synthesise by the photo-chemical cyclisation of lY2,4,6-tetraphenylpyridinium perchlorate and its multi-cyclic structure provides very good fluorescent characteristics.Recently several photochemical cyclisations of polyarylpyridinium salts have been de-scribed Tymyanskii et aZ.13 reported the photocyclisation of 1,2-diarylpyridinium salts in ethanol to yield the dibenzo[c] [1,8]naphthyridinylium ring system; however the 1,2,6-triphenyl and lJ2,4,6-tetraphenyl derivatives in methanol readily undergo double cyclisations to give derivatives of the fused hexacycle benzo [8,9]quinolizino [4,5,6,7-$ed)]phenanthri-dinylium ring system under 300-nm irradiation.l49l5 Q ($&+$,*- / \ \ / \ TPPP PQPP PQPP forms ion-association complexes with a small number of metal halogen anions.These complexes may be used for the spectrofluorimetric determination of the metal after extractio 734 P~~REZ-RUIZ et a,?. FLUORIMETRIC DETERMINATION Analyst Vol. 108 with the appropriate organic solvent. This paper discusses the formation and extraction of the ion-association complex of gold(II1) in a hydrochloric acid medium and a fluorimetric method for the determination of gold in lead. Experimental Apparatus Fluorescence spectra and quantitative spectrofluorimetric measurements were obtained with a Perkin-Elmer Model 3000 spectrofluorimeter equipped with a quantum counter.Excitation spectra were corrected but emission spectra were not. A Perkin-Elmer Elemental Analyzer Model 240 B a Perkin-Elmer 177 grating infrared spectrophotometer a Hewlett Packard 5980 A mass spectrophotometer and a Varian FT-80 nuclear magnetic resonance spectrometer were used for identification of PQPP. Reagents All inorganic chemicals used were of analytical-reagent grade. Acetophenone benzalde-hyde(A1drich) and isoamyl acetate (Carlo Erba) were used as received. Doubly distilled water was used exclusively. Chalcone (1,3-di~kenyZ~rop-2-en-l-one) (I). This was synthesised from acetophenone and benzaldehyde by the procedure of Kohler and Cha.dwell.16 2,4,6-Triphenylpyrilium perchlorate (11). This was synthesised from I with acetophenone and perchloric acid by the procedure of Ba1aban.l' 1,2,4,6-Tetmphenyl~yridinium perchlorate.This was synthesised from I1 and aniline by the procedure of Diethey and Dierich.lS 2-Phenyl benxo [ 8,9] quinolixino [ 4,5,6,7-f ed]p henanthridiny lium Perchlorate. This was syn-thesised by the following procedure. 1,2,4,6-Tetraphenylpyridinium perchlorate (0.5 g) was dissolved in 250 ml of methanol and irradiated using a UV lamp in a vessel equipped with an oxygen bubbler for 3 h. The precipitate was filtered and the filtrate was irradiated again to give second and third crops. The combined solids were recrystallised from methanol to give a yield of 70% with a melting-point of 340 "C. Elemental analysis was as follows calculated, C 72.5 H 3.8 N 2.9%; found C 72.9 H 3.8 N 2.8%.The data obtained from the IR and NMR spectra confirmed the structure. M. This was prepared by dissolving 0.012 g of the reagent in 250 ml of ethanol. The solution is stable for several weeks. This was prepared by dissolving tetrachloroauric(II1) acid in 0.6 M hydrochloric acid and was standardised by the iodimetric method with thio-s~1phate.l~ Working standards were prepared from this solution as required. 2-Phenylbenxo [S 9]quinoZizino [4,5,6,7-f ed]phenanthridinylium Perchlorate solution, Gold(l1l) standard solution 0.01 M. Procedure for Determination of Gold( 111) To a volume of sample solution in a separating funnel containing up to 1.75 pg of gold(III), add 5 ml of 5 M hydrochloric acid and 2 ml of 1 0 - 4 M PQPP solution. Dilute to 50 ml with doubly distilled water and extract the mixture with 5 ml of isoamyl acetate.Shake the funnel vigorously for 2 min allow the phases to separate for 5 min and then transfer the organic layer into a centrifuge tube and centrifuge it to give a water-free organic layer. Activate at 300 nm and read the fluorescence of the complex at 460 nm. Under the recommended conditions the calibration graph is linear over the range 0.2-1.75 pg of gold per 5 ml. Results and Discussion Fluorescence Spectra The excitation and emission spectra of the reagent (in ethanol) and the PQP+AuCl,- (in isoamyl acetate) are shown in Fig. 1. The excitation spectra have maxima at 300 346 and 428 nm and the emission spectra have a maximum at 460 nm. The fluorescence quantum efficiency of PQPP (Aex. = 300 nm) in ethanol was determined by the Gains and Dawson method20 and a value of 0.59 was obtained June 1983 L I 75 >.cn al c .-4- .-50 C 0 c $ 25 / * * A -<.' '1' -OF TRACE AMOUNTS OF GOLD 735 Fig. 1. Excitation and emission spectra of (A and C) gold complex PQP+AuCI,- in iso-amyl acetate and (B and D) reagent in ethanol. Effect of Acidity and Chloride Ion Concentration The effect of the concentration of hydrochloric acid on the formation of the PQP+AuCl,-complex and its extraction into isoamyl acetate was studied using fixed concentrations of gold(II1) (30 p.p.b.) (parts per lo9) and PQPP (3.3 x M) and varying the hydrochloric acid concentration from 0.01 to 1 M. The fluorescence intensity remained constant between 0.1 and 0.5 M in hydrochloric acid.The influence of the chloride ion concentration was studied using 30 p.p.b. of gold(II1) solution in 0.5 N sulphuric acid and sodium chloride solutions ranging from 0.01 to 1 M. Maximum and constant fluorescence intensities were obtained in isoamyl acetate extracts for chloride ion concentrations in the range 0.1-0.5 M. Taking all these results into account a 0.5 M hydrochloric acid medium was selected. Composition of the Complex To establish the composition of the complex the continuous variation21*22 and molar ratio23 methods were applied in 0.5 M hydrochloric acid. The molar ratio of gold to PQPP was found to be 1 1 by these two methods (Figs. 2 and 3). Fig. 2 shows that a molar ratio of [PQPP] to [Au(III)] of higher than 7 is necessary for the complete formation and extraction of the complex.The apparent stability constant of the complex was calculated from the results of the molar ratio and Job's method and an average value of logK = 7.7h0.1 at 20 "C was obtained. 75 z I I I 2 8 14 20 Ratio [PQPI/[Aul Fig. 2. Stoicheiometry of gold - PQPP complex determined by the molar-ratio method; gold(II1) con-centration 1.46 x lo-' M 736 P~~REZ-RUIZ et al. FLUORIMETRIC DETERMINATION Aaalyst Vol. 108 Extraction Efficiency and Stability The selected ratio of aqueous phase to isoamyl acetate was 10 1 (V/V) and the extraction efficiency was 90.7%. The fluorescence intensities of the complex extracted in isoamyl acetate remain constant for at least 30 min. t 0.2 0.5 0.8 Ratio [PQPl/([PQPl+ [Aul) Fig.3. Stoicheiometry of gold - PQPP corn-plex determined by Job's method ; concentration of gold(II1) plus ligand 5 x 10-'M. Calibration Graphs The slope of the calibration graphs increases with increase in gold concentration and under the recommended conditions in the procedure the calibration graph was linear over the con-centration range 6 3 5 p.p.b. The coefficients of variation obtained from ten measurements of 27 and 8 p.p.b. of gold(II1) were rt3.4 and 5.6y0 respectively. TABLE I INTERFERENCE OF OTHER IONS IN THE DETERMINATION OF GOLD Molar ratios at which significant variation in fluorescence was first detected. The concentration of gold(II1) was 16 ng ml-I. Ion added &#) . . As(V) . . Cd(I1) Cu(I1) Cr(II1) . . Ga(II1) .. In(II1) Fe(II1) . . Pb(I1) Mn(I1) Molar ratio [ion added] / [Au (I 11) ] Ion added 5 Hg(I1) 3 Mo(V1) . . . I 3 Pd(I1) . . 15000 Pt(1V) . . $$\;)I) 150 150 300 Zn(I1) . . 150 Nitrate . . . . 0.3 Perchlorate . . 1500 Phosphate . . 150 Sulphate * . . . Molar ratio 0.05 60 3 0.3 1500 0.02 150 6 x lo6 3 x 10' [ion added] / [Au (111) ] 3 x 103 3 x 10 June 1983 OF TRACE AMOUNTS OF GOLD 737 Effect of Other Ions In the determination of 16 p.p.b. of gold(II1) extraneous ions can be tolerated at the levels given in Table I. The limiting value of the concentration for each ion was taken as that value which caused an error of not more than 4% in the fluorescence intensity. Anions were added as chlorides nitrates or sulphates and cations in the form of sodium or potassium salts.The positive interference can be attributed to the fact that those elements also form ion-pair compounds with the reagent in hydrochloric acid and so are slightly extracted into the organic solvent. Thallium mercury platinum and iron cause serious interference. Application The method was checked for its application to the determination of low concentrations of gold in lead by the following procedure. The samples were dissolved in nitric acid and boiled twice nearly to dryness with doubly distilled water to reduce the acidity. The solutions were transferred into 50-ml calibrated flasks and the hydrochloric acid concentration was adjusted to 0.1-0.5 M and diluted to volume with doubly distilled water.To avoid the interference due to the high concentration of lead(II) a calibration graph was prepared using a similar concentration of lead(I1) to that of the sample solution. The results are shown in Table 11. TABLE I1 DETERMINATION OF GOLD IN LEAD Sample 1 Sample 2 Gold content yo* . . . . 0.0006 0.001 8 * Values obtained by atomic-absorption spectrometry. Gold found yo . . . . 0.0005 0.001 7 Conclusion The suitability of the PQP+AuCl,- system for the development of a rapid simple accurate and sensitive fluorimetric method for determining small amounts of gold has been demonstrated. In particular this method could be useful in routine analytical work for the determination of gold without prior separation procedures. 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Fogg A. G. Burgess C. and Thorburn Burns D. Analyst 1970 95 1012. Panchev B. Bulg. Akad. Nauk. Izv. Geol. Inst. Sofia 1965 14 231. Kothny E. L. Analyst 1969 94 198. Podberezskaya N. K. Shiushokova V. A. and Shilenko E. A. Zavod. Lab. 1967 33 152. Marinenko J. and May I. Anal. Chem. 1968 40 1137. Podberezskaya N. V. and Shushkova V. A. Zavod. Lab. 1970 36 1048 and 1197. Blyum I. A. Pavlova N. N. and Kalupina F. P. Zh. Anal. Khim. 1971 26 55. Grigoryan L. A. Mikaelyan D. A. and Tarayan V. M. Arm. Khim. Zh. 1976 29 929. Naganuna T. Shizen Kagaku 1980 29 7 7 . Podberezskaya N. K. Shilenko E. A. and Sheherbov D. P. Zavod. Lab. 1970 36 661.Mureta A. and Ujihara T. Bunseki Kagaku 1961 10 497. Grases F. Garcia-Sanchez F. and Valcarcel M. Anal. Lett. 1979 12 803. Tymyanskii Y . R. Knyazhanskii N. I. Andreichikov Y . P. Trukhan G. E. and Dorofenko, Katritzky A. R. Zakaria Z. Lunt E. Jones P. G. and Kennard O. J . Chem. SOC. Chem. Katritzky A. R. Zakaria Z. and Lunt E. J . Chem. SOC. Perkin Trans. I 1980 1879. Kohler E. P. and Chadwell H. M. Org. Synth. Coll. Vol. I 1941 78. Balaban A. T. C. R. Acad. Sci. Ser. C 1963 256 4239. Dielthey W. and Dierich H. J . Prakt. Chem. 1935 144 1. Beamish F. E. “The Analytical Chemistry of the Noble Metals,” Pergamon Press Oxford 1966. Gains N. and Dawson A. P. Analyst 1979 104 481. Job P. Justus Liebigs Ann. Chem. 1928 9 113. Irving H. and Pierce T. B. J . Chem. Soc. 1959 2565. Yoe J. H. and Jones A. L. Ind. Eng. Chem. Anal. Ed. 1944 16 111. G. N. Zh. Org. Khim. 1976 12 1126. Commun. 1979 268. Received November 1 lth 1982 Accepted December 14th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800733

