摘要:

FEBRUARY, 1969 THE ANALYST Vol. 94, No. I I15 Multi-element Atomic-absorption Analyses with a Gas-stabilised Arc as Primary Light Source BY H. G. C. HUMAN, L. R. P. BUTLER AND A. STRASHEIM (Natio.lza1 Physical Research Laboratory, CSIR, Pretoria, South Africa) The use of a gas and wall-stabilised arc as primary light source for multi-element atomic-absorption measurements was investigated. A solu- tion containing the elements of interest was injected, in aerosol form, into the arc, and the resonance lines excited. The concentration of the solution sprayed into the arc was the only source parameter appreciably influencing the sensitivity (i.e., the degree of atomic absorption by the flame). It was found for fourteen arbitrarily chosen elements that the sensitivity was equal to two thirds of the sensitivity obtained with hollow-cathode lamps as primary sources.The reproducibility of measurement, when non-absorbing reference lines were used, was similar to that obtained with hollow-cathode lamps. The intensity of the background emission of the arc was so low in comparison with the line intensities that it had little effect on the sensitivity. With this source, the atomic-absorption working range can be extended by the variation of the solution concentration in the arc, and this proved useful when lead and zinc were determined simultaneously in a brass sample. THE ideal primary light source for multi-element atomic-absorption spectroscopy is one capable of emitting the resonance lines of several elements simultaneously with high stability, low background and narrow line widths, together with the ability to select any element or element combination required.Various attempts to produce such a light source have been made,lJJJ and the gas and wall-stabilised arc developed by Kranz6S6 was investigated for the following reasons : the high stability of spectral-line emission ; the low background continuum, which is probably caused by the failure of the portion of the plasma used as the light source to conduct the current that maintains the arc; and the ability to excite simultaneously the resonance lines of any number of elements introduced simultaneously into the plasma in solution. EXPERIMENTAL APPARATUS- The construction of the arc chamber was similar to that described by ICranz.696 For this investigation thoriated tungsten was used for both electrodes, while nitrogen was used for stabilising the arc and for nebulising the solutions introduced into the arc. The aerosol was introduced into the plasma by two horizontal tubes along the optical axis, as shown in Fig.1. Because of the high rate of gas consumption of the pneumatic atomiser used, the gas velocity through the sample feed-tubes is high, and the sample is forcibly injected into the plasma. A current-stabilised, direct current power supply, manufactured by Messrs. R.S.V., Hechendorf, Pils., W. Germany, which gives currents in 5-amp steps up to 35 amps, was used. In the optical system used (see Fig. 2) a 1-to-1 image of the arc was formed on a dia- phragm. Two regions could clearly be distinguished in the plasma flame, viz., a luminous central core and a light blue surround.The diaphragm enabled the three regions indicated in Fig. 1 as h,, h2 and h,, of 4-cm height, to be separated and the most suitable to be used 81 0 SAC and the authors.82 HUMAN, BUTLER AND STRASHEIM : MULTI-ELEMENT ATOMIC-ABSORPTION [AmzEyst, Vol. 94 as the primary light source. A second lens rendered the light parallel before transmitting it through either an air - acetylene or a nitrous oxide - acetylene flame. The light was finally focused on the collimator of the spectrograph, and quartz lenses were used. Sample Sample Fig. 1. Diagram showing method of sample intro- duction into the arc plasma, and the three regions investigated A medium Hilger spectrograph, equipped with a four-channel direct-readmg Strasheim attachment,2 was used.The entrance slit was set at 3 O p with fixed exit slits of 6Opm. Lens Direct reading spectrometer no I Fig. 2. Optical system INVESTIGATION OF PARAMETERS INFLUENCING SENSITIVITY- The influence of certain source parameters on the sensitivity of absorption was investi- gated. Fig. 3 shows the influence of the solution concentration on the sensitivity at various heights, h, of the plasma flame. These results were obtained for magnesium, with a constant concentration of the solution in the absorbing flame. It is evident that an optimum concen- tration is obtained for each height, but that the maxima are all equal.Februarv. 19691 ANALYSES WITH A GAS-STABILISED ARC AS PRIMARY LIGHT SOURCE 83 I t Concentration of arc solution, p.p.m. Variation of absorption witb concentraticm of the arc solution, with three different regions used as primary light source (for magnesium) Fig.3. The arc current was found to have a slight influence on the optimum concentration as well as on the sensitivity. The minimum current at which the arc would burn was 25 amps; by varying the current between this value and a maximum of 35 amps, the results in Table I were obtained. A slight increase in sensitivity was found with increasing current. TABLE I OPTIMUM CONCENTRATION OF ARC SOLUTION AT DIFFERENT HEIGHTS FOR DIFFERENT CURRENTS (FOR MAGNESIUM) Optimum concentration of arc solution, p.p.m. Current, amps hl h, h, 25 28.2 40.7 62.5 30 31-6 41.7 56-2 35 33.9 43.7 58.9 For all of these tests the tangential gas flow was kept constant at 8 litres per minute.No difference in sensitivity could be detected for gas flows of 6.9 and 9.2 litres per minute at heights of h, or h,. Concentration of absorbing solution, p.p.m- Fig. 4. Working curves: A, with hollow-cathode lamp as primary source: B, with Kranz arc its primary source From these results it was concluded that the only parameter critically affecting the sensi- tivity was the element concentration in the arc solution. Therefore, for all further tests, the tangential gas flow was kept constant at 8 litres per minute, the current was set at 30 amps, and only the region h, cd the arc was used as the primary light source. An example of a magnesium working graph, shown in Fig. 4, is obtained by using the Kranz arc as primary84 HUMAN, BUTLER AND STRASHEIM : MULTI-ELEMENT ATOMIC-ABSORPTION [AfidJASt, VOI.94 source with these settings and spraying a solution of 30 p.p.m. of magnesium into it. The working graph obtained by using a hollow-cathode lamp at 5 mA is also shown. The ratio of the two sensitivities obtained with the two different sources was 0.65. (The sensitivity ratio is defined as the ratio of the two concentration values giving the same absorption signal, which is taken throughout as 10 per cent. absorption.) It can be seen that the working graph, with the arc as source, deviates from linearity slightly more than the hollow-cathode lamp graph. The line-to-background ratio in the arc was 27 : 1, while with the hollow-cathode lamp it was better than 100: 1.REPRODUCIBILITY- Magnesium was used as the test element in an investigation of the reproducibility of measurement, with both the Kranz arc and a hollow-cathode lamp as primary light sources. With the arc, 30 p.p.m. of magnesium were used in the primary solution, together with 30 p.p.m. of calcium added as internal standard. The calcium 422-7 nm line was used as reference line. The standard deviation of the absorption values was determined when various concentrations of magnesium were sprayed into the absorbing flame, and a mean standard deviation of 0.488 per cent. absorption was obtained. When converted into concentration values by means of the working graph, curve A of Fig. 5 was obtained, the shape of which is in accordance with the general shape of curves of this type.' 20 40 60 80 Absorption, per cent. Fig.6. Relative standard deviation curves : A, with Kranz arc as source: B, with hollow- cathode lamp as source With the hollow-cathode lamp as source, the non-absorbing magnesium 383.8 nm line was used as reference line, and a mean standard deviation of 0.584 per cent. absorption was obtained, which resulted in curve B of Fig. 5. The standard deviation of the absorption values is lower with the Kranz arc as source than with the hollow-cathode lamp. This is probably because the correlation coefficient between the magnesium 285.2 nm and calcium 422.7 nm line intensities in the arc is better than the coefficient obtained between the magnesium 285.2 nm and 3834 nm lines from the hollow-cathode lamp. The values obtained were 0.988 and 0.972, respectively.From Fig. 5 it is clear that the coefficient of variation of the concentration values with the arc as source is lower for relatively small absorptions, and it reaches its minimum value of 1.77 per cent. at about 40 per cent. absorption. The minimum coefficient of variation of 1-65 per cent. with the hollow-cathode lamp as primary source is obtained at a higher absorption level of 55 per cent. This is as a result of the greater deviation from Beer's law of the working graph obtained with the Kranz arc as primary source. BACKGROUND- For insight into the possibilities of universal application of this arc as a primary light source for atomic-absorption spectroscopy, a knowledge of the background emission of the arc is essential.The background emission was photographed with the same spectrograph,February, 19691 ANALYSES WITH A GAS-STABILISED ARC AS PRIMARY LIGHT SOURCE 85 I I I I I I I 290 300 320 350 370 390 420 500 Wavelength, nm Fig. 6. Densitometer tracing showing background emission from the arc, between 280.0nm and 500.0nm. The regions below 280-0nm and above 500.0nm are essentially free from background and the result is shown in Fig. 6. The band systems shown in Table I1 were identified with the aid of tables by Pearse and Gaydon." The regions of heavy background emission are indicated in Table 111, together with resonance lines of the elements in these regions. Of the 65 elements for which atomic-absorption sensitivities have already been measured, only twelve have their most sensitive resonance lines in these high background regions.TABLE I1 BAND SYSTEMS IDENTIFIED IN SPECTRUM OF NITROGEN PLASMA (MOST INTENSE HEADS OF EACH SYSTEM IN BOLD FIGURES) Wavelengths of band heads detected, nm System r A > 2nd positive system of N, 405.9 399.8 394.3 380.4 375.5 371.0 367.1 364.1 357.6 353*6* 350.0 337.1. 333.9 330.9 328.5 326-8 315-9 313.6 311.6 310.4 297.6 296.2 295.3 281.9 281.4 383-5 358.2 356.3 354.8 353-8 281.9 282.9 1st negative system of N,+ 427-8 423.6 419-9 416.6 414.0 391.4 388.4 306.4 nm system of OH 302.1 306.4 306.7 307.8 308-9 281.1 338.0 nm system of NH 336.0 337.0 * Uncertain because of superposition by N,+ and NH heads. TABLE I11 REGIONS OF HEAVY BACKGROUND Resonance lines involved Region, nm r A > 291.0 to 297.8 305.1 to 316.0 Aluminium 309-2 Molybdenum 313.3 328.0 to 358.3 Silver 328.0 Chromium 357-9 Lanthanum 357.4 Lutetium 335.9 Rhenium 346-0 Rhodium 343.5 Ruthenium 349-9 371.0 to 391-4 Scandium 391.1 415.0 to 427.8 Calcium 427.7 Dysprosium 421-1 Although the line-to-background ratio of the silver resonance line at 328.0 nm under optimum sensitivity conditions (200 p.p.m.in primary solution) was only 13 : 1, it still gave 67 per cent. of the sensitivity obtained with a hollow-cathode lamp as primary source. Up to an absorbance of 0.30, the working graph shows no larger deviation from linearity than was observed for magnesium (see Fig. 4). For the elements chromium, molybdenum, alu- minium and calcium the sensitivity ratios found were 0.65, 0.71, 0.70 and 0.63, respectively.86 HUMAN, BUTLER AND STRASHEIM : MULTI-ELEMENT ATOMIC-ABSORPTION [AmZySt, VOI.94 The corresponding he-to-background ratios, with optimum concentrations of the arc solutions, were 50:1, 40:1, 1 9 : l and 60:1, respectively. The slight influence of the background is shown by the different sensitivity ratios of the three chromium and two aluminium lines (see next section and Table IV). For aluminium a slightly higher sensitivity was found for the 396.1 nm line than for the 309.2 nm line with the arc as source. With the hollow-cathode lamp as source, the reverse was found, hence the high sensitivity ratio for the 396.1 nm line. Because of the high sensitivity ratios for those elements with resonance lines in the high background regions, it can be accepted that the background emission is so low as to have practically no effect on absomtion measurements. and that the Kranz arc can be applied successfully as a primary ligk source for most elements.TABLE IV RESULTS OF SENSITIVITY COMPARISON TESTS Element Aluminium . . Silver . . .. Barium .. Calcium . . Cadmium .. Cobalt .. Chromium . . Copper .. Dysprosium . . Erbium .. Iron . . .. Magnesium . . Molybdenum , . Nickel . . .. Lead .. .. Scandium .. Strontium . . Yttrium .. Zinc . . .. .. .. .. .. .. .. .. . . . . .. .. .. .. .. .. .. .. .. .. Line 309.2 396- 1 328-0 563.5 422.7 228.8 240.7 367.0 369.4 360.6 324.8 421.1 400.8 248-3 285.2 313.3 232.0 283-3 391-1 402.0 460.7 410.2 213.8 Optimum concentration of arc solution, p.p.m. 1000 500 200 500 30 300 500 200 200 200 160 1000 1000 200 30 1000 300 760 1000 1000 30 1000 200 Sensitivity ratio 0.70 0.87 067 0.71 0-63 0.46 0.68 0.65 0.68 0.70 0.72 - - 0.72 0.65 0-7 1 0.63 0.80 - - 0.68 0.60 - SENSITIVITY COMPARISON WITH HOLLOW-CATHODE LAMPS- The maximum sensitivities obtained with the Kranz arc as primary source were compared with the sensitivities obtainable with hollow-cathode lamps.The hollow-cathode lamps were always run with currents giving maximum absorption sensitivity coupled with sufficient light intensity, these values ranging from 5 to 15mA. Table IV shows the results for several arbitrarily chosen elements. For the elements dysprosium, erbium, aluminium, molybdenum, scandium and yttrium, a 5-cm nitrous oxide - acetylene flame was used as the absorbing medium, while the remaining elements were investigated with a 10-cm air - acetylene flame.For dysprosium, erbium, scandium and yttrium no hollow-cathode lamps were available and, instead of the sensitivity ratios, the detection limits obtained are given (defined as the con- centration giving an absorption signal equal to twice the standard deviation of the background value). Attempts to obtain absorption signals for samarium and gadolinium were unsuccessful. The samarium resonance line at 429-6 nm was not sufficiently excited to be detected, even with 5000 p.p.m. of samarium in the arc. The gadolinium resonance line at 368-4 nm could easily be detected, but even 1000 p.p.m. of gadolinium in the nitrous oxide - acetylene flame gave no absorption signal. DETECTION LIMITS- As shown under Reproducibility it is clear that, if a proper reference line is used, the reproducibility of measurement with the Kranz arc as source is just as good as that with a hollow-cathode lamp.Thus, the same ratios listed in Table IV for the relative sensitivities are applicable when considering the relative detection limits.February, 19691 87 EXTENSION OF WORKING RANGE- A disadvantage of the atomic-absorption technique is the relatively narrow working range. With a Kranz arc as source, the sensitivity can easily be decreased by using solution con- centrations other than the optimum in the arc. Fig. 7 shows that the working range for magnesium can be extended to the lower sensitivity side by a factor of 8, by increasing the concentration of the arc solution from 30 to lo00 p.p.m.This procedure, however, inevitably results in a working graph with greater curvature because of self-absorption of the emission line. ANALYSES WITH A GAS-STABILISED ARC AS PRIMARY LIGHT SOURCE 2 4 6 8 10 Concentration of absorbing solution, p.p.m. Fig. 7. Working graphs: A, 30 p.p.m. of magnesium in arc; B, 300 p.p.m. of magnesium in arc; C, I000 p.p.m. of magnesium in arc APPLICATION TO METAL ANALYSIS Lead, nickel and iron were determined simultaneously in the brass sample NBS 124b, with the Kranz arc as a primary source. With the optimum concentrations of the three elements in the arc, a 0.030 per cent. w/v solution of the sample gave absorption signals on the linear part of the working graph, and the concentration values in Table V were obtained.TABLE V ANALYSIS OF BRASS SAMPLES Results obtained by using Sample Element present technique, Chemical values, per cent. per cent. NBS 124b .. . . Lead 4.53 4.64 Nickel 0-74 0 7 6 Iron 0.28 0.26 BCS 183 .. .. Lead Zinc 1.80 1.82 1.83 1.86 In the simultaneous determination of lead and zinc in sample BCS 183 (brass), it was found that a 0.015 per cent. w/v solution of the sample absorbed 6 per cent. of the lead resonance line and 53 per cent. of the zinc resonance line, with optimum concentration values in the primary source. The latter absorption signal was considered to be too high for good accuracy. The zinc concentration in the primary source was, therefore, increased from 200 p.p.m. to 2400 p.p.m., resulting in an absorption signal of only 12.2 per cent.In this way the values in Table V were obtained for this sample. For both of these determinations the lead 405.8nm line wits used as reference, and a reproducibility test was carried out on the sample NBS 124b. The standard deviations (obtained from twenty-six measurements) of the percentage absorption values were 0-54 per cent. absorption for iron, 0.83 per cent. absorption for nickel and 0.27 per cent. for lead. These values indicate that the lead 405-8 nm line is excellent for the lead resonance line, but is less suitable for the iron resonance line and is even worse for the nickel resonance line. This low standard deviation of the lead percentage absorption values transforms to a relative88 HUMAN, BUTLER AND STRASHEIM standard deviation (of the concentration values) of 1.35 per cent.at an absorption level of 20 per cent., which was the actual absorption signal that the sample gave. A minimum relative standard deviation of 1.22 per cent. was obtained at about 40 per cent. absorption. DISCUSSION Of the fourteen elements for which the Kranz arc sensitivities were compared with sensi- tivities obtained with hollow-cathode lamps as primary sources, the mean sensitivity ratio is 0.66 (the secondary lines of aluminium and chromium are not considered). Rann9 showed that, when a plasma at atmospheric pressure is used as primary light source, the sensitivity depends on the value of the damping constants of the resonance lines in the primary source and in the absorbing medium, as well as on the degree of sklf-absorption in the primary source.Thus, to be able to predict the relative sensitivity expected for a particular element, these parameters need to be known. The reason for the optimum value of concentration of the solution in the primary source (see Fig. 3) is that a compromise is achieved between decreasing line-to-background ratios for lower concentrations, and increasing self-absorption at higher concentrations. Although the emissivity of the arc in the higher regions (h, and h3) is less, the same absorption signal can be obtained by increasing the concentration of the solution sprayed into the arc. This indicates that the temperature and arc region are not critical for obtaining maximum sensi- tivity, but that the concentration of the element is the only critical parameter.CONCLUSION The Kranz arc can be used most successfully as a primary light source for atomic-absorp- tion spectroscopy. The advantages of using such a source are the possibility of determining several elements simultaneously; only solutions of the desired element combinations are required; and the working range can be extended to lower sensitivities. In comparison with hollow-cathode lamps, the arc source has the disadvantage of having poorer sensitivities and more curved working graphs. However, the reproducibility of measurement, when using a suitable reference line, is just as good as with hollow-cathode lamps as primary sources. The authors thank Dr. K. Laqua of the Institut fur Spektrochemie und angewandte Spektroskopie, Dortmund, Germany, for useful suggestions and assistance with details of the Kranz arc source. 1. 2. 3. 4. 6. 6. 7. 8. 9. REFERENCES Massmann, H., 2. InstrumKde, 1963, 71, 225. Butler, L. R. P., and Strasheim, A., Spectvochim Acta, 1965, 21. 1207. Fassel, V. A., Massotti; V. G., Grossman, W. E. L., and ICniseley, R. N., Ibid., 1966, 22, 347. Strasheim, A., and Human, H. G. C., Ibid., 1968,23B, 266. Kranz, E., in Ritschl, R., and Holdt, G., Editors, “Emissionspektroskopie,” Akademie-Verlag, Kranz, E., in “XI1 Colloquium Spectroscopicurn Internationale, Exeter,” Hilger and Watts Ltd., Hughes, H. K., Appl. Opt., 1963, 2, 937. Pearse, R. W. B., and Gaydon, A. G., “The Identification of Molecular Spectra,” Chapman and Rann, C. S., Specfvochim. Acta, 1968, 23B, 245. Received May 23vd, 1968 Berlin, €964, p. 160. London, 1966, p. 674. Hall, London, 1963.

ISSN:0003-2654    

DOI:10.1039/AN9699400081

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, February, 1969, Vol. 94, +@. 89-93 89 A Comprehensive Scheme for the Analysis of Cement by Atomic-absorption Spectrophotometry* B Y J. T. H. ROOS AND W. J. PRICE (Pye Unicam Ltd., York Street, Cambridge) Atomic-absorption spectrophotometry has been applied to the deter- mination of aluminium, calcium, iron, magnesium, manganese, potassium, silicon, sodium, strontium and zinc in cement. Only one sample weighing is necessary, and the results, which agree well with standard values obtained by classical methods of analysis, can be obtained within a few hours. METHODS currently used for the analysis of cement samples1 frequently involve lengthy procedures and tedious separations. Sample preparation normally involves hydrochloric acid attack followed by dehydration of the precipitated silica, which is then separated by filtration.The determination of individual elements is usually performed gravimetrically or titri- metrically, although emission-flame photometry has been used for several elements.2g5 Takeuchi and Suzuki4 have determined sodium, potassium, magnesium, manganese and calcium in cement by atomic-absorption spectrophotometry, reporting good precision for all but calcium. Capacho-Delgado and Manning6 have described the determination of several elements, including aluminium, silicon and titanium. The determination of aluminium, calcium, iron, magnesium, manganese, potassium and sodium has also been reported by Crow, Hinie and Connolly.6 To compensate for possible matrix effects these authors used standard cement samples for the preparation of calibration standards.Price and ROOST have described the determination of silicon in cement by atomic-absorp- tion spectrophotometry after dissolution of the sample in a mixture of hydrochloric and hydrofluoric acids. They reported enhancement of the silicon absorption by calcium, aluminium and iron in the samples. Addition of vanadium to both standard solutions and samples compensated for this effect. EXPERIMENTAL APPARATUS- The present work was carried out with a Unicam SP90 atomic-absorption spectrophoto- meter equipped with a recorder, lamp turret and nitrous oxide flame accessories, and a standard set of interchangeable stainless-steel burner heads. Hollow-cathode lamps were obtained from Pye Unicam Limited, Cambridge. High spectral output lamps were used for aluminium, silicon and zinc.Polythene or PTFE apparatus was used whenever solutions contained uncomplexed hydrofluoric acid. REAGENTS- Aluminium solutiof6, 2 per cent. AP+-A 20.0-g sample of pure aluminium metal was dissolved in the minimum amount of hydrochloric acid (1 + l), with heating, and diluted to 1 litre. * Paper presented at the Second SAC Conference 1968, Nottingham. 