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

738 Analyst June 1983 VoZ. 108 $y5. 738-741 Preparation of Fatty Acid Methyl Esters from Olive Oil and Other Vegetable Oils Using Aqueous Hydrochloric Acid - Methanol Nikolaos B. Kyriakidis and George Dionysopoulos State Chemical Laboratories Research Department An. Tsocha 16 Ambelokipi Athens Greece Fatty acid methyl ester analysis is used for the examination of olive oil adulteration. Because of the toxicity of boron trifluoride - methanol reagent, which is currently used for the esterification of fatty acids the use of aqueous hydrochloric acid - methanol as an esterification reagent has been studied. The method involves hydrolysis of lipids with a 2% sodium hydroxide in methanol solution followed by esterification with aqueous hydrochloric acid in methanol (3 + 2) for 10 min on a steam-bath.A detailed study of the esterification of olive oil and of other vegetable oils with a high content of unsaturated fatty acids has been undertaken. Comparison of this esterifica-tion with that using boron trifluoride and hydrochloric acid has given excellent agreement of results. Methyl ester hydrolysis has been found to take place during the esterification step. The hydrolysis does not affect the reliability of the method. Keywords Fatty acids ; olive o i l ; methyl esters; gas chromatography The fatty acid composition of olive oil and other vegetable oils is still used by lipid analysts for the detection of adulteration. The mildest reagent used for the preparation of fatty acid methyl esters (FAMe) is anhydrous hydrochloric acid in methano1.l The procedure gives excellent results but needs anhydrous conditions and long reaction times.Concentrated sulphuric acid in methano12s3 transesterifies lipids in the same manner and at the same rate as methanol - hydrochloric acid and it is easily prepared but if the reagent is not used with care, some decomposition of unsaturated fatty acids may occur. Boron trifluoride in methanol* (12-14% m/V) is also used as a transesterification catalyst and in particular as a rapid esterification reagent for free fatty acids. The reagent results in the production of artifacts when it is old or too concentrated.5 In general it is also extremely toxic for the respiratory tract and the mucous membranes. The toxic effects come from the joint action of boron and fluorine.A study has recently been published6 on the determination of fatty acids using a solution of hydrochloric acid in methanol. Owing to the interest of substituting the boron trifluoride - methanol reagent with another non-toxic inexpensive easily prepared reagent it was decided to test the above reagent for the examination of the composition of fatty acids in corn oil sunflower oil etc. Experimental Gas Chromatography A Perkin-Elmer Sigma 2B gas chromatograph equipped with a flame-ionisation detector (FID) and a 6 ft x & in i.d. stainless-steel column packed with 10% EG SS-X on Chromosorb W HP 100-200 mesh was used. The results were processed on a Perkin-Elmer Sigma 10, integrator. The injection temperature was 200 "C the detector temperature was 200 "C and the column temperature was 180 "C.Nitrogen was used as the carrier gas. Thin-layer Chromatography Silica gel 60 without fluorescent indicator and pre-coated high-performance thin-layer chromatographic (HPTLC) plates (E. Merk Darmstadt) were used. Light petroleum (b.p. 60-80 "C) diethyl ether and acetic acid (90 + 10 + 1 V/V)' was used as the developing solvent. The spots were rendered visible with 2,7-dichlorofluorescein. Reagents Analytical-reagent grade chemicals were used throughout. Boron trijuoride 14% in methanol. E . Merk Darmstadt. PerchZoric acid 70%. Ferek Berlin KYRIAKIDIS AND DIONYSOPOULOS 739 Methyl Ester Preparation Hydrolysis and heated for 10 min in a steam-bath under reflux. E sterijcation To the above mixture 10 ml of hydrochloric acid (36% density 1.19 g ~ m - ~ ) -methanol (3 + 2) were added and boiling was continued for 10 min.Turbidity occurred at this stage owing to the formation of insoluble sodium chloride which was quickly settled by precipitation of the crystalline material 10 ml of heptane or light petroleum were added and the boiling was continued for 2 min. After cooling the reaction mixture was diluted with water and the hydrocarbon layer after drying was used for gas-chromatographic (GC) determination of methyl esters. Esterification with Boron Trifluoride - Methanol and Perchloric acid - Methanol Hydrolysis For both reagents the procedure is the same as described under Methyl Ester Preparation. E sterijcation In the boron trifluoride - methanol method 7 ml of a 14% m/V solution of boron trifluoride in methanol were added and refluxed for 3 min.In the perchloric acid - methanol method, 2 ml of perchloric acid were added and refluxed for 5 min. After the esterification 10 ml of heptane or light petroleum were added and the procedure was continued as described under Methyl Ester Preparation. Ester Hydrolysis Qualitative monitoring of the ester formation and hydrolysis was effected by HPTLC. Quantitative studies of hydrolysis during esterification were carried out using TLC (as described under Thin-layer Chromatography) to separate methyl esters from unesterified fatty acids. Esterification times of 5 10 20 and 25 min were used. The ratio of methyl esters to free fatty acids was calculated by GC measurements] after quantitative esterification of free fatty acids with boron trifluoride.Margaric acid methyl ester was used as an internal standard. Hydrolysis was checked further and more accurately using the Iatroscan TH-10 (Iatron Instruments Japan) for quantitative TLC evaluation. The reaction mixture was applied on silica gel rods and developed in the solvent system. The rods were then brought into the Iatroscan and quantitatively evaluated. About 0.2 g of oil was dissolved in 6 ml of 2% m/V methanolic sodium hydroxide solution Results and Discussion The main interest in this study was to find out the possibility of using the hydrochloric acid - methanol esterification technique for the preparation of olive oil methyl esters. The method of Jham et was compared with two established techniques] namely esterification with boron trifluoride and perchloric acid.Two different proportions of hydrochloric acid to methanol 4 + 1 and 1 + 4 were tried. To calculate the range of experimental error a pre-cision study of the different methods was undertaken. The results of the study (Table I) indicated a very good correlation of the aqueous hydrochloric acid - methanol (4 + 1) with the boron trifluoride - methanol reagent. TABLE I COMPARISON OF DIFFERENT ESTERIFICATION METHODS FOR METHYL ESTER SYNTHESIS Catalyst BF HCIOl HCI - MeOH (1 + 4) HCI - MeOH (4 + 1)’ PPPP Standard Standard Standard Standard Fatty acid Mean % deviation % Mean % deviation % Mean % deviation] yo Mean % deviation % 16 0 12.1 0.4 13.4 0.5 13.5 0.2 11.8 0.2 16 1 1.0 1.5 1.1 0.05 0.8 0.03 0.8 0.01 18 0 3.1 1.2 3.6 0.4 2.5 0.01 2.9 0.03 18 1 78.1 0.3 76.2 1.3 77.3 0.2 77.7 0.4 18 2 6.8 0.9 5.7 0.4 5.9 0.04 6.2 0.740 KYRIAKIDIS AND DIONYSOPOULOS PREPARATION OF FATTY Analyst VoZ. 108 TABLE I1 INFLUENCE OF REACTION TIME ON THE PREPARATION OF METHYL ESTERS Esterification reagent was hydrochloric acid - methanol (4 + 1). Composition at a given reaction time % Fatty acid 10 min 15miA 16 0 11.9 12.8 12.9 13.0 16 1 1.0 1.3 0.9 1.0 18 0 2.3 2.6 2.7 2.8 18 1 71.7 74.9 74.9 75.5 18 2 8.4 7.1 7.6 6.8 18 3 4.0 1.2 0.7 0.6 A combined GC and TLC study to evaluate the minimum time required for the completion of the esterification step has shown that at 3 min the esterification reaction was still incomplete but at about 7 min it was complete.A 10-min reaction time for the esterification step was used in the subsequent experiments but it was found that up to 15 min had no detrimental effects on the ester produced (Table 11). The hydrochloric acid - methanol (4 + 1) reagent is TABLE I11 PERCENTAGE COMPOSITION OF MAJOR FATTY ACIDS FROM OLIVE OIL LIPIDS, USING HYDROCHLORIC ACID - METHANOL (3 + 2) AND BORON TRIFLUORIDE - METHANOL AS ESTERIFICATION REAGENTS f Sample 1 * Fatty acid HC1 BF, 16 0 9.9 10.0 16 1 0.8 0.8 18 0 3.2 3.2 18 1 78.8 79.5 18 2 6.1 5.5 Composition % A Sample 2 Sample 3 7- * HC1 BF HC1 BF, 10.4 10.6 11.4 11.4 0.7 0.7 1.1 0.9 3.4 3.3 2.5 2.8 78.9 78.6 72.9 73.5 5.8 5.7 10.8 10.1 - Sample 4 7-HC1 BF, 11.0 12.1 0.8 0.8 2.9 2.8 76.8 75.8 7.5 7.4 concentrated in hydrochloric acid and consequently dangerous during handling.Because of this more dilute hydrochloric acid-methanol (3 + 2) was tested. Results of comparative studies on four samples of olive oil esterified with hydrochloric acid - methanol (3 + 2) and boron trifluoride - methanol are presented in Table 111 showing excellent agreement of results. In order to study the effects of hydrochloric acid on unsaturated fatty acids samples of soybean oil sunflower oil and corn oil were esterified with hydrochloric acid - methanol (3 + 2) for 10 rnin and compared with methyl esters of the same oils made with boron trifluoride. The results are given in Table IV showing a very good agreement in the relative proportions of 18 1 and 18 2 fatty acids (Fig. 1). Owing to the water content of the esterification reagent, TABLE IV PERCENTAGES OF FATTY ACID METHYL ESTERS OF VEGETABLE OILS: COMPARISON OF THE ESTERIFICATION REAGENTS BORON TRIFLUORIDE - METHANOL AND HYDROCHLORIC ACID - METHANOL Composition yo Sunflower oil & Fatty acid BF HC1- MeOH 16 0 6.7 7.1 16 1 0.2 0.1 18 0 4.7 4.1 18 1 20.0 19.4 18 2 68.1 68.9 18 3 0.3 0.4 7 BF3 10.8 0.2 2.0 26.0 59.9 1.1 Corn oil F HC1- MeOH 11.1 0.2 1.7 23.8 60.8 1.4 3 Soybean oil & BF HC1-MeOH 10.2 10.4 0.2 0.2 3.7 3.5 24.9 24.1 53.7 54.0 7.5 7.Jzcrte 1983 ACID METHYL ESTERS FROM VEGETABLE OILS 741 hydrolysis of the methyl esters produced was very probable. Indeed TLC studies have shown that hydrolysis during esterification is taking place.The results of the TLC study are presented in Fig. 2. The hydrolysis was measured quantitatively by two different methods as described. The two methods gave slightly different percentages ranging from 18 to 18.8% for the GC method and from 18.9 to 19.5% for the Iatroscan TLC method with a crude mean value of 18.8%. 0 2 t 2 0 2 Fig. 2. Hydrolysis study olive oil esterification with hydrochloric acid - methanol 3 + 2. (A) Methyl esters and (B) fatty acids. Fig. 1. Sunflower fatty acids esterified (A) with boron trifluoride - methanol and (B) with hydrochloric acid - methanol. Conclusions One of the main objectives of this work was to find a cheap and efficient method for the preparation of olive oil methyl esters. The esterification with boron trifluoride is not recom-mended for everyday routine use because of its high toxicity and there is a growing reluctance among analysts to use it especially in poorly ventilated laboratories.The esterification with hydrochloric acid - methanol is a cheaper and much safer alternative for the preparation of methyl esters. Esterification with hydrochloric acid - methanol and with boron trifluoride - methanol give very similar percentages and it is clear that the hydro-chloric acid does not affect the unsaturated fatty acids under the reaction conditions. Because of the concurrent hydrolysis of methyl esters during the esterification step it is not possible to use this method for quantitative lipid studies. However it is recommended as a very safe routine method of GC analysis for the detection of adulteration in vegetable oils. References 1 . 2. 3. 4. 5. 6. 7 . Stoffel W. Chu F. and Ahrens E. H. Anal. Chem. 1959 31 307, Rogozinski M. J . Gas Chromatogr. 1964 2 136. Sheppard A. J. and Iverson J. L. J . Chromatogr. Sci. 1975 13 448. Morrison W. R. and Smith L. M. J . Lipid Res. 1964 5 600. Long A. K. Bioclaem. J . 1964 90 40. Jham G. N. Telen F. F. R. and Campos L. G. J . Am. Oil Chem. SOC. 1982 59 132. Metcalfe L. D. and Wang C. N. J . Chromatogr. Sci. 1981 19 530. Received November 15th 1982 Accepted January loth 198