0 SAC and the authors.90 [Analyst, Vol. 94 Lanthanzcm chloride solution, 5 per cent. La3+-A 58.6-g sample of lanthanum oxide was dissolved in 800ml of 20 per cent. hydrochloric acid, with heating, and diluted to 1 litre. Vanadium chzloride solution, 2-5 per cent.-A 25.0-g sample of vanadium trichloride was dissolved in 500 ml of water containing 20 ml of analytical-reagent grade hydrochloric acid (sp.gr.1.16); the solution was filtered and diluted to 1 litre. Boric acid solzction, 4 per cent.-A 40-g sample of analytical-reagent grade boric acid was dissolved in 800 ml of water, with heating, and the solution made up to 1 litre. Stock solzctiolzs, lo00 p.P.m.-Stock solutions of aluminium, iron, manganese, magnesium and zinc were prepared by dissolving 1WOg of the pure metal in the minimum volume of hydrochloric acid (1 + 1) and making up to 1 litre. Calcium, potassium, sodium and strontium stock solutions were prepared from analytical-reagent grade salts, while the stock silicon solution was prepared from sodium silicate, and standardised gravimetrically for silica. For the preparation of calibration graphs, the stock solutions were further diluted as required.All solutions were made up with de-ionised water and stored in polythene bottles, and all acids used were of analytical-reagent grade. INTERFERENCES- The effects of the other constituents of cement on the determination of a particular element were investigated over those concentration ranges likely to occur in cement samples. As expected, silicon was found to suppress the absorption by aluminium, iron, calcium, magnesium and zinc. Addition of vanadium (to give an over-all concentration of 0.5 per cent. of vanadium chloride) was found to overcome the effect of silicon on aluminium, iron and zinc; addition of lanthanum (to give a final concentration of 0.6 per cent. of La3+) instead of vanadium overcame the silicon interference on all the above elements.Similarly, interference of aluminium on calcium and magnesium was eliminated by the addition of lanthanum. With the nitrous oxide-acetylene flame the only interference observed in the deter- mination of strontium was from calcium, which caused considerable enhancement of the strontium absorption. The addition of lanthanum to both standards and samples fully com- pensated for this effect, and also for the enhancement, in an air - acetylene flame, of potassium absorption by calcium. It was also found that the presence of lanthanum compensated for the enhancement of silicon absorption by calcium, aluminium and iron in the samples. Hydrofluoric acid in the amounts used in this investigation were not found to interfere in any of the determinations.ROOS AND PRICE: A COMPREHENSIVE SCHEME FOR THE ANALYSIS DEVELOPMENT OF THE METHOD- Addition of lanthanum chloride solution to a solution containing even a small concen- tration (0.5 per cent. v/v) of hydrofluoric acid causes precipitation of insoluble lanthanum fluoride. This precipitate is not re-dissolved, even in the presence of relatively high concen- trations of nitric or hydrochloric acids. It was, therefore, necessary to complex the excess of hydrofluoric acid (after dissolution of the sample) prior to addition of lanthanum. Aluminium and beryllium are reported8 to be the most effective complexing agents for higher concen- trations of hydrofluoric acid. As the use of beryllium in routine analysis is precluded by its toxicity and, as aluminium is normally determined in cement, the possible use of boric acid was investigated.Addition of 2 g of boric acid to 1 ml of hydrofluoric acid (40 per cent.) in 50 ml of water was found to prevent precipitation of lanthanum fluoride when 20 ml of lanthanum solution were added. The presence of boric acid effectively prevented hydro- fluoric acid attack of glassware by such solutions. The presence of boric acid was not found to influence the absorption of light by any of the elements determined in cement; nevertheless, it is recommended that boric acid aIso be added to the standard solutions to compensate for any impurities in the boric acid solution. With a 10-cm air - acetylene flame, the most suitable concentration ranges for calcium and magnesium are 3 to 30 p.p.m.and 0.2 to 2-5 p.p.m., respectively. These concentration ranges require a 50 or 100-fold dilution of the original sample solution as described under Procedure. With an emission burner head giving an absorption path of 1 cm, however, the method becomes less sensitive for both calcium and magnesium by a factor of 5 to 10, thus making it possible to work with much higher concentrations of these elements. A 10-fold dilution only of the original sample solution is then required, leading to both an increase in precision and a saving in time.February, 19691 OF CEMENT BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 91 PROCEDURE STANDARDS- respective stock solutions. thanum (La3+) solution. solution. cent. lanthanum solution. solution. thanum solution. lanthanum solution.solution. hydrochloric acid (sp.gr. 1-16) per 1OOml of solution. lanthanum solution. cent. lanthanum solution. lanthanum soh tion. acid (sp.gr. 1-16) per 100 ml of solution. SAMPLE PREPARATION- Weigh 0.500 g of the powdered sample into a 100-ml polythene or PTFE beaker. Wash down the sides of the beaker with about 20 ml of water. Add, with stirring, 10 ml of hydro- chloric acid (sp. gr. 1-18), breaking up any gritty particles with the end of the stirring rod. When the sample has dissolved (except for any precipitated silica) rinse down the glass rod, remove it from the beaker and add 1.0 ml of hydrofluoric acid (40 per cent.). Carefully swirl the mixture until all precipitated silica is dissolved, then add 50ml of boric acid solution and mix thoroughly.Transfer the solution quantitatively to a 200-ml calibrated flask, add 20 ml of stock lanthanum chloride solution and dilute to the mark with water (solution A). Transfer 10.0 ml of solution A to a 100-ml calibrated flask, add 5 ml of hydrochloric acid and 9.0 ml of stock lanthanum chloride solution, and dilute to the mark with water (solution B). Solution A is used for the determination of aluminium, iron, manganese, silicon, sodium, strontium and zinc, while solution B is used for the determination of calcium, magnesium and potassium. The instrumental conditions for these determinations are given in Table I. Prepare the following series of calibration standards by appropriate dilution of the Aluminium solutions-Equivalent to 0 to 125 p.p.m. of alumina in 0.5 per cent.lan- Iron solutions-Equivalent to 0 to 125 p.p.m. of iron(II1) oxide in 0-5 per cent. lanthanum Manganese solutions-Equivalent to 0 to 2.5 p.p.m. of manganese(II1) oxide in 0.5 per Silicon solutions-Equivalent to 0 to 750 p.p.m. of silica in 0.5 per cent. lanthanum Sodium solutiom-Equivalent to 0 to 12.5 p.p.m. of sodium oxide in 0.5 per cent. lan- Strontium solutions-Equivalent to 0 to 12.5 p.p.m. of strontium oxide in 0-5 per cent. Zinc solutiorrzs-Equivalent to 0 to 0.5 p.p.m. of zinc oxide in 0.5 per cent. lanthanum Each member of the above series should also contain 25ml of boric acid and 5ml of Calcium solutions-Equivalent to 0 to 200 p.p.m. of calcium oxide in 0-5 per cent. Magnesizlm solzitiorts-Equivalent to 0 to 12.5 p.p.m. of magnesium oxide in 0-5 per Potassium solutions-Equivalent to 0 to 2.5 p.p.m.of potassium oxide in 0.5 per cent. These solutions should also contain 2-5 ml of boric acid solution and 5 ml of hydrochloric Element Aluminium Calcium Iron Magnesium Manganese Potassium Silicon Sodium Strontium Zinc Wave- length, nm 309.3 422-7 248.3 285.2 279.5 766.5 251.6 589.0 460.7 213.9 TABLE I INSTRUMENTAL CONDITIONS FOR TEN ELEMENTS Slit width, mm 0.10 0.08 0.10 0.08 0.15 0.15 0.10 0.10 0.10 0.10 Burner 5-cm N,O - C,H, Air - CaHa emission 10-cm air - C,H, Air - C,H, emission 10-cm air - C,H, 10-cm air - C,H, 10-cm air - C,H, 10-cm air - C,H, 5-cm N,O - C,H, 5-cm N,O - C,H, Acetylene flow-rate, Oxidant, litres per litres per minute minute 4-0 Nitrous oxide, 5-0 1.4 Air, 5.0 1.4 Air, 5-0 1.4 Air, 5.0 1.4 Air, 5.0 1-0 Air, 5.0 4-3 Nitrous oxide, 5-0 1.0 Air, 5.0 4.0 Nitrous oxide, 5.0 1-0 Air, 5.0 Burner height, cm 1.0 2.0 1.0 2.0 1-0 1.0 1-0 1.0 1.0 1.0 Lamp current, mA 10 12 15 4 12 12 15 12 12 1292 ROOS AND PRICE : A COMPREHENSIVE SCHEME FOR THE ANALYSIS [ArtdySt, VOl.94 RESULTS The results obtained for the analysis of standard cement samples by the proposed method are given in Table 11. Also included in this table are figures for the precision (expressed as the coefficient of variation) for the determination of seven of the elements. These were calculated from the results obtained for six replicate analyses of the same standard cement sample (N.B.S. 1015). TABLE I1 ANALYSIS OF STANDARD SAMPLES Constituent Alumina, A1,0, . . Calcium oxide, CaO .. Iron(II1) oxide, Fe,O, Magnesium oxide, MgO Manganese oxide, Mn,O, Potassium oxide, K,O Sodium oxide, Na,O Strontium oxide, SrO Zinc oxide, ZnO . . Silica, SiO, . . .. N.B.S. 1013 Atomic' Certificate absorption . . 3.30 3.18 3.21 . . 64.26 64.6 63.9 . . 3-07* 3-21 3-17 . . 1-39 1-37 1-37 . . 0.05 0-052 0.052 . . 0-32 0.38 0.38 . . 24.2 24.7 23.7 . . 0.20 0-19 0.19 . . 0.08 0.080 0.085 .. - 0-008 0.008 N.B.S. 1018 Atomicl Certificate absorption 61-37 60.4 59.5 5.04 4-99 5-06 3-27* 3.36 3.38 4.25 4-35 4.24 0.06 0.059 0-059 0.87 0.85 0.85 0.16 0.17 0.16 0.11 0.095 0.098 - 0.011 0.011 20.6 21.0 20.3 Coefficient of variation, per cent. 1.0 0.7 0.7 0-7 1.3 1.2 0.9 - - - * Capacho-Delgado and Manning6 found 3.17 and 3-37 per cent. of Fe,O, for N.B.S. 1013 and N.B.S.1016, respectively. The complete analysis of five samples was performed in about 1 normal working day made up as follows. Dissolution of five samples required 14 hours; the preparation of dilutions for atomic-absorption analysis 4 hour; the setting up of the instrument took 5 minutes; the measurement of five samples for one element 5 minutes and the measurement of six standards for one element 6 minutes, thus the time required for ten elements was about 3 hours; and the preparation of calibration graphs and the calculation of results required 34- hours. The total time, therefore, required was 8 hours. No significant increase in the time required for analysis would be caused by an increase in the number of samples analysed. DISCUSSION Suppression of the absorption of several elements in the flame by silicon (as silicate ions) is a well established phenomenon, having been noted by several a u t h o r ~ .~ ~ ~ O ~ ~ ~ However, the suppression of aluminium absorption appears not to have been noted previously. This effect operates only if, in addition to aluminium and silicon, a third element such as calcium is present. The interference is, therefore, dependent on the formation of a complex silicate, such as calcium aluminium silicate. This does not occur in the presence of an excess of vanadium or lanthanum. Although either vanadium or lanthanum chlorides can be used as releasing agents for aluminium, iron and zinc in the presence of silicate ions, lanthanum is recommended because vanadium does not adequately release calcium, magnesium and manganese from refractory silicates, whereas lanthanum does ; lanthanum fully compensates for the enhancement of the strontium absorption, by calcium in the samples, while vanadium, because of its higher ionisation energy, compensates only partially ; and vanadium salts are not readily available in a sufficiently pure form.The use of lanthanum salts as releasing agents for calcium and magnesium is frequently criticised on the ground that laboratory-reagent grade lanthanum compounds may contain significant amounts of magnesium and calcium impurities, resulting in large values for the reagent blanks. However, by using the shorter path-length emission burner with consequent reduction in sensitivity, the calcium and magnesium concentrations can be increased by a factor of five or ten over those normally used in the long-yath burner.However, the same amount of lanthanum is found to give complete releasing action, and the readings for the blank solutions are proportionately reduced to acceptable limits.February, 19691 OF CEMENT BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 93 14tomic absorption has proved to be an extremely rapid and versatile method for the analysis of cement. Not only is it possible to determine ten elements after one sample weigh- ing, and without any chemical separations, but the reproducibility is comparable with that expected for the accepted methods of analysis. The only possible exceptions are the deter- minations of the major constituents, silicon and calcium. For example, a standard deviation of 0-7 per cent. was obtained for calcium, for which standard gravimetric procedures are usually expected to give nearer 0.2 per cent. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES “Analysis of Calcareous Materials,” S.C.I. Monograph No. 18, Society of Chemical Industry, Diamond, J. J., Analyt. Chenz., 1956.28, 328. Schuhknecht, W., and Schinkel, H., Ibid., 1963, 36, 161. Takeuchi, T., and Suzuki, M., Talanta, 1964, 11, 1391. Capacho-Delgado, L., and Manning, D. C., Analyst, 1967, 92, 553. Crow, R. F., Hime, W. G., and Connolly, J. D., J. Portland Gem. Ass. Res. Dev. Lab., 1967, 60. Price, W. J., and Roos, J. T. H., Analyst, 1968, 93, 709. Graff, P. R., and Langmyhr, F. J., Analytica Chim. Acta, 1958, 21, 429. Gidley, J. A. F., and Jones, J . T., Analyst, 1960, 85, 249. Belcher, C. B., and Kinson, K., Analytica Chim. Acta, 1964, 30, 483. Slavin, W., Sprague, S., and Manning, D. C., Atomic Absorption Newsletter, 1963, 15, 1. London, 1964. Received August 26th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400089

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

94 Analyst, February, 1969, Vol. 94, pp. 94-104 Separate and Simultaneous Determination of Zirconium and Hafnium in Nickel-base Alloys with Xylenol Orange” BY H. J. G. CHALLIS (International Nickel Limited, Wiggin Street, Birmingham 16) The use of xylenol orange as a spectrophotometric reagent for zirconium and hafnium has been investigated for their determination in the range 0.002 to 0.2 per cent. in complex nickel-base alloys, and the effects of major alloying elements, and likely impurities, have been studied. A simple, direct procedure, based on the formation of the red xylenol orange complexes in 0.8 N hydrochloric acid, has been successfully applied to the determination of either metal in nickel alloys containing chromium, cobalt, iron, molyb- denum, titanium and aluminium.For alloys containing both zirconium and hafnium, a procedure de- veloped for their simultaneous determination is based on the relative effect of acid concentration on the xylenol orange complexes. Preliminary mercury cathode and hydroxide separations are followed by measurement of the total optical densities at three levels of acidity, 0-35, 1.12 and 2.0 N per- chloric acid. This “three-point” method has proved satisfactory with syn- thetic alloy solutions and when applied to complex nickel alloys containing both zirconium and hafnium. Confirmatory evidence of the results was obtained by X-ray fluorescence, emission and mass spectrometry. The simultaneous procedure provides a simple and sensitive chemical method of differentiating between microgram amounts of zirconium and hafnium, and should be capable of wider application to other alloy systems.The simpler direct method should also prove advantageous when mutual interference does not arise. ADDITION of minor amounts of zirconium or hafnium, or both, beneficially affects the mechanical properties, weldability and mechanical working of various nickel-base alloys. However, because of their close chemical similarity, the separate determination of these two elements in alloys containing aluminium, chromium, cobalt, iron, manganese, molybdenum and titanium presented a difficult analytical problem. Consideration of the numerous spectrophotometric reagents available for zirconium and hafnium112 indicated that xylenol orange and arsenazo I11 offered distinct advantages in high sensitivity and selectivity.As arsenazo 111 was not available for the initial tests, xylenol orange was chosen for detailed examination, particularly as exploratory tests indicated that this reagent not only offered the possibility of a direct method for both elements, but also their simultaneous determination appeared feasible. Following lengthy separations, Chen@*s5 used xylenol orange to determine zirconium or hafnium in high-temperature alloys, but his attempt to determine hafnium in the presence of zirconium by control of acidity was unsuccessful. In the investigation reported below, initial experiments were concerned with the de- velopment of simple, rapid and direct methods for the separate determination of each of the elements and later, after studies of the effect of acidity and alloying metals on colour reaction, methods were established covering the range 0.002 to 0.2 per cent.of zirconium or hafnium in complex nickel alloys. However, the problem of mutual interference remained. On a practical basis, alloys in which zirconium alone is used could safely be assumed to be hafnium free, but the converse was not true, because commercial grades of hafnium contain 2 t o 3 per cent. of zirconium. These initial experiments, together with the information available in the literature, formed a useful background to the development of a procedure for the simultaneous determination of each element in the presence of one another. * Paper presented at the Second SAC Conference 1968, Nottingham. 0 SAC and the author.CHALLIS 95 EXPERIMENTAL APPARATUS- made at 535nm, with 1-cm cells.A Unicam SP600 spectrophotometer was used for all optical-density measurements SEPARATE DETERMINATION OF ZIRCONIUM AND HAFNIUM- Efect of acidity-The effect of acidity on 50 pg of zirconium or hafnium was studied by preparing solutions containing varying amounts of hydrochloric or perchloric acids (0-3 to 2 N) per 25 ml and 2 ml of xylenol orange solution (0.05 per cent.). The results (Fig. 1) confirmed the optimum acidity for zirconium to be 0-8 N and that for hafnium 0.35 N. Careful control of acidity was obviously essential, for although the zirconium and hafnium complexes were only slightly affected between 0.7 and 0-9 N, and 0.3 and 0.4 N, respectively, the reagent blank values were high, with optical-density readings of 0.10 at 0.35 N and 0-14 at 0.8 N.These values were, however, stable and reproducible. In subsequent tests, a preliminary acid determination was carried out on an aliquot of each solution by tittation, in order that another aliquot could be exactly adjusted to the desired acidity before addition of the xylenol orange reagent. Perchloric acid, N Fig. 1. Effect of perchloric acid on zirconium - xylenol orange and hafnium - xylenol orange com- plexes. Curve A, 50 pg of zirconium; curve B, 40 pg of zirconium; curve C, 50 pg of hafnium; and curve D, reagent blank Calibration graphs-Based on Cheng’s re corn mend at ion^,^^^ satisfactory calibration graphs were constructed for zirconium up to 50 pg per 25 ml in 043 N hydrochloric or perchloric acids, and for hafnium up to 100 pg per 25 ml in the same acids at 0.35 N concentration.The graphs for zirconium were almost identical in the two acids and complied with the Beer - Lambert law; similarly, graphs for hafnium at 0.35 N concentration were equally satisfactory. The similarity of graphs obtained in hydrochloric and in perchloiic acids was convenient, as the use of inixed acids was made practical.96 CHALLIS : SEPARATE AND SIMULTANEOUS DETERMINATION OF ZIRCONIUM [Arta@St, VOl. 94 Efect of other metals-The effect of nickel and of other metals likely to occur in nickel - chromium high-temperature alloys was studied by analysis of synthetic solutions of Specpure (Johnson, Matthey Limited) metals in 0-8 N hydrochloric acid. After correcting for any background colour, metals in the amounts shown in Table I were found to have no apparent TABLE I EFFECT OF NICKEL AND OTHER METALS ON THE DIRECT DETERMINATION OF ZIRCONIUM Per cent.Zirconium (calculated Zirconium found Metals Nickel . . .. .. .. Aluminium . . .. .. Cerium .. .. .. .. Cobalt . . .. .. .. Chromium . . .. .. Copper . . .. .. .. Nickel - chromium . . .. Nickel - chromium - cobalt . . Manganese . . .. .. Magnesium . . . . .. Thorium .. .. .. Yttrium. . .. .. .. Titanium . . .. .. Weight, g 0.05 0.05 0.003 0.02 0.02 0.05 0-05 0-05 0.05 0.05 0-004 0.02 0.05 on 0-OS-g sample) 100 100 6 40 40 100 80 nickel - 20 chromium 60 nickel - 20 chromium - 20 cobalt 100 100 8 40 100 added, ccg 10 20 50 10 10 10* - - - - - - 10 10 10 - - 10 10 10 10 10 60 - - - - (apparent), Pg N.D.(<0-6) 10.2 20.3 51 N.D. (<0.5) 10, 10.2 N.D. (<0-5) 9.8, 10.1 9-8 N.D. N.D. N.D. (<0*5) 10.0, 9-8 N.D. 10.2 9.8, 10.2 N.D. 10.0, 10.1 N.D. 9.9, 9-8, 10.2 S.D. 9-9, 10.3 N.D. 10.0, 10.5 N.D. (<0-5) 9.8, 9.9 49.0 * After fuming with perchloric acid and reduction with hydrochloric acid. N. D .-Not determined. effect on xylenol orange or on the determination of added zirconium (10 pg). Iron and molybdenum, however, produced slight positive errors, equal to an apparent zirconium content of 0.001 per cent. for each 1 per cent. of iron, and 0.001 per cent. for 10 per cent. of molybdenum (Table 11). Reduction of iron by ascorbic acid was not entirely satisfactory, but with tin(11) TABLE I1 INTERFERENCE OF IRON AND MOLYBDENUM IN DIRECT METHOD Per cent.(calculated on Apparent zirconium, Weight, g 0.06-g sample) per cent. 0.0005 1 0.001 0.001 2 0.002 0.0026 5 0.007 0.006 10 0.016 0.05 100 0-13 0.006 10 0.001 0.01 20 0.003 0.025 60 0.006 0.05 100 0.010 Iron- Molybdenum-February, 19691 AND HAFNIUM IN NICKEL-BASE ALLOYS WITH XYLENOL ORANGE 97 chloride the error was almost eliminated, and it was thus possible to make determinations in the presence of at least 10 per cent. of iron (calculated on a 0-05-g sample). Unfortunately, tin(I1) chloride could not be used in the presence of molybdenum because a dark brown colour was produced; this colour faded rapidly. However, in the samples to be analysed, iron and molybdenum did not exceed about 5 per cent. ; it was, therefore, considered preferable to leave any iron in the iron(II1) state and apply slight corrections for both of the interfering elements.Although the alloys concerned did not contain tungsten, niobium or tantalum, it should be noted that alloying amounts of these elements can interfere because of pre- cipit at ion. In attempting to apply the direct method to hafnium two complications arose; first, the mutual interference of zirconium and hafnium ; secondly, the considerably increased interference of iron at the optimum acidity of 0.35 N, amounting to an apparent 0.2 per cent. of hafnium for each 1 per cent. of iron present. Obviously, direct determination at 0.35 N acidity would be unsatisfactory for alloys containing minor or alloying amounts of iron. Provided that zirconium was present only as an impurity arising from the hafnium present (ie., about one fiftieth of the hafnium content), a more satisfactory alternative was to sacrifice some sensitivity and determine hafnium in 0-8 N hydrochloric acid.The direct method for zirconium or hafnium was checked by analysis of synthetic nickel- base solutions with a basic composition similar to that of the commercial hirrh-temperature alloy “Nimonic”* alloy 105. Sitisfactory recoveries were obtained of 0.02 to’ 0.10 of added zirconium or hafnium (Table 111). TABLE I11 DIRECT DETERMINATION OF ZIRCONIUM OR HAFNIUM IN SYNTHETIC NICKEL-BASE Zirconium, per cent. Test & N O . Added Found No. Hafnium, per cent. Test r- Added Found 1 Nil < 0.00 1 5 Nil <0*001 2 0.020 0.020 6 0.020 0.020 3 0.040 0.040 7 0.040 0.039 4 0.100 0.102 8 0.100 