ISSN:0003-2654    

DOI:10.1039/AN9830800738

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

742 Analyst JHne 1983 Vol. 108 pp. 742-747 Modified Gas - Liquid Chromatographic Method for Determining Bromide/Total Bromine in Foodstuffs and John A. Roughan Patricia A. Roughan and John P. G. Wilkins Ministry of Agriculture Fisheries and Food Harpenden Laboratory Hatching Green Harpenden Hertford-shire AL5 2BD The widespread use of methyl bromide as a soil fumigant has necessitated the development of convenient and specific analytical methods for determining bromide/total bromine in foodstuffs and soils subsequent to fumigation. The gas-chromatographic method described by Heuser and Scudamore was initially adopted by this laboratory. However for the substrates under investigation i.e. salad crops and soils we found that the method lacked resolution and reproducibility and was hindered by tailing and long retention-time peaks.The modified method for dried ground substrates is described. Mass spectro-metry was used on a non-routine basis to identify the chromatogram peaks. The mean recovery for dried vegetable substrates is 97% for a wide range of bromide levels equivalent to approximately 20-1 000 mg kg-1 on a fresh mass basis. The method can be used to determine bromide down to 0.1 mg kg-l of substrate fresh mass. The method proved suitable as a basis for development. Keywords Gas chromatography - mass spectrometry ; bromide/total bromine determination ; 2-bromoethanol ; foodstufls ; soils The method described by Heuser and Scudamorel for the determination of ionisable bromide in stored products was selected as being particularly suitable as a basis for development for the determination of bromide/total bromine in other foodstuffs and soils.In this method the bromide ion is converted into 2-bromoethanol by reaction with ethylene oxide in acetonitrile -diisopropyl ether under acidic conditions. The 2-bromoethanol thus obtained is determined in processed extracts by gas - liquid chromatography with an electron-capture detector. The procedure has been established by collaborative study as suitable for the specific determination of residues of inorganic bromide in wheat and maize.2 When the method was applied to salad crops and soils in this laboratory the determinations were hampered by tailing and the presence of many interfering peaks in the gas chromatograms of the final extracts (also observed by Stijve3).Some of the interfering peaks were found to be associated with the use of 4% V/V ethylene oxide in diisopropyl ether. When the ethylene oxide was prepared in acetonitrile the resolution of the 2-bromoethanol was improved without any loss of response. Tailing was further reduced when Carbowax 20M TPA was used to prepare the gas-chromato-graphic column. Examination of the 4% ethylene oxide solution in acetonitrile by gas chromatography revealed a large peak of ethane-l,2-diol (see Identipcation of gas-chroma-tographic peaks). The magnitude of this peak was reduced by bubbling ethylene oxide into cold acetonitrile rather than decanting from the cylinder. Greve and Grevenstuk4 eliminated interferences by ashing lettuce for 15 min at 500 "C.The ash was mixed with 0.6 N sulphuric acid after which the ethylene oxide treatment was carried out as described in the original procedure. The method of Greve and Grevenstuk4 has since been used in interlaboratory studies5 We observed that this pre-treatment gave some improvement but often the substrate was only carbonised and not completely ashed. It was shown that it was necessary to ash for 2 h at 500 "C (600 "C for oily substances) to achieve the light grey powdery ash that we associated with recoveries in excess of 90%. The method at this stage did not give us the desired level of reproducibility; however we found that by drying the substrate and grinding prior to digestion the problem was overcome. A few peaks with long retention times remained which have been shown by mass spectro-metry to be derivatives of ethylene oxide and acetonitrile.The method has been used success-fully in studies to monitor a wide range of foodstuffs for natural bromide and also bromide derived from methyl bromide used as a pre-plant soil fumigant (to be published) ROUGHAN ROUGHAN AND WILKINS 743 Experimental Caution-In view of possible allergic reactions to finely divided proteinaceous material and the highly toxic nature of acetonitrile and ethylene oxide all work where practicable should be carried out in a fume cupboard. Apparatus Perkin-Elmer with an electron-capture detector (g3Ni foil) ; a 2 m 3 mm i.d. glass column packed with 10% Carbowax 20M TPA on Chromosorb W-HP, 80-100 mesh; oven temperature 130 "C; injector temperature 160 "C; detector temperature 320 "C; carrier gas nitrogen 60 ml min-l through the column and detector.Trilab Computing Integrator (Trivector) optional. Varian 1400 gas chromatograph interfaced to a VG Micromass 12B spectrometer (gas chromato-graph - mass spectrometer GC - MS). To record low resolution electron impact (30 eV) spectra, with an accelerating potential of 3 kV; source temperature 180 "C; glass column 0.9 m 2 mm i.d. packed as for the Sigma 4; helium carrier gas. Sigma 4 gas chromatograph. Drying oven. With forced ventilation for sample drying. Centrifigal mill. Oven. Set at 110 "C for overnight storage of digested samples. Mz@?e furnace. Thermostatically controlled hot-plate. Pre-heated to 500 "C (or 600 "C for oily substrates such as tomatoes).Set at 110 "C. Reagents All reagents should be of analytical-reagent grade and tested for interfering impurities. A cetonitrile. Absolute ethanol. Ammonium sulphate crystalline. Sodium sulphate anhydrous crystalline. Sodium hydroxide solution 0.2 N. Ethylene oxide. Cambrian Chemicals lecture bottle. Prepare a 4% solution of ethylene oxide in acetonitrile by bubbling ethylene oxide gas into 96 ml of cold acetonitrile until 100 ml of solution are obtained (decantation introduces large amounts of interfering substances). Sulphuric acid 0.6 N. 2-Bromoethanol stock solution. Dilute 0.1560 g to 100 ml with acetonitrile equivalent to a bromide concentration of 1000 pg ml-l. Gas chromatographic standard solutions are pre-pared by diluting the stock solution to give 0.2 0.4 0.8 1.2 1.6 and 2.0 pg ml-l.Potassium bromide stock solutions. Dissolve 2.978 5 g of oven-dried potassium bromide in distilled water and dilute to 100 ml equivalent to a bromide concentration of 20000 pg ml-l. Dilute to give solutions containing 2000 800 500 400 and 50 pg ml-l. Sample Preparation Weigh each sample before and after drying in order to determine the moisture content (necessary if results of analysis are reported on a fresh mass basis). Prepare the crop e.g. by cutting tomatoes in half and placing skin side down in a porcelain dish or spreading lettuce leaves out on a wire tray. Grind the dried vegetable samples, transfer the resultant powder into a self-seal polyethylene bag for storage in a cool place until required for analysis.Grind dried soil samples using a ball mill and sieve through a 100-mesh sieve; discard the stones. Dry all samples at 110 "C. Store as for vegetable samples. Analysis Add 5 ml of sodium hydroxide solution and stir the mixture thoroughly with a glass rod. Add 10 ml of absolute ethanol gradually whilst continuing to stir using the final portion to rinse the glass rod. Place the crucible on the hot-plate and leave to evaporate to dryness then transfer into the pre-heated oven and leave overnight. Cover the crucible with a silica lid unless the sample is tomato or some other oily substrate, when the lid is omitted place in the pre-heated muffle furnace (see Apparatus; caution-some Weigh a 1-g sub-sample of dried ground material into a silica crucible 74-4 ROUGHAN et al.MODIFIED GLC METHOD FOR DETERMINING Analyst VoZ. 108 substrates ignite spontaneously a t this stage) for 2 h withdraw it and allow to cool. Add 10 ml of sulphuric acid swirl gently and allow the effervescence to subside (soils require thorough mixing with a glass rod). Transfer the sulphuric acid solution/slurry quantitatively into a 100-ml conical flask rinsing the crucible with 4 x 10 ml of acetonitrile and adding the washings to the conical flask. Add 10 ml of ethylene oxide solution to the conical flask, securely stopper shake vigorously for 1 min and allow to stand for 15 min. Decant the acetonitrile solution into a 25-ml cylinder (containing ammonium sulphate levelled to the 2-ml mark) to the 10-ml mark and shake for 15 s. Decant the upper acetonitrile layer into a 10-ml cylinder (containing sodium sulphate to the 1-ml mark) to the 5 ml-mark stopper shake for 15 s and allow to stand overnight (Le.12-24 h). It has been proposed by Stijves and con-firmed by us that the bromide - ethylene oxide reaction is complete after 15 min. We also have determined that there is no loss of bromoethanol on standing for a period of up to 24 h. This period of drying has been found desirable to preserve the performance of the gas-chroma-tographic column though it probably results in increased amounts of extraneous ethylene oxide reaction products (see Table I). TABLE I SUMMARY OF GC - MS DATA Approximate relative gas-chromato-graphic peak area Relative Mass spectrum (most abundant Electron-capture Compound retention time ions in order of intensity)/a.m.u.Total ion current detector response Acetronitrile (solvent) . . . . . . - 41 (M+) 40 42 - -2-Chloroethanol 0.35 31 43 51 80 (M+) 10 200 2-Bromoethanol. . . . . . . . 