0.098 * Cr 15, Co 20, Mo 6, Ti 2, A1 5, Fe 2, Mn 1 per cent.; balance Ni.per cent. ALLOY * SIMULTANEOUS DETERMINATION OF ZIRCONIUM AND HAFNIUM- The major problem of the separate determination of the two elements remained, and the final solution was based on a study of the relative effect of acid concentration on their relevant xylenol orange complexes. In an attempt to determine hafnium in the presence of zirconium, Cheng5 used the difference in optical densities in 0.6 and 1.0 N perchloric acid but obtained results 20 to 30 per cent. high, which he attributed to difficulty in control of acidity. In the present investi- gation, control of acidity by the preliminary titration technique has presented no difficulty. Further study of his acidity curves, and of those shown in Fig.1, suggested that measurement of optical densities at other levels of acidity would be more advantageous. Chene s6 also suggested an alternative procedure involving the masking of zirconium with hydrogen peroxide, but tests have shown that such a procedure would be inapplicable in the presence of titanium. Experiments were, therefore, concentrated on the effect of acidity based on the following observations. The optical density of the pure zirconium complex is the same at 0.35 N (optimum for hafnium) and 1-1 N. Consequently, in solutions containing both elements, any difference between the measurements at these two levels of acidity will be a function of the hafnium present. A value for hafnium in such solutions should, therefore, be obtainable from a calibration graph constructed from the difference between optical-density readings at 0.35 and 1.1 N.This value, converted into a corresponding optical-density reading for hafnium at 0.35 N, and subtracted from the total reading at this acid concentration, should give the optical density attributable to zirconium at 0.35 N, thus enabling the amount present to be determined. * “Nimonic” is a trade mark.98 CHALLIS : SEPARATE AND SIMULTANEOUS DETERMINATION OF ZIRCONIUM [Andyst, Vol. 94 TABLE IV SIMULTANEOUS DETERMINATION OF HAFNIUM AND ZIRCONIUM IN SYNTHETIC SOLUTIONS BY “TWO-POINT” METHOD Added , pg Recovered, p g Hafnium Z i r c o n i t m Hafnium Z i r c o ; n i u m 60 20 48.6, 62, 46 22, 20, 23 20 40 26, 22 37, 40 30 40 33-6 37.6 10 40 14 38 20 20 24, 18 19,20 Appropriate calibration graphs were constructed for hafnium -(in micrograms) plotted against optical density at 0.35 N minus optical density at 1.1 N also for zirconium at 0.35 N.The proposed method was checked by determinations of the total optical density at 0.35 and 1.1 N for synthetic solutions containing 10 to 50 pg of hafnium and 20 to 40 pg of zirconium. Results obtained by this “two-point” method (Table IV) were promising, but only approximately correct, probably because of the steep slope of the hafnium acidity curve at 1.1 N concentration. From a further consideration of the curves in Fig. 1, measurement of optical density at a third point, 2.0 N, was introduced. This full procedure was designated the “three-point” method.At that level of acidity, the optical density due to hafnium is considerably reduced, to about one tenth of that at 0.35 N, whereas readings for zirconium are about one third. Correction for hafnium at 2.0 N can, therefore, be made on the basis of the approximate content calculated from the readings at 0.35 and 1.1 N described in the “two-point” method, to provide a more accurate zirconium value. In turn, the latter value can be converted into an equivalent reading at 0.35 N which, subtracted from the total optical-density readings at this point, enables the hafnium content to be calculated at the optimum acidity. Briefly, the “three-point” method involves prior construction of calibration graphs for zirconium and hafnium at 0-35 and 2.0 N, also for hafnium readings corresponding to the optical density at 0.35 N minus optical density at 1.1 N.Zirconium and hafnium contents can then be calculated from the total optical-density readings obtained in perchloric acid solutions at 0-35, 1.1 and 2-0 N. This procedure was checked by analysis of synthetic solutions containing 20 to 50 pg of zirconium and 10 to 50 pg of hafnium. The recoveries obtained (Table V) showed a definite improvement over corresponding values calculated from “two- point” readings, and were sufficiently promising to justify application of the method to actual samples. TABLE V SIMULTANEOUS DETERMINATION OF HAFNIUM AND ZIRCONIUM IN SYNTHETIC SOLUTIONS : COMPARISON OF “TWO-POINT” AND “THREE-POINT” METHODS Recovered, pg A I -l Added, p g “Two-point” method “Three-point” method H Z i Z F Z Z Z n m Hafnium - Zirconium Hafnium r-U* 60 20 50 20 60 20 20 20 20 60 10 60 48 46 49 23.6 26 14-6 22.6 20.6 20.3 18.6 46 46 60-6 20.6 61.6 20.5 61 19 21 19-6 21 49 10 49 The experimental work on the direct method showed clearly that removal of interfering elements, particularly iron and molybdenum, would be essential.It was, therefore, decided to include mercury cathode and hydroxide separations.* s6 The full, proposed procedure was tested by addition of varying amounts of zirconium and hafnium to synthetic solutions containing the equivalent of 1 g of a nickel-base alloy, and satisfactory recoveries were obtained at the equivalent of 0.01 to 0.05 per cent. of zirconium and 0-02 to 0.20 per cent. ofFebruary, 19691 AND HAFNIUM IN NICKEL-BASE ALLOYS WITH XYLENOL ORANGE 99 hafnium (Table VI).Attempts were made to electrolyse in perchloric acid solution to avoid the hydroxide separation, which had been included to eliminate the interference by sulphates and sulphuric acid, but results were then less satisfactory. The procedures described below were, therefore, finally adopted. TABLE VI NICKEL-BASE ALLOY * BY “THREE-POINT” METHOD DETERMINATION OF HAFNIUM AND ZIRCONIUM IN SYNTHETIC SOLUTIONS OF A Added, per cent. Found, per cent. Hafnium Zirconium f A 1 Test - No. Hafnium Zirconium 1 2 3 4 5t 6t 7 t 8t 9t - 0.02 0.03 0.05 0.10 0.20 0.02 0.05 0.10 - N.D. t0.002 - 0.021, 0-022 - 0-032 - 0-052 0.0 1 0.098, 0.102, 0.100 0.02 0.20 1 0-02 0.022 0.03 0.052 0-05 0.096 0.003 0.002, 0-003 0.002 0.002 0.010, 0.011, 0.009 0.020 0.019 0.030 0.050 * Alloy composition, per cent.: Cr 20, Co 16, Ti 2.5, A1 1-5; balance Ni.t Tests 5 t o 9 corrected for Zr in alloy blank (0.003 per cent.). DETERMINATION OF ZIRCONIUM AND HAFNIUM IN NICKEL-BASE ALLOYS METHOD I : DIRECT DETERMINATION OF ZIRCONIUM OR HAFNIUM This method is suitable for either zirconium or hafnium in amounts up to about 0.2 per cent. when present alone in nickel-base alloys, substantially free from tungsten, niobium and tantalum. REAGENTS- Standard zirconium solzltion-Transfer O*lOOO g of high-purity zirconium to a platinum dish (3 inches in diameter is suitable). Add 20 ml of water, then hydrofluoric acid (40 per cent.), dropwise, until the metal is dissolved (about 0.5 ml will be required).Add 5 ml of sulphuric acid (sp.gr. 1-84), evaporate to dense fumes of sulphuric acid and fume for 10 minutes. Cool, wash the sides of the dish with water and fume again for 5 minutes. Cool, wash the sides of the dish with water and then fume almost to dryness. Cool, dissolve the salts in 20 ml of hydrochloric acid (1 + 1) and heat to boiling for 5 minutes, cool, transfer to a 100-ml calibrated flask and dilute to the mark. Dilute 5 ml of this solution to 500 ml with 0-8 N hydrochloric acid. 1 ml of final solution = 10 pg of zirconium. Stafl,dard hafnium solution-Transfer O-lOOO g of high purity hafnium to a platinum dish and proceed as described for the preparation of 100 ml of standard zirconium solution. Dilute 10 ml of this solution to 500 ml with 0.8 N hydrochloric acid.1 ml of final solution = 20 pg of hafnium. Xylenol orange solution, 0.05 per cent.-Dissolve 0*0500 g of xylenol orange in 50 ml of water and dilute to 100ml (Note 1). PREPARATION OF CALIBRATION GRAPHS FOR ZIRCONIUM AND HAFNIUM- Add 6.0 ml of standard zirconium (or hafnium) solution to a 25-ml graduated flask and, to another 25-ml flask, add 1.75 ml of hydrochloric acid (sp.gr. 1-18). Dilute each solution to the mark and check the normalities of the respective solutions by titrating 1 ml with 0.1 N sodium hydroxide (Note 2). Prepare a series of calibration solutions by the separate addition of 0.0, 1.0, 2.0, 3.0, 4.0, 5-0 and 6.0 ml of the standard zirconium (or hafnium) solution to each of seven 25-ml cali- brated flasks; add the amount of hydrochloric acid (sp.gr.1.18) calculated to produce a final acidity of 0.80 N (Note 2). Dilute the solutions to 20 ml, add 2 ml of 0.05 per cent. xylenol orange solution and dilute to the mark. Measure the optical densities of the solutions at 535 nm, with a 1-cm cell, against water in the compensating cell. Correct for the reagent blank and construct a calibration graph for optical densities against micrograms of zirconium (or hafnium).100 CHALLIS : SEPARATE AND SIMULTANEOUS DETERMINATION OF ZIRCONIUM [AIZabSt, VOl. 94 Dilute 1 ml of each solution to 10 ml and check the normaiities against 0.10 N sodium 0.01 N, the determination hydroxide. If the normality of a solution is not within 0.80 should be repeated, with appropriate correction for acid addition.PROCEDURE- Determine a reagent blank value with each batch of samples. For zirconium or hafnium contents up to 0.2 per cent., transfer a 0.5-g sample to a 150-ml beaker and dissolve in 15 ml of hydrochloric acid (sp.gr. 1.18) plus 5 ml of nitric acid (sp.gr. 1.42). Add 5 ml of perchloric acid (sp.gr. 1.54) and evaporate to fumes, with a glass cover on the beaker, until carbides are destroyed and chromium salts oxidised. Cool to room temperature, add 15 ml of hydro- chloric acid (sp.gr. 1-18), warm to dissolve salts and boil for 5 minutes, with the cover on, to reduce chromium salts. Add 10 ml of water, boil and filter through a 9-cm No. 42 Whatman filter-paper into a 50-ml calibrated flask. Wash any precipitate on to the filter-paper with water, continue washing and dilute the filtered solution to 50 ml.Examine the filter-paper for zirconium or hafnium (Note 3). Place three separate 5-ml aliquots (Note 4) of the sample solution into 25-ml calibrated flasks (A, B and C). Dilute the solution in flask A to 25ml and check for normality as described under Preparation of calibration graphs for zirconium and hafnium. To flasks B and C add the amount of hydrochloric acid (spgr. 1.18) calculated to produce final acidities of 0.80 N. To solution C add 10 ml of water and 2 ml of xylenol orange solution; dilute solutions B and C to 25ml. Measure the optical densities of solutions B and C at 535nm against water, with 1-cm cells, correct the optical density of solution C for reagent blank and back- ground (B), and then calculate the zirconium or hafnium content from the calibration graph.Make corrections for iron and molybdenum, if necessary (Note 5). NOTES- 1. Prepare a calibration graph with each new batch of xylenol orange. 2. The 1.75 ml of hydrochloric acid (sp.gr. 1-18) added should, after dilution with water to 25 ml, produce a 0.80 N solution. Subsequent acidity corrections of the calibration and sample solutions are made on the basis of these titrations. 3. Ignite the filter-paper, fuse the residue with sodium carbonate (0.5 g), dissolve the cooled melt in 10 ml of hydrochloric acid (1 + l), boil the solution for 3 minutes and make up to 25 ml. Check the acidity of a 5-ml aliquot and examine for zirconium or hafnium as described under Procedure. 4. For 0.1 to 0.2 per cent. of zirconium take a 2-ml aliquot. 