0.60 31 45 (95 97) (124,126) (M+) 2 200 Ethane-l,a-diol . . . . . . . . 1.0 31 33 32 43 45 62 (M+) 1000 1000 Acetamide . . . . . . . . 1.8 59 (M+) 44 43 42 32 2 < 10 Unknown . . 3.2 113 (M+?) 66 57 70 10 200 2-Iodoethanol . . . . . . . . 1.1 45 172 (M+) 31 <1 50 2,2 -Dihydroxydkhyl eiher . . 4.4 45 75 76 44 107 (M + 1+) 10 40 Gas Chromatography The 2-bromoethanol derived from the reaction between the bromide anion and ethylene oxide is determined by electron-capture gas chromatography interfaced to the computing integrator. Calibrate with 5-pl injections of standard solutions of 2-bromoethanol containing from 1 to 10 ng of bromine using peak height as the measured parameter.The method is equally suitable for conventional measurements from a chart recorder. Sample solutions should be diluted if necessary to bring them within the calibration range using 5-pl injections. It is our experience that the resolution of the 2-bromoethanol peak is good even in the presence of large amounts of extraneous ethylene oxide reaction products. A typical chromatogram is shown in Fig. 1; the magnitude of the peaks is variable depending on the substrate and the characteristics of each individual column. Identi$cation of gas-chromatographic peaks To produce a solution with a sufficiently high concentration of the compounds of interest for GC - MS study (about 10 pg ml-l) a batch of solutions from the routine analysis of samples was pooled and concentrated by evaporation.Comparison of the gas-chromatographic results from this solution with those obtained prior to the evaporation stage showed that none of the compounds of interest had been lost and relative concentrations were approximately the same. 2-Bromoethanol was shown to be present in this solution by its GC - MS behaviour using reference material for comparison. In the mass spectrum of 2-bromoethanol the molecular ion was observed at m/z 124 and exhibited the characteristic intensity distribution of a species containing a bromine atom with M+ and (M + 2)-+ being of similar intensity. Other bromine doublets observed were due to (M-OH)+ (M-CHO)+ (M-CH,O)+ and HBr+.The most intense ions occurred at m/z 31 CH,OH+ and m/z 45 CH,CH,OH+. Six other gas-chromatographic peaks (apart from bromoethanol and the solvent acetonitrile) were observed the identities of five of these were deduced and verified by comparison with reference material (see Table I). Both the ethane-l,2-diol and the 2,2’-dihydroxydiethyl ether are produced by acid-catalysed hydrolysis of ethylene oxide the former being by far the most abundant by-product J m e 1983 BROMIDE/TOTAL BROMINE IN FOODSTUFFS AND SOILS 745 > E . v) 0 v) ?? L c 0 0, 4- 0" G Fig. 1. Typical chromatogram con-taining A acetonitrile ; B 2-chloro-ethanol; C 4 ng of 2-bromoethanol; D ethane-1,Z-diol ; E 2-iodoethanol ; F acetamide; G unknown; and H, 2,2'-dihydroxydiethyl ether.0 4 8 12 16 20 Time/m in The chloro- and iodoethanols are generated from chlorine and iodine in the samples in a similar fashion to the 2-bromoethanol. Their presence in the gas chromatogram suggests that, with a little development this procedure could be used to determine chlorine bromine and iodine (and possibly fluorine) simultaneously. The acetamide is formed from the acid-catalysed hydrolysis of the solvent acetonitrile. The only unidentified component of the solution was that with the characteristic m/z 113 ion in its mass spectrum. Assuming that this is the molecular ion it is probably a nitrogen com-pound and is possibly an impurity from the reagents used. Results and Discussion Verification of Analytical Method Recovery experiments were carried out by spiking 1-g sub-samples of various dry substrates with pipetted 1-ml volumes of standard potassium bromide solution at levels within the ranges of residues currently being determined.Table I1 indicates the bromide background levels of TABLE I1 BACKGROUND LEVEL OF BROMIDE IN DRIED SUBSTRATES USED FOR RECOVERY WORK Values are for micrograms of bromide per gram of dry mass. Replicate determinations of background bromidelpg g-1 Standard r L Mean/ deviation/ Coefficient of Standard Substrate 2 3 4 5 6 7 wg g-' wg g-' variance yo error/vg g-' Soil 10.8 11.2 11.6 11.5 11.9 11.5 11.4 11.4 0.34 2.98 f0.13 Lettuce 45.1 47.0 46.3 45.8 46.0 44.2 45.1 45.6 0.92 2.02 i0.36 Cucumber 17.6 17.0 17.8 16.3 17.7 17.3 18.4 17.4 0.67 3.85 fO.25 Tomato 15.8 15.7 15.2 14.5 13.8 14.1 14.6 14.8 0.77 5.20 -+0.2 746 ROUGHAN et at!.MODIFIED GLC METHOD FOR DETERMINING Analyst VOl. 108 TABLE I11 RECOVERIES OF BROMIDE FROM SPIKED l-g SUB-SAMPLES OF VARIOUS DRIED SUBSTRATES Br- Statistical treatment of mean recovery determined r A 3 Br- added corrected for Mean Standard Coefficient as KBrl background1 Recovery recovery deviation of variance Standard Substrate Soil. . Soil Lettuce . . Lettuce . . Cucumber . . Cucumber . . Tomato Tomato PLg g-' 50 50 50 50 50 50 50 . . 500 500 500 500 500 500 500 . . 2000 2 000 2 000 2 000 2 000 2 000 2 000 . . 20000 20000 20 000 20 000 20 000 20 000 20 000 800 800 800 800 800 800 800 .. 2000 2 000 2 000 2 000 2 000 2 000 2 000 400 400 400 400 400 400 400 800 800 800 800 800 800 800 Pg g-l 40.9 39.2 40.2 41.3 40.7 40.8 41.2 464 460 463 470 481 483 463 1933 1936 1941 1965 1948 1962 1952 19888 19 897 19696 19711 19 527 19 271 19829 782 773 778 765 789 797 782 1942 1948 1916 1978 1952 1881 1896 391 397 379 379 375 382 394 777 790 759 790 770 758 764 % 81.4 81.6 82.4 96.2 96.6 92.6 97.4 98.1 97.6 97.6 96.4 99.1 98.6 99.6 97.8 97.1 97.6 94.1 94.8 97.8 93.8 95.5 98.5 96.3 94.8 95.5 % % % error % 81.2 1.44 1.77 f 0.66 93.8 97.4 98.4 97.6 96.5 96.4 96.6 1.86 1.98 A0.70 0.62 0.64 f0.23 1.11 1.13 f0.42 1.30 1.33 f0.49 1.70 1.76 f0.64 2.14 2.22 f0.81 1.70 1.76 f0.6 Jww 1983 BROMIDE/TOTAL BROMINE IN FOODSTUFFS AND SOILS 747 the various substrates prior to analysis calculated on dry mass.Recovery data for soil, lettuce cucumber and tomato are presented in Table 111 on dry mass. A check sample study was initiated between five laboratories carrying out bromide residue determinations (or having recently carried them out). A bulk lettuce sample was prepared at the Harpenden Laboratory from dried ground lettuce by homogenising 2-g amounts of 121 of the samples obtained for a survey. After re-drying the sample was sub-divided for distribu-tion to participating laboratories. Subsequently replicate determinations of bromide were carried out using the individual laboratories own method on the composite lettuce sample.A compilation of results is presented (Table IV). The mean of the means from Table IV [5316 mg of bromide ion per kilogram of lettuce (dry mass)] compares well with the mean [5280 mg of bromide ion per kilogram of lettuce (dry mass)] calculated from our individual residue results for the original 121 samples. TABLE IV RESULTS OF CHECK SAMPLE ANALYSIS BY FIVE LABORATORIES Laboratory and method 1 2 3 4 Modification method Titrimetric method Titrimetric method Bromide specific Harpenden Parameter in reference 1 in reference 7 in reference 7 electrode* 2-BrEtOH Bromide in lettuce dry masslmg kg-1 5 330 5 630 5130 5670 5 390 5 740 5 700 Meanlmg kg-l 5513 Standard deviatkh/mg'kg-l 230.99 Coefficient of variance yo .. . . 4.19 Standard errorlmg kg-1 . . . . -+87.50 * Individual results not supplied. t Assuming f i = 7. 5 313 - 5407 - 5 371 5672 5 212 - 6 264 5 701 5237 - 5407 5 691 5 262 - 5 228 5 690 5 351 - 5228 5675 5 275 5 713 5 326 - 5228 5667 5687 5 282 4 795 5 305 16.66 50.05 153.40 86.25 0.29 0.95 3.20 1.63 k6.31 f18.96 *58.117 f 32.67 Sensitivity strate fresh mass. The method as described will allow determination of bromide down to 0.1 mg kg-l of sub-Conclusion The gas-chromatographic method described by Heuser and Scudamorel for the determina-tion of bromide as 2-bromoethanol has been modified so that determination may be made on a wide range of substrates.When the original method was applied to fruit vegetables and soils it showed a number of associated problems i.e. severe tailing lack of resolution and poor recoveries; these have been overcome. The mean recovery for a selection of salad crops is 97% for a wide range of bromide levels equivalent to approximately 20-1 000 mg kg 1-1 on a fresh mass basis. The method can be used to determine bromide down to a 0.1 mg kg 1-1 of substrate fresh mass. A few long retention peaks remain; these however have been shown by mass spectrometry to be derivatives of ethylene oxide. 1 . 2. 3. 4. 5. 6 . 7 . References Heuser S. G. and Scudamore K. A. Pestic. Sci. 1970 1 244. Report by the Panel on Fumigant Residues in Grain of the Committee for Analytical Methods for Residues of Pesticides and Veterinary Products in Foodstuffs of the Ministry of Agriculture, Fisheries and Food Analyst 1976 101 386. Stijve T. Dtsch. Lebensm. Rundsch. 1977 73 321. Greve P. A. and Grevenstuk W. B. F. Meded. Fac. Landbouwwet. Rijksuniv. 1976 41 1371. Greve P. A. and Grevenstuk W. B. F. J . Assoc. 08. Anal. Chern. 1979 62 1155. Stijve T. Dtsch. Lebensm. Rundsch. 1981 77 99. Turner A. J . Sci.Food Agric. 1964 15 265. Received February 26th 1982 Accepted January 12th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800742

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

748 Analyst June 1983 Vol. 108 $9. 748-753 Inter-laboratory Calibration for Pesticide Analysis in South Africa Louis P. van Dyk Laurraine Lotter Pieter R. de Beer Andre de Klerk, Awie J. Viljoen and Susan M. Prinsloo Task Group ICE Working Group on Pesticide Analysis Private Bag X134 Pretoria South Africa 0001 The organising and running of four inter-laboratory calibration exercises in South Africa in 1981 are described. Analyses of a solution containing pesti-cides margarine and fruit pulp fortified with pesticides and a potato sample with incurred residues were carried out. The results indicate that most labora-tories produced acceptable results. This type of exercise should be continued. Keywords Inter-laboratory calibration ; pesticide analysis The South African Working Group on Pesticide Analysis (WGPA) is a technical sub-committee of the Inter-departmental Advisory Committee Safeguarding Man against Agricultural Poisons (INDAC).l One of the functions of the WGPA is to run annual inter-laboratory calibration exercises (ICEs).Previous attempts by other organisations in South Africa to undertake ICEs produced mixed results. One of the factors contributing to the disappointing results was lack of complete co-operation from the various participating laboratories. In an attempt to overcome this problem a Task Group was formed to direct the inter-laboratory exercises. The way this task group organised ICEs and the results of those completed in 1981 are reported. Procedure Organisation Members of the Task Group were drawn from central and local government laboratories the South African Bureau of Standards and an agrochemical company laboratory.The Task Group met regularly to plan strategy and discuss the progress of each exercise. There is a relatively small number of pesticide analytical laboratories in South Africa and of these most specialise in a particular field. One of the first problems encountered was to plan the ICEs so that the largest possible number of laboratories could participate. None of the pesticide analysts could be compelled to participate as their laboratories are not centrally funded. Motivation of analysts and continuous efforts to stress the value of ICEs were therefore two of the main tasks of the Task Group. The argument used was that no analyst or laboratory is perfect and that continuous participation in ICEs is the only sure way of proving analytical integrity.Initial efforts were directed at supervisors of pesticide laboratories in the hope that a motivated leader would ensure the participation of his laboratory in ICEs. A preliminary letter was sent to all pesticide analysis laboratories explaining the value and importance of regular ICEs. These exer-cises were planned so that they progressed from a single analysis of pesticides in an organic solvent to more complex samples. The exercises were designed to enable each laboratory to take part in at least one of the four planned ICEs. For the laboratories that participated see the Appendix. The laboratory supervisor was requested to indicate in which of the ICEs his laboratory could participate.Fifteen laboratories out of a total of 24 indicated that they would be prepared to participate in one or more of the ICEs. Of these four laboratories entered two participants for some of the exercises. ICE 1/81 was to be carried out by 19 analysts; ICE 2/81 by 8; ICE 3/81 by 10; and ICE 4/81 by 6 analysts. A second letter, reiterating the importance of ICEs was included in the information mailed to each partici-pating analyst. A protocol for the analytical procedure and reporting of the results was supplied to each participant with a standard form on which instrument operating parameters, methodology calculations and final results had to be submitted. To determine whether participants have carried out the exercise in an acceptable manner, data may be presented graphically2 or in the traditional numerical ~ a y .~ - ~ These tests, The first programme for 1981 was also supplied VAN DYK LOTTER DE BEER DE KLERK VILJOEN AND PRINSLOO 749 however basically examine the normality of the distribution of results. Results falling out-side a satisfactory distribution for a particular standard deviation and confidence limit are usually discarded as unsatisfactory. Results are unsatisfactory in this sense and only as a consequence from the statistical testing point of view if participants have not carried out an exercise in an acceptable manner. All of the results received were first reviewed for outliers using the ASTM procedure5 to a 5% significance level before analysing for mean standard deviation and relative standard devia-tion.The results were then assessed according to the method of Greenberg et aL4 (1) results falling between the mean and & 1 standard deviation are acceptable; (2) results falling between &l and &-2 standard deviations from the mean are acceptable but questionable; and (3) results outside the limits of &2 standard deviations from the mean are unacceptable. This differs from IS0 test 572EiJ6 in which results outside the 95% confidence limit critical values for the Dixon and Cochran tests for the width of the distribution are “stragglers” and those outside the 99% confidence limit are outliers. It was decided to permit up to two analysts per laboratory to participate in a specific ICE provided that the supervisor made certain that these analysts worked independently.Each analyst was allocated a code number known only to the participating laboratory and the Task Group. Each analyst was requested to submit a single value. The preparation of ICE samples had previously presented some difficulties and it was decided to appoint one laboratory for this purpose to ensure consistent expertise in this field. Analyti-cal standards would also be made available from this central laboratory. Programme Four ICES were undertaken in 1981 and these proceeded as follows. ICE 1/81 Parathion and diazinon were dissolved in hexane at concentrations of 10.20 and 5.09 mg 1-1, respectively. Both pesticides can be detected by most of the available detectors and the concentration is also high enough for flame-ionisation detectors to be used.Dilution of the sample could be made by the analyst if necessary. The identity of diazinon was given to the analyst and the other pesticide parathion had to be identified. A number of laboratories only analyse known compounds and do not have the ability to determine unknown compounds. To give these laboratories the opportunity to participate in this exercise they were given the choice of determining the concentration of diazinon only or to also quantitate the unknown using a standard obtained from the central laboratory reporting the result as milligrams per litre of unknown. ICE 2/81 A margarine sample was analysed to establish that it was free from pesticide residues. All of the water was removed from the sample the margarine melted and a volume of hexane, containing dieldrin and gamma-BHC was added.The hexane was removed under vacuum. The margarine was fortified to levels of 0.54 mg k g l of gamma-BHC and 1.049 mg kg-l of dieldrin. Participants were given the information that the sample contained dieldrin and were asked to determine the concentration. The gamma-BHC had to be identified and the concentration determined. The multi-residue method for fatty foods of the WGPAl was recommended; however each laboratory was free to use a different method which then had to be reported. ICE 3/81 Homogenised apple pulp was analysed to establish that it was free from pesticide residues. The pulp was frozen in 600-g masses and each laboratory received one of these and an ampoule containing about 5 ml of a hexane solution of diazinon and dimethoate.The identify of di-methoate was not divulged. The sample was analysed by taking 100 g of de-frosted apple pulp and adding to it 1 rnl of the hexane solution. It was then homogenised and left for 30 min. It was recommended that the sample be extracted with 100 ml of a 30% ethyl acetate in hexane mixture. The recovery of the compounds from the pulp had to be used to correct th 750 Analyst Vol. 108 final result. The concentration range of the pesticides in the pulp was also given as between 0.5 and 5 mg k g l for diazinon and between 5 and 15 mg kg-l for the unknown (dimethoate). The concentrations were calculated and pre-determined to be 10.0 mg kg-1 dimethoate and 1.5 mg k g l diazinon. The concentrations were relatively high to facilitate the use of less sensitive flame detectors.VAN DYK et al. CALIBRATION FOR PESTICIDE ICE 4/81 Potatoes growing under normal field conditions were treated with a granular aldicarb formulation. When the results indicated that the concentration in the tubers had dropped to about 1 mg k g l , expressed as aldicarb sulphone a large sample of 50 kg was collected taken to the laboratory and washed. The potatoes including peel were chopped and then ground on a food mincer. Sub-samples of 500 g each were frozen and supplied to each laboratory. The confidential method of analysis was supplied by Union Carbide SA (Pty) Ltd. Samples were taken at monthly intervals after treatment and analysed. Results Results were not received from all of the laboratories so that the total number of analyses reported are sometimes less than expected.ICE 1/81 Two of the laboratories could not undertake the analysis owing to lack of staff and even after repeated enquiries no comment was received from the other laboratories. TABLE I RESULTS OF ICE 1/81 Results of this exercise are given in Table I. Only 13 analysts supplied results. Pesticide concentration/ mg I-' c E F1 F2 G1 G2 H1 H2 J * . N Laboratory code 01 0 2 P . . . . . . . . Mean Standard'deviatkn . . Relative standard deviation, Truevalue ' Parathion Diazinon' (unknown) (known) . . 13.82 8.43 . . 11.64 13.16t % 10:90 10.29 10.68 10.70 9.42 10.29 10.30 Not detected 12.53 12.01 5.89 11.40 1.25 10.96 10.20 5107 5.11 5.11 5.10 4.04 5.39 5.10 5.07 5.62 5.41 4.90 5.36 1.04 19.40 5.09 Detector used Electron capture Electron capture Flame photometric Flame photometric Flame photometric Flame photometric Thermionic Flame photometric Flame photometric Flame ionisation Flame photometric Flame photometric Electron capture Column used OV-17 + OV-210 SE-30 Dc-200 DC-200 SE-30 + SP-2401 SE-30 + SP-2401 OV-17 ov-210 ov-101 ov-1 SE-30 + QF-1 SE-30 + QF-1 OV-17 + QF-1 Method of quantitation Integrator Peak height Peak height Peak height Peak height Peak height Integrator Integrator Integrator Peak height Integrator Peak height Peak height Dilution 1 + 1 1 + 99 1 + 9 1 + 9 1 + 9 1 + 9 1 + 9 1 + 9 None None 1 + 4 1 + 4 None Standard used* A Unknown A A A A B B A A A A A * A Supplied by the Department of Agriculture.f Outlier not included in calculation. B Supplied by the US Environmental Protection Agency. ICE 2/81 Results of this exercise are given in Table 11. Six analysts supplied results whereas origin-ally eight indicated that they would participate. The two non-participating analysts were from one laboratory and they indicated that they had experienced problems in analysing the sample. ICE 3/81 results. All laboratories used the recommended extraction procedure. ICE 4/81 supplied results. Results of this exercise are given in Table 111.Nine of the ten original participants supplied Results of this exercise are presented in Table IV. Five of the six original participant J m e 1983 ANALYSIS I N SOUTH AFRICA 751 TABLE I1 RESULTS OF ICE 2/81 Pesticide concentration/ Gamma-BCH Dieldrin mg kg-l w Method of Method Standard Laboratory code (unknown) (known) Detector used Column used quantitation used used* c . . . . . . . . . . 1.10 0.43 Electron capture OV-101 + OV-210 Peak height WGPA A F1 . . . . . . . . 0.93 0.68 Electron capture OV-17 + QF-1 Peak height WGPA A F2 . . . . . . . . 0.76 0.69 Electron capture OV-17 + QF-1 Peak height WGPA A G1 . . . . . . . . 1.20 0.70 Electron capture OV-17 + QF-1 Peak height WGPA A G2 . . . . . . . . 1.10 0.60 Electron capture OV-17 + QF-1 Peak height WGPA A 0 0.79 0.65 Unknown Unknown Unknown Unknown Unknown Arithmetic*mean ' 0.98 0.63 Standard deviation .. . . 0.18 0.10 Relative standard deviation % 18.4 15.9 True value . . . . . . 1.049 0.541 * A Standard supplied by the Department of Agriculture. Discussion The level of participation by members of the WGPA in the ICEs was disappointing in many instances. The task of motivation should be continued in the hope that it may eventually prove successful. Failure to submit results after an analyst had indicated his willingness to participate also presents a serious problem. Although this problem may be due to staff short-ages in some instances it makes planning and execution of an exercise extremely difficult. Nevertheless the sustained efforts by some members of the WGPA are commendable and the level of co-operation achieved is probably the best that can be expected under a system of voluntary participation.TABLE I11 RESULTS OF ICE 3/81 Laboratory code B . . . . . . . . . . E . . . . . . . . . . F1 . . . . . . . . F2 . . . . . . . . G1 . . . . . . . . G2 . . . . . . . . H . . . . . . . . . . 0 . . . . . . . . . . P . . . . . . . . . . Mean Standard'deviatkk : Relative standard deviation % Truevalue . . . . . . * Not identified conclusively. Pesticide concentration/ mg kg-1 (unknown) (known) v . . 26.10* 3.09 . . 5.89 2.76 . . 12.50 1.57 . . 9.90 1.52 . . 19.23 2.75 . . 15.99 2.53 . . 10.91 1.42 . . 5.08* 1.30 ND 1.19 . . 13.20 1.79 . . 7.04 1 .oo . . 53.3 55.9 .. 10.00 1.50 Detector used Column used Flame photometric Electron capture Flame photometric 3% OV-17 Flame photometric 3% OV-17 Flame photometric Flame photometric Flame photometric 5% QF-1 Electron capture 1% DC-11 + 0.5% DEGS 3% SE-30 + 6% QF-1 4% SE-30 + 6% SP-2401 4% SE-30 + 6% SP-2401 1.5% OV-17 + 1.95% QF-1 Thermionic 1.5% OV-17 + 1.95% QF-1 Method of quantitation Peak height Peak height Peak height Peak height Peak height Peak height Peak height Unknown Peak height Although the assessment system used resulted in a low percentage of the results submitted being regarded as unacceptable this may be due to the use of statistical methods not suitable for the small number of results available for evaluation. These methods were used because similar statistical techniques are recommended by organisations such as the Environmental Protection Agency'; however in South Africa far fewer laboratories participate in ICEs.No published statistical methods for evaluating a limited number of ICE results were known and therefore it was decided that in future exercises of this kind a different approach will be adopted. Success of a specific exercise will not be evaluated but laboratories will be provided with their results the true value and the relative error. The relative error is defined as (I N-xi I / N ) x 100 where N = the true value and xi is an individual result. No value judge-ment will be made on the results and it will be left to the individual analyst and his supervisor to assess his performance.These ICEs were not undertaken with the primary aim of demonstrating the state of pesti-cide analysis in South Africa but rather as an effort to create a feeling of confidence responsi-bility and accountability in a diverse group of analysts. In this respect it may be considere 752 VAN DYK et al. CALIBRATION FOR PESTICIDE Analyst Vol. 108 TABLE IV RESULTS OF ICE 4/81 Pesticide concentration (aldicarb sulphone)/ Detector Method of mg kg-l used* Column used measurement 0.66 FPD (S) Carbowax 20M Peak height 0.67 FPD (S) Carbowax 20M Peak height 0.67 FPD (S) Carbowax 20M Peak height 0.68 FPD (S) Carbowax 20M Peak height 0.68 FPD (S) Carbowax 20M Peak height 0.67 0.01 1.5 Laboratory code B F . . F . . FJ N Mean . . Standard deviation .. Relative standard deviation % * FPD (S) flame-photometric detector in the sulphur mode. to have achieved limited success. Each laboratory reporting a result that deviated signifi-cantly from the mean was supplied with a possible explanation of his error. Explanations given were as follows insufficient dilution when using an electron-capture detector; the volume injected was either too large or too small giving rise to errors in precision; the peaks measured were sometimes very small and also outside the range of linearity of the detector; errors may have been made in preparing standard solutions or in diluting ICE samples; sometimes simple arithmetic errors were noted but in other examples it was found that the analysts used com-puting integrators incorrectly ; some laboratories did not properly clean the glassware or chemicals and therefore had dirty samples as could be seen from their chromatograms and this interfered with their quantitation.This feedback service to the participating laboratories was intended to help them improve their general performance. Each ICE sample preparation involved considerable work because the prepared samples had to be analysed in the central laboratory before they were sent to the participants. The time taken for this was well spent because it ensured the integrity of the ICE sample. Previous experience with hastily prepared samples showed that they may be the cause of error in an ICE. Conclusion A large input of time and other resources must be made to run a successful ICE; neverthe-less the value obtained from such an exercise outweighs the costs.The aim of a sustained ICE programme is to give users of analytical results confidence in results they are presented with if such results originate from a laboratory regularly taking part in ICEs. As no certified reference materials are available in the field of pesticide residues analyses ICEs provide a useful tool for indicating to participating laboratories shortcomings of which they may have been unaware. Analysts also gain confidence by participation in an ICE and learn to trust other analysts in their field. Although no laboratory or analyst will ever be perfect this type of continuous ICE programme is the only insurance against complacent and inefficient analysts. This programme is to be continued undertaking 3 4 ICEs annually.APPENDIX Participating Laboratories Laboratories of the following organisations indicated their willingness to participate in ICEs Agricultural Product Standards Division Department of Agriculture ; City Health Department Johannesburg City Council ; Food and Nutritional Products SA (Pty) Ltd. ; Form-Chem (Pty) Ltd. (two laboratories) ; Health Chemistry Laboratory Department of Health and Welfare ; Hydrological Research Institute Department of Environmental Affairs; Marine Pollution Group University of Port Elizabeth ; Plant Protection Research Institute, Department of Agriculture ; Rand Water Board; South African Bureau of Standards; South African Transport Service; Union Carbide SA (Pty) Ltd. ; Veterinary Research Institute, Department of Agriculture ; and Water Research Institute Council for Scientific and Industrial Research J w e 1983 ANALYSIS IN SOUTH AFRICA 753 References 1. 2. 3. 4. 6. 6. 7. Van Dyk L. P. Wiese I. H. and Mullen J. E. C. Residue Rev. 1982 82 42. Youden W. J. and Steiner E. A. “Statistical Manual of the AOAC,” Association of Official Smith R. Water SA 1978 4 161. Greenberg A. E. Moskowitz N. Tamplin B. R. and Thomas J. J. Am. Water Works Assoc., “Annual Book of ASTM Standards Part 31 Water,” American Society for Testing and Materials, International Organisation for Standardisation IS0 5725-1981 (E). US Environmental Protection Agency EPA 600/1-76-017 1979. Analytical Chemists Washington DC 1975. 1969 61 599. Philadelphia PA 1975. Received August 13th 1982 Accepted January 4th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800748