5.Iron and molybdenum cause slight positive errors, which can be corrected by subtracting 0.001 per cent. of zirconium (or 0.002 per cent. of hafnium) for each 1 per cent. of iron present and 0.001 per cent. (or 0.002 per cent. of hafnium) for each 10 per cent. of molybdenum. METHOD I1 : SIMULTANEOUS DETERMINATION OF ZIRCONIUM AND HAFNIUM IN NICKEL-BASE ALLOYS This method is suitable for the determination of zirconium and hafnium in amounts up to about 0.2 per cent. when both metals are present in nickel-base alloys substantially free from tungsten, niobium and tantalum. REAGENTS- with 0.8 N perchloric acid. PREPARATION OF CALIBRATION GRAPHS- Add 6 ml of standard zirconium (or hafnium) solution to a 25-ml graduated flask and, to another similar flask, add 3.0 xnl of perchloric acid (sp.gr.1-54). Dilute each solution to the mark and determine normalities by titrating 1 ml with 0.1 N sodium hydroxide (Note 6). Based on Method I, determine the optical densities of calibration solutions containing- (a) 0 to 60 pg of zirconium in 0.35 N perchloric acid; ( b ) 0 to 60 pg of zirconium in 2.0 N perchloric acid; (c) 0 to 100 pg of hafnium in 0.35 N perchloric acid; (a) 0 to 100 pg of hafnium in 1-12 N perchloric acid; and (e) 0 to 1OOpg of hafnium in 2 . 0 ~ perchloric acid. As described in Method I, except that the final standard solutions are diluted to 500 mlFebruary, 19691 AND HAFNIUM IN NICKEL-BASE ALLOYS WITH XYLENOL ORANGE 101 Check the acidities of the solutions and correct for reagent blanks in 0.35, 1.12 and 2.0 N perchloric acid.Construct calibration graphs for solutions (a), ( b ) , (c) and (e) to give graphs 1, 2, 3 and 4, respectively. Construct a calibration graph for differences between optical densities at 0.35 N rninzts optical densities at 1.1 N against micrograms of hafnium (graph 5) * PROCEDURE- Determine a reagent blank value with each batch of samples. Transfer a 1.0-g sample to a 250-ml beaker and dissolve it in 30 ml of hydrochloric acid (sp.gr. 1.18) and 10 ml of nitric acid (sp.gr. 1.42). When dissolved, cool, add 20 ml of sulphuric acid (1 + 1) and evaporate slowly to fumes of sulphuric acid. Cool, dissolve in 40 ml of water by boiling for about 15 minutes, with a cover on the beaker. Filter the solution through a 9-cm NO.42 Whatman filter-paper and wash the beaker with water, transferring any precipitate to the paper. Ignite the filter-paper in a platinum crucible, fuse the residue with sodium carbonate (0.5 g), dissolve the melt in 10 ml of sulphuric acid (1 + 3), boil and, if clear, add the solution to the original filtrate (Note 7). Dilute the solution with water to 160ml; transfer it to a mercury-cathode cell and electrolyse at 4 amps until spot tests confirm the absence of iron and chromium. Boil the electrolyte until any yellow colour caused by titanium is discharged, then evaporate to fumes of sulphuric acid. Cool, dissolve in 40 ml of hydrochloric acid (1 + l), boil for 5 minutes, cool and adjust the pH to between 8 and 9 by addition of ammonia solution (1 + 1). Boil the solution for 5 minutes to precipitate zirconium and hafnium together with the aluminium and titanium present in the alloys (Note 8).Allow the precipitate to settle, filter it on to a 11-cm No. 42 Whatman filter-paper and wash it free from sulphates with ammonium nitrate (1 per cent. w/v). Dissolve the precipitate off the paper with four 25-ml portions of hot 0.8 N perchloric acid, collecting the solution in a 250-ml beaker, and wash the paper with water until free from perchloric acid. Reserve the filter-paper to test for zirconium and hafnium (Note 9). Evaporate the solution just to fumes of perchloric acid, cool, add 20 ml of hydrochloric acid (1 + 1) and boil for 5 minutes to ensure complete breakdown of zirconium and hafnium complexes. Add 20 ml of water and filter through a 9-cm Whatman No.42 filter-paper into a 100-ml graduated flask. Wash the beaker and paper with water and dilute to 100ml (Note 9). DETERMINATION OF ZIRCONIUM AND HAFNIUM BY “THREE-POINT” METHOD- Transfer 5-ml aliquots of blank and sample solutions to a 25-ml graduated flask, dilute to the mark and determine acidity by titration of 1 ml with 0.1 N sodium hydroxide. Add the calculated amount of perchloric acid (sp.gr. 1-54) to a further 5 ml of sample solution to produce a final acidity of 0.8 N. Add about 10 ml of water and 2 ml of xylenol orange solution (0.05 per cent.) ; dilute to 25 ml. Measure the optical density of the sample solution at 535 nm, with a 1-cm cell, against water in the corresponding cell to determine whether the total absorbance exceeds about 0.8 (Note 10).Transfer three separate 5-ml (or appropriate) aliquots of the sample and reagent blank solutions to 25-ml graduated flasks and add the amount of perchloric acid (sp.gr. 1-54) calcu- lated to produce final acidities of 0.35, 1-12 and 2 . 0 ~ in the sample and blank solutions (Note 6). Prepare three acidic solutions for determination of the xylenol orange reagent blank at 0.35, 1-12 and 2.0 N. Dilute the solutions to about 20m1, add 2ml of xylenol orange solution (0.05 per cent.) and dilute to 25ml. Measure the optical density of each solution at 535 nm, with a 1-cm cell. Finally, check the normality of the solutions by titrating 1 ml with 0.1 N sodium hydroxide solution. CALCULATION OF ZIRCONIUM AND HAFNIUM CONTENTS- Correct the optical density values at 0-35, 1.12 and 2.0 N for the respective acid and reagent blanks (Note 11).From the optical-density reading of the sample at 1-12 N, subtract that for 0.35 N, and from this difference value read off from graph 5 the approximate hafnium content in micro- grams (Note 12).102 CHALLIS : SEPARATE AND SIMULTANEOUS DETERMINATION OF ZIRCONIUM [AWZbSt, VOl. 94 B By using graph 4, convert hafnium content into the corresponding optical-density reading at 2.0 N, subtract this reading from total absorbance value at 2.0 N, and convert the difference into micrograms of zirconium by using graph 2. With graph 1, convert micrograms of zirconium into an equivalent optical-density reading at 0.35 N, subtract that value from total absorbance at 0.35 N and, from the difference, obtain the hafnium content in micrograms by using graph 3.Calculate the percentages of zirconium and hafnium. NOTES- 6. The 3.0 ml of perchloric acid (sp.gr. 1.64) added should, after dilution with water to 26 ml, produce a 1.12 N solution. Subsequent acidity corrections of the calibration, blank and sample solutions are made on the basis of these titrations. 7. If the solution is not clear (because of traces of silica), filter the precipitate on to a 7-cm No. 42 Whatman filter-paper and wash with water. Ignite the pager in a platinum crucible, cool, add one drop of sulphuric acid (sp.gr. 1.84) and 1 ml of hydrofluoric acid (40 per cent.). Evaporate to dryness; heat to 7OOOC to expel all of the acid. Fuse any residue with sodium carbonate, etc., as before, and add the extract and earlier filtrate to the main solution.8. If insufficient titanium or aluminium is present to act as co-precipitant, add a solution containing 40 mg of pure titanium or aluminium to the samples and blanks. 9. Ignite the filter-papers in platinum crucibles, fuse with sodium carbonate (0.6 g), dissolve the melt in 10 ml of hydrochloric acid (1 + l), boil for 3 minutes and make up to 60 ml. Examine the solution for zirconium pZus hafnium a t an acidity of 0.8 N. Only traces are normally detected (<0.001 per cent.) ; larger amounts of zirconium and hafnium should be determined by the “three-point” method. 10. If the total absorbance exceeds 0.8 use smaller aliquots for the determinations. 11. The acid and reagent blanks are practically identical at the three levels of acidity.12. An approximate zirconium content can be calculated by subtracting the equivalent optical- density reading for hafnium a t 0.35 N (obtained from graph 3) from total absorbance reading a t 0.36 N. The difference is caused by zirconium, and the content can be obtained from graph 1. TABLE VII DETERMINATION OF ZIRCONIUM IN NICKEL-BASE ALLOYS * Zirconium, per cent. Zirconium r A I Sample (nominal) , Separation method Alloy No. per cent. Direct method (Cheng) A 1 0.02 0.019, 0.020, 0.020 0.019 2 0.026 0.029, 0,029, 0.029 0.030 3 0.05 0.061, 0.048, 0,061 0.048 4 0.10 0.096, 0.097, 0.100 0.103 6 0.20 0.189, 0.200, 0.202 0.185 1 - 0.003, 0.004, 0.0034, 0.0032, 0.003, 0.003 0-0027, 0.0032 2 0.02 A . 0.029, 0-0296t - 3 0.06 A .0.064, 0.0636 - 4 0.10 A . 0.1036, 0.1035 - B. 0.0286, 0.028 B. 0.0666, 0.0646 B. 0*1016, 0.1016 C 1 0.06 0.067, 0-069 0-064 D 2 0.10 0.104, 0.106, 0.108 0.106, 0.100, 0.096 1 0.08 0.090, 0.090 - 2 0.06 0-066, 0.067 3 0.10 0.106, 0.106 - 4 0.12 0-126, 0-127 - 5 0.14 0.143, 0.142 - - E 1 - 0.001 0.0016 Compositions of alloys tested, per cent.- Alloy Cr Co Mo Ti A1 Fe Mn Si cu Ni 2 6 1 0.6 1 0.2 Balance A 11 20 6 B 20 16 - 2.6 1.6 As for A C 20 - - 2.3 1-3 As for A D 15 20 6 1.2 4-6 As for A E 20 - - - - - - - Balance max. m a . max. max. - t A and B separate samples tested in duplicate on aliquots from same solution.February, 19691 AND HAFNIUM IN NICKEL-BASE ALLOYS WITH XYLENOL ORANGE 103 APPLICATION OF METHODS The proposed direct Method I was applied to the analysis of complex nickel-base alloys containing chromium, cobalt, manganese, molybdenum, aluminium, titanium, copper, silicon and iron, with zirconium up to 0.2 per cent.Typical results are given in Table VII. The replicate values showed satisfactory reproducibility and close agreement with results obtained by lengthy separation methods.4,s It was interesting to note that with sample B1, to which no zirconium had been added, a mass-spectrographic test confirmed the presence of about 0.002 per cent. of zirconium compared with the average chemical value of 0.003 per cent. The accuracy of the direct method was confirmed by the determination of zirconium in a British Chemical Standard Magnesium alloy B.C.S. No. 307 containing 0.56 per cent. of zirconium, for which duplicate values of 0.55 per cent.were obtained. TABLE VIII DETERMINATION OF HAFNIUM AND ZIRCONIUM IN HAFNIUM STANDARDS (ALLOY B) Nominal hafnium, Mark per cent. B5 Nil B6 0.02 B7 0.05 B8 0.10 B9 0.20 7 Number of tests 5 4 3 3 5 Simultaneous “three-point” method Hafnium - 7 Range, Average, Number Range, per cent. per cent. of tests per cent. N.D. <0.002 N.D. t0.002 5 0.003 to 0.004 0.022 to 0.026 0.025 4 0.002 to 0.006 I A \ 2 ir c o ni u m 0.050 to 0-055 0.054 2 0.002 to 0.005 0.099 to 0.102 0.100 2 0.003 to 0.005 0.161 to 0.173 0.167 4 0.004 to 0.007 Direct 1 method Average, hafnium, per cent. per cent. 0.003 N.D. <0.005 < 0.006 0.022 (0.003) < 0.005 0.050 (0.003) <0*005 0.11 (0.004) 0.005 0-16 * Direct results after applying correction for zirconium present.For the determination of hafnium in a series of nickel-base alloy samples, the full “three-point” procedure was used to ascertain the impurity levels of zirconium (Table VIII). Considering the complexity of the problems involved, replicate tests showed reasonable reproducibility up to 0.2 per cent., at which level the zirconium content was about 0.005 per cent. Confirmation of the hafnium values was obtained by mass-spectrographic tests, which also confirmed the presence of zirconium, but indicated a sensitivity factor greater than 1 (Table IX). The direct Method I was successfully applied also to this series of alloys and, after making slight corrections for the zirconium present, the results were in close agreement with those produced by the more lengthy Method I1 (Table VIII).TABLE IX COMPARISON OF CHEMICAL AND MASS-SPECTROGRAPHIC RESULTS ON HAFNIUM STANDARDS Nominal hafnium, Mark per cent. B5 Nil B6 0.02 B7 0-05 B8 0.10 B9 0.20 Hafnium, per cent. r MSS- N.D. <0*002 <0.001 0.025 0.02 0.054 0.04 0.100 0.1 0.167 0.2 Chemical spectrographic Zirconium, per cent. r Mass- <0.005 (0.003) (0-002) < 0.005 (0.003) (0.01) <0.005 (0.003) (0.01) < 0.005 (0.004) (0.02) 0.005 (0.02) Chemical spectrographic Finally, several complex nickel-base alloy samples containing added zirconium and hafnium were examined by the full Method 11, with the results given in Table X. In the absence of any alternative chemical method, critical evaluation of these results is difficult, but the values are within working limits of the nominal additions.Confirmation of the results was, however, obtained by emission and X-ray fluorescence spectrometry, by using chemically analysed standards containing either hafnium or zirconium (Table X) .104 CHALLIS TABLE X DETERMINATION OF HAFNIUM AND ZIRCONIUM IN NICKEL-BASE ALLOY D Nominal N o ~ n a l hafnium, zmonium, Mark per cent. per cent. D6 - 0.05 D7 0.01 D8 0.01 - - D9 0.10 0-06 D10 0.03 0.08 D11 0-03 0.12 Simultaneous “three- point” method Hafnium, Zirconium, per cent. per cent. - < 0.002 0-056 0.012 (5) 0.001 0.010, 0-011, 0.002, 0.002, 0.01 1 0-002 0.094, 0.098, 0.065, 0.062, 0.105 0.060 0,042, 0.040 0.081, 0.085 0.040, 0.041, 0.12, 0,125, 0.040, 0-046 0.122, 0.115 Spectro- graphic method Hafnium, per cent. <om01 0.01 1 0.01 1 0.11 0.037 0.037 X-Ray method Zirconium, per cent. 0.06 <0.005 < 0-005 0.07, 0.05 0.09, 0.09 0.13, 0.14 CONCLUSIONS It was concluded that the direct xylenol orange procedure (Method I) offered a simple and reproducible means of determining zirconium (0.002 to 0-2 per cent.) in complex nickel alloys, with the further advantage of considerable saving in time over separation methods.Direct determination of hafnium is, however, complicated by the presence of alloying amounts of iron and of zirconium present in commercial grades of hafnium. By sacrificing sensitivity and determining hafnium at an acid concentration of 0.8 N, instead of the optimum of 0.35 N, results have been obtained on alloys low in zirconium that were in good agreement with those obtained by the more lengthy simultaneous Method 11.Information obtained during development of the direct method has provided the back- ground necessary for the successful evolution of a simultaneous procedure for determination of zirconium and hafnium. Careful control of acidity is essential, but the check procedure described makes possible easy adjustment to the required normality. From analysis of synthetic solutions, the accuracy has been shown to be within 5 per cent. of the amounts present. In the absence of alternative chemical methods, the accuracy obtained on actual samples could not be ascertained, but replicate tests on alloys were reproducible to within about 5 per cent. which, in view of the difficulties involved and the levels of contents present, is considered acceptable. Confirmatory evidence of blanks and of the magnitude of the zirconium-to-hafnium ratios in the alloys was provided by X-ray fluorescence, emission and spark-source mass spectrometry. Although the methods have been developed primarily for analysis of complex nickel-base alloys, they are capable of wider application, particularly if interfering metals are absent or can be removed by mercury-cathode electrolysis. As shown in the preliminary tests, the direct method can be applied in the presence of many other base materials (not iron) when mutual interference does not arise. It has been used for the analysis of copper, titanium-base and magnesium-base samples containing up to 0.5 per cent. of zirconium, and higher values (up to 5 per cent.) have also been determined in nickel alloys, showing fairly good agreement with gravimetric results. Alloys rich in tung- sten, niobium and tantalum, however, present special problems, and further experimental work would be necessary before the method could be applied to such materials. The simultaneous procedure provides a simple chemical means of differentiating between microgram amounts of zirconium and hafnium and, by inclusion of mercury cathode and hydroxide separations, the method has been successfully applied to a series of complex nickel- base alloys. It also offers the possibility of wider application to other alloy systems. The author is indebted to International Nickel Limited for permission to publish this paper. REFERENCES 1. Elinson, S . V., and Petrov, K. I., “Analytical Chemistry of Zirconium and Hafnium,” Israel Distributed in Great Britain by the Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience Cheng, K. L., Talanta, 1969, 2, 61. - , Ibid., 1959, 2, 186. - , Ibid., 1959, 3, 81. - , Analytica Chim. Ada, 1963, 28, 41. Programme for Scientific Translations, Jerusalem, 1965. Oldbourne Press, London. Publishers Inc., New York and London, 1959, p. 966. 2. 3. 4. 5. 6. Received August 4th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400094

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Artahst, February, 1969, Vol. 94, $$. 105-109 105 An Ultraviolet Spectrophotometric Method for the Characterisation of Some Phenolic Stabilisers in Extracts of Polymer Compositions BY L. H. RUDDLE AND J. R. WILSON (Imperial Chemical Industries Limited, Plastics Division, Bessemer Road, Welwyn Garden City, Herts.) A method for extending the usefulness of ultraviolet spectrophotometry in the characterisation of some stabilisers used in polymer compositions is described. As well as the spectra of the stabilisers in ethanol and alkaline ethanolic solutions, the ultraviolet absorptions of the solutions after reaction with nickel peroxide, and after making these solutions alkaline, are also presented. These four spectra together give a more positive identification of the antioxidant. As examples of the method, three sets of spectra are given for phenolic antioxidants, the normal ultraviolet absorptions of which are identical.These spectra show marked differences by which the compounds can now be identified. ULTRAVIOLET-ABSORBING stabilisers normally added to plastic materials are divided into two classes, antioxidants and ultraviolet absorbers. Antioxidants, as the name implies, are added to polymers to hinder oxidation of the polymer chain during the heating processes in the manufacture and subsequent fabrication of a polymer composition. The antioxidants most often used are alkyl-substituted phenols, but substituted aromatic amines and other compounds, such as phosphites and thio-esters, are also used. Each polymer type usually has its own series of antioxidants, with properties suitable only for that polymer, although some antioxidants are in common use for many polymers.Antioxidants also have applications as additives in petroleum and lubricating oil, and some are permitted for use in foodstuffs. The scheme described in this paper has been used only for examination of extracts of plastic materials, especially polyolefins, and appli- cation of the procedure to oil or food may not be possible because of interference from the raw materials. Ultraviolet absorbers are used to protect the polymer from degradation by both sunlight and artificial light. They are also substituted phenols, the most common types being hydroxy- benzophenones, hydroxybenzotriazoles and salicylate esters. These compounds have wave- lengths of maximum absorption in the ultraviolet region of the spectrum between 300 and 350 nm, whereas the antioxidants absorb at shorter wavelengths, with A,,,.between 260 and 290 nm. The additives may be present at any concentration from 100 p.p.m. to 0.5 per cent., although 0.1 to 0.2 per cent. is the usual range. The determination of the stabiliser content of polymer compositions must be carried out to correlate moulding or weathering properties with additive content and, with com- petitive materials, the stabilisers must also be identified. The determination is carried out, after extraction of the additive with a suitable solvent, either by measurement of the ultraviolet absorption of the extract or, for antioxidants, by the Metcalfe and Tomlinsonl iron(II1) chloride reduction procedure.The stabiliser content can also be determined by direct ultraviolet absorption measurement of a thin film of the polymer.2 The identification of unknown stabilisers is more complicated and may involve the use of most of the physical methods of analysis. The stabiliser must first be obtained as a pure compound, usually by thin-layer chromatography.3 After elemental analysis and molecular weight determination, the fraction can be examined by colour tests4 and by measurement of the ultraviolet, infrared, nuclear magnetic resonance and mass spectra of the compound. This full treatment is required only for new stabilisers; for a characterisation of well known compounds the simplest method is by direct comparison of the ultraviolet absorption spectra 0 SAC and the authors.106 RUDDLE AND WILSON : AN ULTRAVIOLET SPECTROPHOTOMETRIC METHOD [Autalyst, Vol.94 with those of a series of known stabilisers. For some compounds this will probably be suffi- cient, but many substituted phenols have similar spectra, and for three of the most frequently used antioxidants the ultraviolet spectra are identical. Topano1 OC, Ionox 330 and Binox M (see Table I for their chemical constitution) in ethanolic solution all have A,,,. = 277 nm, with a shoulder at 282nm.2 To extend this procedure the spectra of alkaline solutions of the phenols have then been measured either directly against a solvent blank or as “difference spectra’’ measured against the neutral solution.6 This still gives almost identical spectra for the three compounds mentioned above.TABLE I WAVELENGTHS OF MAXIMUM ABSORPTION IN ETHANOL, ETHANOL - POTASSIUM HYDROXIDE, ETHANOL - NICKEL PEROXIDE AND ETHANOL - NICKEL PEROXIDE - POTASSIUM HYDROXIDE Alkaline Ethanolic ethanolic solution, solution, Trade name Chemical constitution Amax. (nm) hmsx. (nm) Topanol OC 2,6-Di-t-butyl-4-methylphenol 277 303,274,257 Binox M Bis-(3,6-di-t-butyl-4-hydroxy- 277 303,265 Ionox 330 1,3,6-Trimethyl-2,4,6-tris- 277 303,274 phenyl) methane (3,6-di-t-butyl-4-hydroxy- benzyl) benzene Nickel Absorption peroxide change Alkaline reaction after nickel reaction product, peroxide product, hmsx. (nm) reaction Amax. (nm) 340,286 x8 Absorption suppressed 428 x 30 678 336,304 x8 Nochange This paper introduces two further stages in this procedure for extending the use of ultraviolet spectrophotometry in the characterisation of these compounds. They consist in (a) measuring the ultraviolet absorption spectrum of the stabiliser solution after reaction with solid nickel peroxide and (b) re-measuring it after making the reaction products alkaline.Cooks obtained the substituted stilbene quinone after reaction of 2,6-di-t-butyl-4-methylphenol with lead dioxide, and Braithwaite and Penketh’ have used lead dioxide for the determination of Topanol OC in liquid paraffin. They obtained an absorption peak with Amax. = 420 nm. Stafford* used air for the oxidation and obtained a product with Amax. = 365 nm. Kharasch and Joshis have reported on the oxidation of bis-(3,6-di-t-butyl4hydroxy- phenyl) methane, carried out by pumping oxygen into an alkaline ethanolic solution of the phenol.They obtained a dark purple solution produced by the anion of (CH3)o C \ ,C (C H3 ) 3 (CH3)3 O P H q o H C C(CH3)a Our early work was carried out by using lead dioxide for the determination of very small amounts of Topanol OC in extracts from polythene compounds.10 However, when new batches of lead dioxide were purchased we found that they did not give quantitative recoveries of Topanol OC, and in some instances the reagent would not react at all. In 1960, twelve samples of different grades from different suppliers were examined and shown to have differing reactivities. Other higher oxides, including mercury( 11) oxide and manganese dioxide, did not give the reaction, but it was shown that nickel peroxide, which was first made com- mercially available in January, 1963, was in fact more reactive than the original lead dioxide.Nickel peroxide consists of the higher oxides of nickel, and has been used by Nagakawa, Konaka and Nakatall for the oxidation of aliphatic and aromatic alcohols and, more recently, Sugita12 has studied the reaction of some alkyl and dialkyl phenols. The reaction at room temperature of dilute solutions of Topanol OC, about 1 mg per 