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

754 SHORT PAPERS Analyst June 1983 Spectrophotometric Determination of Exchangeable Calcium in Soils by Chlorophosphonazo-mA Qiu Xing-chu and Zhang Yu-sheng Agricultural Science Research Institute of Ganzhou Prefecture Jiangxi China and Zhu Ying-quan P.O. Box 82 Chengdu China Keywords Chlorophosphonazo-mA ; spectrophotometry ; calcium determina-tion; soils Chlorophosphonazo-mA (CPA-mA) was first proposed by East China Normal University as a spectrophotometric reagent for rare earth elements1 Recently it was used to determine uranium2 and thorium,3 but has previously not been used to determine calcium spectrophoto-metrically. In this work the optimum conditions for the spectrophotometric determination of calcium and the removal of the interference from diverse ions were studied in detail.CPA-mA Experimental Apparatus Analytical Instruments Factory). Instruments Factory). Spectra and absorbances were recorded with a Model 72 spectrophotometer (Shanghai The pH measurements were made with a Model pH S-73 pH meter (Tianjin Analytical Reagents Standard calcizcm solzction 100 pg ml-1. Dissolve 249.8 mg of calcium carbonate (dried at 110 "C) in 40 ml of 2 M hydrochloric acid. Remove the carbon dioxide by boiling and dilute the solution with water to 1 1 and mix well. Dilute an aliquot to give a final calcium con-centration of 20.0 pg ml-l in the working standard solution. CPA-mA aqueous solzction 0.015~0. Triethanolamine (TEA) 1 + 4 solzction. Quinolin-8-ol solzction 0.25% in water. All other reagents were prepared from analytical-reagent grade chemicals.East China Normal University Chem. Indus. Mfg. Procedure Transfer a solution containing not more than 60 pg of calcium into a 25-ml calibrated flask. Add with stirring 2 ml of ammonium acetate (1 M) 1 ml of ammonia (1 + 1) and 3 ml of CPA-mA (0.015%). Dilute to the mark with water. Measure the absorbance of the coloured solution at 630 nm in a 2.0-cm cell against a reagent blank prepared similarly SHORT PAPERS Results and Discussion n z 0.1 a” Q 755 0.2 -n tn 5 0.1 -Absorption Spectra cium. exhibits an absorption maximum at 630 nm. measurements. Fig. 1 shows absorption spectra of aqueous solutions of CPA-mA and its complex with cal-The absorption maximum of the reagent is at 560nm and the calcium complex This wavelength was employed for further 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Effect of Acidity and Salt Medium The optimum pH range in which the absorbance of calcium complex with CPA-mA can be measured is shown in Fig.2. The pH was adjusted with 1 + 1 ammonia by using the pH meter The absorbance is maximal and nearly constant at pH 7.0-9.3. A pH of 9.3 is pre-ferred. In borate medium the sensitivity of this method decreases. 0.6 8 .ff 0 0.4 2 0.2 500 540 580 620 660 Wavelengthlnm Fig. 1. Absorption spectra of CPA - mA and its calcium complex. (1) CPA - mA (against water as reference) 1.0-cm cell. 4.0 0.3 (II +2 n tn a 0.2 7.0 8.0 9.0 10.0 11.0 12.0 PH (2y Ca- CPA-mA complex (against reagent blank as reference) ; Ca taken 60 pg; 2.0-cm cell.Fig. 2. Effect of acidity on absorbance of Ca-CPA-mA complex. Ca taken 40 pg; 3.0-cm cells. (1) Ammoniacal medium; (2) borate medium. Composition of the Complex The composition ratio (calcium to CPA-mA) of the complex was determined by the molar ratio and continuous variation methods. The results obtained showed that the ratio of calcium to reagent in the complex was 1 2 (see Figs. 3 4 and 5). Beer’s Law and Sensitivity of the Reaction 0-60 pg per 25 ml. was 3.9 x lo3 1 mol-l cm-l. Linear calibration graphs passing through the origin were obtained for calcium in the range From this straight line the apparent molar absorptivity coefficient ( E 756 SHORT PAPERS Analyst VOI?. 108 Effect of Temperature and Stability It can be seen that the absorbance is unaffected by temperature over the range from 10 to 60°C.Therefore normal variations in laboratory temperature will have no effect. Besides the complex was formed instantaneously at room temperature (about 25 "C) and the absorbance of the complex did not change for 6 h and then faded slightly. The effect of temperature on the absorbance of the complex is shown in Fig. 6. $ 0.4 C % 2 0.2 a 1.0 2.0 2.5 3.0 4.0 Ca/m I g 0.2 2 LL 10 30 50 70 90 Tern peratu rePC Fig. 5. Molar ratio plot. Variation of Ca against constant CPA-mA; 0.5-4.5 ml of 2.0 x 1 0 - 4 ~ Ca added to 5.0ml of 2.0 x M CPA-mA per 25 ml. Fig. 6. of Ca2+ - CPA-mA complex. 2.0-cm cell. Effect of temperature on absorbance Ca taken 40 pg; Interferences of Diverse Ions Diverse ions were added to a solution containing 40 pg of calcium per 25 ml and the colour was developed by the usual procedure.In the presence of TEA and quinolin-8-01 as masking agents the following ions when present in the amounts (in micrograms) shown in parentheses, do not interfere Mg(I1) (80) Fe(II1) (50) Al(II1) (20) Mn(I1) (lo) Cu(I1) (5) Zn(I1) (lo), W(V1) (5) Cr(V1) (5) Mo(V1) (5) and Pb(I1) (10). Application The proposed method has been applied to the determination of calcium in soils. Procedure Weigh 1 g of soil sample (air dried) into a 50-ml centrifuge tube and add 30 ml of 1 M ammonium acetate. Centrifuge the tube for 3-5 min. (The procedures were repeated in triplicate.) Combine the clear centri-fugate and dilute to exactly 100 ml with 1 M ammonium acetate.Shake well. Pipette 2 ml of this extract into a 25-ml calibrated flask add 1 ml of TEA (1 + 4) and 1 ml of 0.25% quinolin-8-01 following the above general procedure. Stir the contents for a few minutes with a glass rod. TABLE I RECOVERY OF CALCIUM Calcium/ pg Sample Soil type 1 Violet soil 2 Alluvial soil 3 Red soil 4 Red soil Source of parent material Original Purple sandstone and 15.84 shale weathering products River drift 14.40 Quaternary period red 15.28 Granite weathering 8.86 clayey soil product Added 4 8 12 4 8 12 4 8 12 4 8 12 Found 20.34 24.04 27.89 18.90 22.59 26.44 19.54 23.55 27.09 12.87 16.81 21.15 Error/ pg + 0.50 + 0.20 + 0.05 + 0.50 + 0.19 + 0.04 + 0.26 + 0.27 - 0.19 + 0.01 - 0.05 + 0.29 Recovery, 102.5 100.8 100.2 102.7 100.8 100.2 101.3 101.2 99.3 100.1 99.2 101.0 Jum 1983 SHORT PAPERS 757 Results It can be seen that the recoveries of calcium from standard solutions were >99% and the relative standard deviation (eight determinations) was (3.08%’ The results shown in Table I1 are in reasonable agreement with those obtained by atomic-absorption spectropho tometry.The results obtained by the recommended method are listed in Tables I and 11. TABLE I1 ANALYTICAL RESULTS AND COMPARISON Calcium/mequiv. per 100 g soil f A \ Atomic-absorption Relative standard Sample spectrophotometry Present method* deviation yo 1 10.60 10.65 0.87 2 2.10 2.21 3.08 3 5.06 5.12 1.4 4 7.51 7.48 1.23 * Mean of eight determinations. The authors are grateful to Zhang Lin P.O. Box 82 Chengdu China for her assistance. References 1. 2. 3. East China Normal University Lihua Jianyan 1979 15 1. Xue Guang Huaxue Shiji 1982 4 332. Xue Guang and Zhu Ying-quan unpublished work. Received October 4th 1982 Accepted January 4th 198

ISSN:0003-2654    

DOI:10.1039/AN9830800754

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Jum, 1983 SHORT PAPERS 757 Histochemical Demonstration of Collagen in Comminuted Meat Products F. Olga Flint and Barry M. Firth Procter Department of Food Science, University of Leeds, Leeds, LS2 9JT Keywords : Specific collagen test ; meat-product microscopy ; connective tissue ; rind The connective tissue associated with meat is largely collagenous and in meat products this intrinsic collagen may be accompanied by processed collagen in the form of cooked chopped pigskin (rind). Rind is a traditional ingredient in many meat products and there is no legal limit to the amount of rind that products may contain. However, in the assessment of meat products only amounts of rind “naturally associated with the flesh used” can count toward meat c0ntent.l Because of their chemical similarity, connective tissue and rind are assessed together.The collagen they contain is rich in hydroxyproline2 and because skeletal muscle and plant materi- als contain only trace amounts of hydroxyproline its determination can be used to measure the collagen content of meat products. Chemical methods for hydroxyproline, such as the widely used Stegemann and Stalder3 technique, measure the total collagen present in meat products but do not distinguish between hydroxyproline derived from connective tissue and that from rind. None of the histological techniques for collagen tested by Hildebrandt et aZ.* distinguished between cooked and uncooked collagen. Their preferred method for the computer-assisted758 SHORT PAPERS Artalyst, VoZ. 108 image analysis of collagen in raw meat products was the van Gieson technique5 in which collagen is stained red by the acid fuchsin component and other tissues are coloured yellow by the picric acid present in the stain.A disadvantage of the van Gieson technique is that the stain tends to fade and for this reason Sweat et aZ.6 replaced the fuchsin component with Sirius Red F3BA, which has better light fastness. They found that Sirius Red gave a more intense colour to the collagen and that its birefringence was affected. The Sweat et al. picro-Sirius method was later examined by Junqueira et aZ.7 who suggested that staining with Sirius Red when combined with enhanced birefringency was a specific histochemical test for collagen. Their explanation is that Sirius Red F3BA is an elongated molecule containing sulphonic acid groups that can react with the basic groups of collagen. The increased birefringence is due to the dyestuff being bound to the collagen so that the axes of both dye and collagen molecules are parallel, thus reinforcing the natural birefringence of the protein.In this investigation the improved picro-Sirius method recommended by Junqueira et al.' has been adapted for unfixed cryostat sections by the introduction of a de-greasing followed by a rehydration stage. To demonstrate any change in birefringence, control staining using acid fuchsin in place of Sirius Red F3BA was carried out simultaneously. Experimental Preparation of Picro-Sirius Red Stain of saturated aqueous picric acid solution. staining solution was allowed to stand for 24 h.A 10-ml volume of 1% m/V aqueous Sirius Red F3BA (C.I. No. 35780) was added to 90 ml A few picric acid crystals were then added and the Preparation of Picro-fuchsin Control Stain No. 42685) for Sirius Red FSBA. This was prepared exactly as the picro-Sirius Red stain but substituting acid fuchsin (C.I. Sample Preparation Materials. Method. A 5-g mass of the dried rind was stirred into 15 ml of distilled water and was allowed to stand for 1 h. A 10-g mass of the moist rind was then thoroughly mixed with 90 g of sausage meat using a glass pestle and mortar. Three samples of approximately 1 g each were taken from the sausage meat - rind mixture and 10-pm cryostat sections were prepared from each sample by rapid freezing in liquid nitrogen and sectioning in a cryostat at -20 "C. Sections cut from the three samples were stored separately until needed for staining.Commercial pork sausage, dried rind supplied by Protein Foods, Tipton. Staining Procedure The sections were placed in acetone for 20 min to remove fat and then in distilled water for 1 rnin before staining in either picro-Sirius or the picro-fuchsin solutions for 60 min. The sections were then rinsed in 0.01 M hydrochloric acid for 2 min and dehydrated by passing through 30, 70 and 90% aqueous ethanol (1 min each) followed by two changes of absolute ethanol (1 min each). The sections were then cleared in two changes of xylene (3 rnin each) and were mounted in DPX (refractive index 1.524). Results and Discussion In sections from the three samples the picro-Sirius and the picro-fuchsin both stained collagen red and muscle cells yellow, leaving starch granules unstained (see Fig.1). Sirius Red coloured collagen more intensely than did acid fuchsin and this enabled the finer collagen fibres to be seen more easily. Both techniques coloured muscle tissue a similar yellow colour and when viewed by polarised light muscle cells showed a weak yellowish white birefringence. The birefringence of the collagen seemed unaffected by the pkro-fuchsin stain, collagen associated with muscle cells retained its birefringence and cooked rind showed no birefringence. In contrast, the picro-Sirius technique caused connective tissue associated with muscle cells to glow a bright golden yellow colour when viewed by polarised light, even the finest collagen(bl (dl Fig. 1.Cryostat sections of commercial sausage with added rind. Bars represents 100 pm. (a) Picro-Sirius stained section viewed by transmitted light: muscle cells coloured yellow, connective tissue (arrows) and rind (r) stained red. (b) Same field as (a) viewed by polarised light: muscle cells show weak yellowish white birefringence, connective tissue exhibits a bright golden yellow birefringent colour and rind a red birefringent colour, starch granules are bright with a well marked cross. (c) Picro-fuchsin stained section viewed by transmitted light: muscle cells coloured yellow, connective tissue (arrows) very weakly stained red and red stained rind (r). (d) Same field as (c) viewed by polarised light: muscle cells and connective tissue show similar weak yellowish birefringence, rind shows no birefringence, starch granules are bright with a well marked cross.[:oface 9. 758 Starch granules unstained.Jzcne, 1983 SHORT PAPERS 769 fibres being detectable. The rind, which showed no birefringence when stained with picro- fuchsin, now appeared birefringent. The colour is much darker and redder than that shown by the raw collagen and although less bright the birefringence is unmistakable. This difference in the birefringent colours of connective tissue and processed rind could form the basis for the positive identification of rind in the presence of connective tissue. Viewed with normal transmitted light the intense red staining of collagen and the sharp contrast with yellow coloured muscle tissue could aid in the quantitative microscopy of connective tissue in meat products and it would be of use in machine-aided analysis where contrast is a prime req~irement.~ Although the results confirm Junqueira et aZ.’s view that only Sirius Red stained collagen exhibits enhanced birefringence, food products contain such a diverse range of ingredients that their behaviour with Sirius Red needs to be investigated.This study and the establishment of optimum conditions for the use of Sirius Red FSBA with processed foods are the subject of current work. References 1. 2. 3. 4. 6. 6. 7. “The Sausage and Other Meat Product Regulations 1967,” SI 1967 No. 862 amended by SI 1968 Egan, H., Kirk, R. S . , and Sawyer, R., “Pearson’s Chemical Analysis of Foods,” Eighth Edition, Stegemann, H., and Stalder, K., Clin. Chim. Acta, 1967, 18, 267. Hildebrandt, G., Konigsmann, R., and Kretschmer, F. J., FZeischwirtschuft, 1977, 57, 689. van Gieson, I., N.Y. Med. J., 1889, 50, 57. Sweat, F., Puchtler, H., and Rosenthal, S. I., Arch. Puthol., 1964, 78, 69. Junqueira, L. C. U., Bignolas, G., and Brentani, R. R., Histochem. J., 1979, 11, 447. No. 2047, HM Stationery Office, London. 1981, Churchill Livingstone. London, p. 399. Received December 20th. 1982 Accepted January 7th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9830800757

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Jzcne, 1983 SHORT PAPERS 769 Evaluation of the Determination of High Levels of Total Cadmium in Foodstuffs Using Flame Atomic-a bsorption Spectrophotometric Measurement Dorothy Dellar De9artment of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SEl 9NQ Keywords : Cadmium determination ; foodstufls analysis ; flame atomic- absorption spectrophotometry There is no evidence that cadmium is essential to human nutrition1; however, the accumulation of cadmium in the body does give cause for concern because of its long biological half-life and the damage that high levels can do to the kidneys. In the environment and generally in food- stuffs levels are low, but they may be high in shellfish and animal kidney. In both these instances, high levels of cadmium can occur from natural causes and may be further elevated by localised pollution.The conventional chelation and extraction procedure for low level determinations of cadmium in digested foodstuffs2~3 is unnecessary for these higher levels. A more direct method, involving no pre-treatment of digests prior to measurement by atomic- absorption spectrophotometry, can be conveniently used. Crown Copyright.760 SHORT PAPERS Experimental Analyst, Vol. 108 Reagents All reagents should be of the grade specified; solutions should be prepared with distilled water. Nitric acid, sp. gr, 1.42. BDH Aristar, Hopkin and Williams Ultrar or an equivalent grade of acid is recommended. SuZphuric acid, sp. gr. 1.84 and dilzcte (1 + 19). Standard solzctions.A primary standard solution, specially prepared for atomic-absorption spectrophotometry, can be obtained from commercial sources and contains 1.0 g 1-l of cad- mium in 1 N nitric acid. Dilute this to give an intermediate standard containing 5 mg I-l, ensuring an acidity of at least 0.2% V/V sulphuric acid. From this standard prepare working standards containing 0 4 . 3 mg 1-1 of cadmium in sulphuric acid (1 + 19), standards O-H. Grades as recommended for nitric acid. Apparatus All-glass apparatus must be kept permanently full of 1 N nitric acid when not in use. Atomic-absorption spectrophotometer. An Instrumentation Laboratory 453 double-beam instrument was used, which was fitted with a Counter-Flo Jet nebuliser and a 100-mm triple- slot burner; the incident beam was 30 mm above the burner.The air and acetylene pressures were 4.5 and 4.8 lb in-2, respectively. Hydrogen background correction was used on both the sample and compensating radiation beams with a single element hollow-cathode lamp for the cadmium. Procedure Sample digests should be prepared according to method (1)C of the Analytical Methods C~mrnittee,~ taking the precautions subsequently described.2 The resulting 100 ml of digest, in nominally 5% V/V sulphuric acid, should be colourless and contain no suspended solids. Two reagent blanks using the volumes of acid used in sample oxidation are prepared at the same time. Tune the emission source to give the maximum sensitivity to noise ratio according to the maker’s instructions at a wavelength of 228.8 nm, allowing the appropriate warm-up time for the hydrogen lamp used for background correction.Obtain a steady base line on a millivolt recorder, aspirating water. Read the samples, standards and blanks in random order, but with a reagent blank in each half of the series. Calibration For the results given for retail foods, reference materials and evaluation of the interferences, nine series of standards (O-H) were measured as described but measurement of each standard was duplicated randomly within each series. Each average net response for standards A-H may be expressed in terms of standard A, and the average net response, expressed as for stan- dard A, calculated for each calibration line. The ratio of the sum of net responses for each standard to the sum of the net average response is shown in Table I.The within-series stan- dard deviation, Sw, may be derived for each standard concentration from the difference between TABLE I CALIBRATION DATA Standard solution . . .. .. .. Cadmium concentration/mg 1-l . . .. A B C D E F G H 0.01 0.02 0.05 0.10 0.15 0.20 0.25 0.30 Ratio of net standard to net average response 0.95 1.05 1.02 1.01 1.00 1.00 0.99 0.98 I A \ Coefficient of variation of measurement, yo, for mean of 2 readings Within-series standard deviation . . .. Between-series standard deviation .. j . 3 3.8 3.9 2.4 1.9 1.9 1.5 l.B 19 8.1 4.4 2.4 1.8 0.9 1.5 2.6June, 1983 SHORT PAPERS 761 the duplicate response for all the nine standard series. The between-series standard deviation, sb, at each standard concentration may be calculated from the difference between the net mean individual responses with the net average response of each of the nine calibration lines.5 Relevant standard deviations are also shown in Table I.For the calculation of the results obtained in application to foodstuffs, the calibration was computerised using a third-order polynomial curve-fitting programme. This is naturally biased for the higher concentration standards, but for the lower concentration standards there will be a deviation for each point in the measurement range from the best fitted curve, from which standard deviations may be calculated. The average for standards 0-F (40 degrees of freedom) was 0.0017 mg 1-l. Any calibration point exceeding the 95% confidence interval (95 CI) of 0.0034 mg 1-1 (calculated from this) was discarded and the calibration graph re- calculated.Interferences The major inorganic species in digests of retail foods will be sodium, potassium, calcium, magnesium and phosphorus, the last element being present as orthophosphate after dilution of a concentrated sulphuric acid digest. From known levels of these elements in foodstuffs,s 100-ml volumes of digests prepared from 10 g of foodstuffs are unlikely to contain amounts exceeding 200 mg of sodium, potassium or phosphorus, 40 mg of calcium and 50 mg of mag- nesium. Only about 5% of the listed retail foods would give digests containing half these amounts of major inorganic species. These major inorganic species were therefore tested individually for interference at the higher amounts listed, at different burner heights from 10 to 30 mm and varying flame conditions for standard concentrations 0, D and G in 5% V/V sulphuric acid.Although the 95% confidence limits (95 CL) were not exceeded for any condition used, a 30-mm burner height and a normal flame gave the most stable conditions. As iron, copper, zinc and tin may also be present at higher levels in animal offal and some shellfish, iron, zinc and tin at 1 mg together with copper and manganese at 0.2 mg were also tested. Standards containing mass ratios of calcium plus magnesium to phosphorus of 1.8 and 0.9 were also measured. Each complete series of solu- tions was measured, with standards 0-H, in duplicate within the same series of readings; 95 CL were those derived from Sw, obtained for the calibration.The results given in Table I1 show that these limits are not exceeded, implying the absence of any direct interferences, i.e., systematic bias, from these inorganic species when measurement was made a t the optimised triple-slot burner height of 30 mm using a Counter-Flo Jet nebuliser. TABLE I1 INTERFERENCE EFFECTS ON MEASUREMENT OF CADMIUM Concentration/mg 1-1 96% confidence limits Added ion Na+ K+ Gas+ Mga+ P Cas+ Mga+ P Ca2+ Mga+ P Fe3+ Zna+ Sn4+ CUB+ Mna+ * . .. .. .. .. .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Over-all average .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Amount/ mg 200 200 40 60 200 :: } :: } 100 60 0.2 '1 0.2 .. 0.1 0.26 f0.05 f0.03 Relative difference in response with and without added element r -0.05 - 0.02 - 0.01 -0.05 - 0.04 -0.01 - 0.02 - 0.03 - 0.03 1 - 0.02 0.00 -0.01 - 0.01 0.00 - 0.03 -0.01 - 0.01 -0.01762 SHORT PAPERS Analyst, Vol.108 Results and Discussion The accuracy of the method in application to foodstuffs was obtained for standard reference materials with high cadmium-containing foodstuffs (bovine kidney and crab) included to give a wider range of measurement. Each sample was digested six times (two triplicates), and each triplicate total analysis was measured within the same series of readings. The mass of reference materials digested reflected an equivalent amount of the corresponding wet sample. Results are summarised in Tables I11 and IV. TABLE I11 EVALUATION OF ACCURACY OF METHOD IN APPLICATION TO FOODSTUFFS All means are for 6 results.Reference material NBS tomato NBS oyste; Bowen’s kale NBS spinach leaves tissue Meanlmg kg-l . . .. .. . . 1.04 1.47 2.66 3.68 Certified valuelmg kg-l . . .. * . 1.01 f 0 . l l t (1.5)* (3.0)* 3.5 f 0.4 f95 CI of meanlmg kg-l . , . . 0.04 0.25 0.19 0.12 * Uncertified. t The value for Bowen’s kale is taken from reference 2. The figures in Table I11 can be used to assess the accuracy of the method in application to reference foodstuff materials. From Table IV, comparison of the standard deviations (ex- pressed as coefficient of variation) with S b for measurement of standard solutions at equivalent concentrations (times 43 to represent single measurement) indicate that ratios fall within relevant 95 CLs of 0.48-2.35 except for the kale sample.This indicates little significant contribution to the variation of results from digestion of different foodstuff matrices. The 95 CI for results for total analysis may be calculated from relevant between series standard deviations and are shown in Table IV for a 10-g sample mass. The limit of detection is obtained as described previously5 and is based on the formula t,.,,,,s. From amounts determined up to 7 pg (spinach, kale, tomato leaves and oyster tissue) this limit of detection is 0.7 pg (20 degrees of freedom) and hence, for 10 g of foodstuffs digested to produce 100 ml of solution, 0.07 mg kg-l in the original sample. Conclusion A method for the determination of high cadmium levels in foodstuffs is described, involving destruction of organic matter by wet oxidation and direct measurement on the sulphuric acid digests. Under optimised measuring conditions with normal levels of interfering species found in foodstuffs, the sulphuric acid digests can be measured without systematic bias.TABLE IV REPLICATE ANALYSES FOR CADMIUM ON FOODSTUFFS AND STANDARD REFERENCE MATERIALS Each row represents six results obtained by one analyst. Sample mass/ Foods tuff g NBS spinach . . . . . . 1 Bowen’s kale .. .. .. 2 NBS tomato leaves . . . . 2 NBS oyster tissue . . . . 2 Crab . . . . . . . . 4 Kidney .. .. .. 10 Beef kidney .. .. .. 10 Mean/ Rangel mg kg-l mg kg-l 1.47 1.13-1.78 1.04 0.98-1.09 2.66 2.52-2.96 3.68 3.58-3.83 1.41 1.38-1.44 3.40 3.33-3.51 2.55 2.44-2.64 Standard deviation, Sb 95% Confidence Amount/ A , intervals*/ Irg % Erg mgkg-’ fmgkg-’ 1.47 16.1 0.237 0.237 0.061 2.08 3.8 0.078 0.039 0.02 5.31 6.8 0.363 0.181 0.093 7.36 3.0 0.220 0.110 0.056 13.6 1.9 0.251 0.063 0.065 14.1 2.3 0.316 0.032 0.081 25.5 2.6 0.665 0.067 0.17 Concentrations are based on a 10-g sample mass.This limit is for a single result derived from the reproducibility.J m e , 1983 SHORT PAPERS 763 The agreement with certified or concensus mean values for reference foodstuffs indicates that the method would give satisfactory results when applied to retail foods containing cadmium levels above 0.07 mg kg‘l with 95% confidence intervals of & 0.10 mg kg-1 up to 1 mg k g l and above this level to corresponding intervals of &-loyo. This paper is published with the permission of the Government Chemist. References 1. 2. 3. 4. 5. 6. WHO Tech. Refi. Ser., No. 532, 1973. Evans, W. H., Read, J. I., and Lucas, B. E., Analyst, 1978, 103, 580. Analytical Methods Committee, Analyst, 1975, 100, 761. Analytical Methods Committee, Analyst, 1960, 85, 643. Evans, W. H., Dellar, D., Lucas, B. E., Jackson, F. J., and Read, J. I., Analyst, 1980, 105, 529. Paul, A. A., and Southgate, D. A. T., “McCance and Widdowson’s The Composition of Foods,” Received January 25th, 1982 Accepted December 29th, 1982 H.M. Stationery Office, London, 1978.

ISSN:0003-2654    

DOI:10.1039/AN9830800759

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

J m e , 1983 SHORT PAPERS 763 Further Studies on the Recovery of Iodine as Iodine425 After Alkaline Ashing Prior to Assay G. Bryan Belling CSIRO Division of Human Nutrition, Kintore Avenue, Adelaide, South Australia 5000 Keywords : Iodine determination ; biological materials assay ; iodine- 125 In an earlier paper1 it was reported that alkaline ashing of plant material a t 600 "C was the most reliable method of oxidation prior to the final determination of iodine using the Sandell and Kolthoff reaction.2 Two questions have arisen concerning this method of ashing since the original study. For most tissues the recommended method of oxidation1 is satisfactory; however, a few plant materials leave an unacceptably large amount of unburnt carbon in the ash and it is necessary to establish whether ashing at a higher temperature to minimise the amount of unburnt carbon would still give satisfactory recoveries of iodine.Also, personal communication has revealed that some laboratories, using various methods of alkaline ashing prior to determination of iodine in urine, have experienced poor recoveries. I t was thought that the ashing method recommended by Jones et aZ.1 might be adapted satis- factorily for urine prior to either a manual determination or to an automated determination3 in instances where interfering substances cause poorly shaped peaks. Ashing of urine is com- plete at 600 "C. The use of added iodine-125 provides a sensitive and accurate means of studying and clarify- ing both of these problems. Experimental Raising the Ashing Temperature Dried plant materials suitable for feedstuffs were finely ground and 300-mg samples of the ground material were placed in 16 x 125-mm Pyrex text-tubes. A 0.5-ml volume of de- ionised glass re-distilled water was added to wet the sample followed by 1 .O ml of 1 M potassium hydroxide solution.A 15-pl aliquot of a solution of Na1251 containing at least 50000 c.p.m. (specific activity 186 MBq pg-l of iodine) was added to the tube and the radioactivity counted. The supplied iodine-125 was used to calibrate the counter. The tubes and contents were dried in an air oven for 18 h at 105 "C. The tubes, in a stainless-steel rack, were placed horizontally764 SHORT PAPERS Analyst, Vol. 108 in a cold muffle furnace and heated at 150 "C for 30 min after which the temperature was raised to 650 "C and maintained for 2 h.The furnace door was opened four times during this period to renew the air in the furnace. The ash was dissolved in 6 ml of water by heating to almost boiling. When cool, the tubes were centrifuged a t 700 g for 10 min. A 1.5-ml aliquot was transferred into a clean 16 x 125-mm test-tube, 300 mg of the original material added to the solution, in order to maintain the same counting conditions as prior to ashing, and the iodine- 125 activity counted and recoveries calculated. This procedure was repeated using ashing temperatures of 680 and 720 "C. Ashing of Urine Thirty one samples of human urine were ashed in batches of 4-5 by the following procedure. Aliquots of 3 ml of urine were placed in 16 x 125-mm Pyrex test-tubes, 1.0 ml of 1 M potassium hydroxide solution and 15 ,u1 of Na1251 solution containing about 50000 c.p.m.(specific activity 186 MBq pg-l of iodine) were added and the radioactivity counted. The tubes were dried for 18 h at 105 "C, then ashed, as described above, at 600 "C. The ash was dissolved in 6 ml of de- ionised glass re-distilled water. A 4-ml aliquot was transferred into a clean 16 x 125-mm Pyrex test-tube, giving the same counting conditions as prior to ashing, the radioactivity was counted and the recovery of iodine-125 calculated. Counting of Iodine-125 Counting of iodine-125 was carried out using a EKCO, Type N644A, scintillation counter, which has a sodium iodide crystal well detector of 150 mm diameter and an EKCO automatic scaler, Type N530F.Corrections were made for radioactive decay. Results and Discussion Plant Material Table I shows the recovery of iodine-125 after ashing the plant material feedstuff at tempera- tures of 650,680 and 720 "C. From these results it can be seen that the recoveries of iodine- 125 from plant material feedstuff (96.6 -+ 2.5-100.5 * 2.3%) ashed at 650 "C are satisfactory and compare favourably with those obtained at 600 "C by Jones et aZ.l (95.4 1.8%) for a range of biological materials. After ashing at 650 "C there appeared to be less unburnt carbon in the ash of the pea pollard and wheat samples, the two materials with most carbon after ashing at 600 "C. When the ashing temperature is raised to 680 "C or above serious losses of iodine occur. It is important that the sample - potassium hydroxide solution mixture does not froth above the volume of the extracting water otherwise apparent losses may occur owing to difficulty in dissolution of the iodide from the sides of the tube above the water level.2.7-99.7 TABLE I RECOVERY OF IODINE-125 FROM PLANT MATERIAL AFTER ASHING Recovery ( f standard deviation),* yo A f \ Sample 650 O C 680 "C 720 OC Maize.. . . .. . . 100.5 f 2.3 (9) 80.2 f 5.4 (3) 87.6 f 6.4 (3) Pea pollard . . .. . . 99.4 f 3.1 (9) 97.6 f 1.1 (3) 91.6 f 2.7 (3) Wheat . . .. . . 97.1 f 4.0 (6) 95.0 f 2.3 (2) - Mung bean .. .. . . 99.4 f 1.7 (3) 89.2 f 2.0 (3) - Low iodine diett . . . . 96.6 f 2.5 (5) 81.5 f 5.6 (3) 78.1 f 5.2 (2) * The figures in parentheses indicate the number of observations. t This diet consists of maize (60%), peas (15%), torula yeast (15%), meat meal (15%) and a mixture of vitamins and minerals. The conclusion is that ashing of plant material at 650 "C reduces the amount of unburnt carbon but does not significantly affect the recovery.J m e , 1983 SHORT PAPERS 765 Urine The mean recovery of iodine from sodium iodide labelled with iodine-125 in the 31 urines tested was 97.2 & 5.3% (& standard deviation). The method described above therefore provides a satisfactory method of ashing urine prior to either manual determination or the AutoAnalyzer m e t h ~ d . ~ The AutoAnalyzer method normally uses dialysis instead of ashing, however, dialysis does not always remove interfering substances and ashing then becomes necessary. References 1 . 2. 3. Jones, G. B., Belling, G. B., and Buckley, R. A., Analyst, 1979, 104, 469. Sandell, E. B., and Kolthoff, I. M., Mikrochim. Ada, 1937, I, 9. Garry, P. J., Lashley, D. W., and Owen, G. M., CEin. Chem., 1973, 19, 950. Received December 7th, 1982 Accepted January loth, 1983

ISSN:0003-2654    

DOI:10.1039/AN9830800763

出版商:RSC    

年代:1983

数据来源: RSC