100m1, with lead dioxide required 20 minutes’ shaking time, whereas the reaction with nickel peroxide is complete within 2 minutes and, if continued for longer periods, the absorb- ance of the strongest band at 286 nm begins to diminish. This speed of reaction hindersFebruary, 19691 FOR THE CHARACTERISATION OF SOME PHENOLIC STABILISERS 107 any attempt to obtain reproducible results for quantitative work with Topanol OC, although the speed of reaction varies with the antioxidant concerned. A recent batch of nickel peroxide has been found to be more reactive than usual, so that any attempts at quantitative work are, therefore, made even more difficult.METHOD APPARATUS- can be used but a recording spectrophotometer is more suitable. Any ultraviolet spectrophotometer capable of measurements over the range 200 to 700 nm REAGENTS- EthamZ-A spectroscopic grade of absolute ethanol must be used. Distillers Co. Ltd. RR grade was used throughout this work. Ethmzolic eotassium hydroxide sohtim-Dissolve 10 g of analytical-reagent grade potas- sium hydroxide in 10 ml of distilled water and dilute the solution to 100 ml with absolute ethanol.Nickel @roxide-B. D .H. laboratory-reagent grade. Wavelength, nm Fig. 1. Spectra for: (a), Topanol OC; (b), Binox M; and (c), Ionox 330. Curve A, ethanolic solution; curve B, alkaline ethanolic solution (10 ml of ethanolic solution plus 1 mi of water plus 2 ml of ethanolic potassium hydroxide solution) ; curve C, ethanolic solution after reaction with nickel peroxide; and curve D, nickel peroxide reaction product made alkaline with 2 drops of ethanolic potassium hydroxide solution108 RUDDLE AND WILSON AN ULTRAVIOLET SPECTROPHOTOMETRIC METHOD [AflUbSt, VOl. 94 EXTRACTION PROCEDURE- Prepare the polymer compound containing the unknown stabiliser by cutting or grinding to small pieces, and extract about 5 g by heating under reflw for 24 hours with 50ml of ethanol.Cool the extract to room temperature and filter off the polymer. Treat a solvent blank of 50ml of ethanol identically. SPECTROPHOTOMETRY- 1. Measure the ultraviolet absorption spectrum of the extract over the range 250 to 450nm, with the solvent blank in the comparison beam. Dilute the extract, if necessary, so that the absorbance of the main band is about 0-6. 2. To 10 ml of the solution used in 1, add 1 ml of distilled water and 2 ml of ethanolic potassium hydroxide solution. Mix well and measure the ultraviolet absorption spectrum of this alkaline solution over the same range against a solvent blank similarly treated. 3. Transfer 20 ml of the solution used in 1 to a 50-ml flask, add about 0.5 g of nickel peroxide, stopper the flask and shake, with a mechanical shaker, for 5 minutes.Filter the solution and measure the absorption spectrum as before, over the range 250 to 450 nm, with a solvent blank in the comparison beam. If necessary, because of the high absorption, dilute the filtered reaction products before measurement and note the dilution used. If the solutions are coloured, also measure the absorption spectrum over the range 450 to 700nm. 4. Add 2 drops of ethanolic potassium hydroxide solution to the cell containing the solution used in 3. Mix well and re-measure the absorption spectrum over the same range as before. RESULTS The above procedures give four spectra for each extract, which are then compared with the spectra obtained by carrying out the procedures with known stabiliser solutions.The sets of spectra for the three compounds previously mentioned, Topanol OC, Binox M and Ionox 330, tbe chemical constitutions for which are given below, are shown in Fig. 1 [ ( a ) , (b) and (c)]. Table I gives the wavelengths of maximum absorption for eaeh procedure, and also some indication of the absorbance change, Le., the dilution required, after nickel peroxide reaction. The initial concentration of the stabilisers was 10 mg per 100 ml of ethanol. (cH3)3c+ C(CH3 )3 CH3 (CH3)3 c Q C ( C H 3 13 C(CH3)3 OH Binox M (cH3)3c+ Topanol OC OH lonox 330February, 19691 FOR THE CHARACTERISATION OF SOME PHENOLIC STABILISERS 109 The reaction product of Topanol OC has A,,,. = 285 nm, and is probably the inter- mediate product obtained by Cook,6 that is, 1,2-bis-(3,5-di-t-butyl-4-hydroxyphenyl) ethane.The highly coloured substituted stilbene quinone is obtained at room temperature only after reaction of much more concentrated solutions. The colour obtained on making the nickel peroxide reaction product of Binox M alkaline was similar to that obtained by Kharasch and J o ~ h i . ~ The band gave Amax. = 578 nm and was presumably caused by the anion mentioned previously. The nature of the reaction products of the other stabiliser is not known. Although only three sets of spectra are given here, it will be appreciated that a collection of such sets of spectra can be acquired for all of those antioxidants and ultraviolet absorbers encountered during day-to-day work. The components of mixtures of ultraviolet-absorbing stabilisers cannot usually be identi- fied by carrying out this nickel peroxide reaction, and such mixtures must first be separated by thin-layer chromatography.The separation of the residue on evaporation of the extract is carried out by a method similar to that previously described for the examination of plasticisers from poly(viny1 chloride) compositions.13 The fractions are scraped from the plate, the stabilisers extracted from the powder with ethanol and the nickel peroxide reaction carried out on these solutions. The high sensitivity of this ultraviolet procedure enables the identification to be made with much smaller samples than would be required for infrared examination. Mixtures of Topanol OC with Ionox 330, and of Topanol OC with a substituted benzophenone-type ultraviolet absorber, have been identified by this method. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. REFERENCES Metcalfe, K., and Tomlinson, R. F., Plastics, Lond., 1960, 25, 319. Luongo, J. P.. Appl. Spectrosc., 1965, 19, 117. Van der Neut, J. H., and Maagdenberg, A. C., Plastics, Lond., 1966, 31, 66. “The Detection and Determination of Antioxidants in Food,” Special Report No. 1, Association Soccek, J., Chemickj Prdm., 1966, 6, 348. Cook, C. D., J . Org. Chern., 1953, 18, 261. Braithwaite, B., and Penketh, G. E., Analyst, 1963, 88, 297. Stafford, C., Analyt. Chem., 1962, 34, 794. Kharasch, M. S., and Joshi, B. S., J . Org. Chem., 1957, 22, 1439. Haslam, J., and Willis, H. A., “Identification and Analysis of Plastics,” The Iliffe Group, London, Nagakawa, K., Konaka, R., and Nakata, T., J . Org. Chem., 1962, 27, 1697. Sugita, J., J . Chem. SOC. Japan, Pure Chem. Sect., 1966, 87, 603 and 607. Ruddle, L. H., Swift, S. D., Udris, J., and Arnold, P. E., in Shallis, P . W., Editor, “Proceedings of the SAC Conference, Nottingham, 1965,” W. Heffer and Sons Ltd., Cambridge, 1965, p. 224. Received June 27th, 1968 of Public Analysts, London, 1963. 1965, p. 307.

ISSN:0003-2654    

DOI:10.1039/AN9699400105

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

110 Analyst, February, 1969, Vol. 94, p p . 110-120 The Absorptiornetric Determination of Silicon in Water Part VIL* Improved Method for Determining the Total Silicon Content of High-purity Water BY H. M. WEBBER AND A. L. WILSON (Central Electricity Research Laboratories, Cleeve Road, Leatherhead, Surrey) A method is described for determining the total silicon content of high- purity waters; this method allows more precise results to be obtained than the method given in Part I11 of this series. Silicon in the water is initially concentrated on a mixture of finely ground cation and anion-exchange resins, which are then ignited and fused with sodium carbonate. The resulting melt is dissolved in water and silicate is determined absorptiometrically as the reduced /%molybdosilicic acid.The standard deviation of analytical results for 1-litre samples containing between 0 and 100 g of silica was about 3 pg of silica. Ten analyses and the necessary blant determinations can be carried out in 8 hours. MORRISON and Wilson1 have described reasons why the total silicon content of high-purity waters (e.g., boiler feed-water and make-up water) is of interest in the chemical control of modern power stations. They have described2 a method that was developed primarily for a preliminary investigation of the occurrence of “non-reactive” silicon? in make-up waters. This method consisted of evaporation of a sample to dryness, fusion of the residue with sodium carbonate and absorptiometric determination of silicate in the melt. When 20-ml samples were analysed, a standard deviation of about 0.015 p.p.m.of silica was obtained, the dominant sources of error being contamination and heterogeneous distribution of silicon in the sodium carbonate. Such precision is inadequate for the analytical control of high- purity water, which may often contain 0.02 p.p.m., or less, of silica. Accordingly, ways of obtaining better precision were considered. Three methods for converting “non-reactive” into “reactive” forms of silicon were con- sidered, namely- (a) Treatment of the sample with sodium hydroxide at high temperatures and pressures in a steel bomb.3 (b) Treatment of the sample with hydrofluoric a ~ i d . ~ ~ ~ (c) Fusion with sodium arbo on ate.^,^,^ Method (a) was rejected because it was considered too inconvenient for normal use.In addi- tion, little evidence was available for the efficiency of the technique for different forms of “non-reactive” silicon. Method ( b ) appeared most attractive because of its relative simplicity as compared with method (c), but the latter appeared2p4v5 more likely to enable all possible forms of “non-reactive” silicon to be determined. Accordingly, it was decided initially to concentrate on the fusion technique. It is possible that high-purity waters can be analysed satisfactorily after treatment by method (b). It is hoped, therefore, in subsequent work to compare results obtained by methods (b) and (c). * For details of earlier parts of this series, see reference list, p. 120. t “Non-reactive” silicon is defined as those compounds of silicon which do not react with ammonium 0 SAC and the authors.molybdate in 10 minutes under the conditions described by Webber and Wilson.6WEBBER AND WILSON 111 The main source of random error in the previously reported2 fusion technique was considered to arise from the evaporation and fusion procedures. The magnitude of errors from the fusion could be simply reduced by increasing the volume of sample evaporated. However, to obtain a worthwhile improvement, long evaporation times would be required, and would, therefore, make the method more cumbersome, as well as leading to proportionally greater danger of contamination of the samples. Thus, we sought an alternative to evapora- tion as a means of concentrating the silicon in the sample. Preliminary tests were made of the possibility of co-precipitating silicon with metal hydroxides (iron, aluminium and thorium were used), but the metal contents of the solutions obtained by dissolution of the precipitates markedly affected the precision of the absorptiometric determination. Attention was, therefore, turned to the use of ion-exchange resins.Wickbold’ has described a method in which an anion-exchange resin was used for con- centrating the silicon present in high-purity waters. The sample is first treated with hydro- fluoric acid to form fluorosilicic acid, which is strongly absorbed by the resin. This technique was rejected for the same reason that hydrofluoric acid was rejected as a means of converting “non-reactive” silicon into “reactive” forms. The approach finally adopted was to adsorb silicon on a mixture of cation and anion-exchange resins, which were then ignited.The residue was fused with sodium carbonate and the melt analysed absorptiometrically. Morrison and Wilson2 used a-molybdosilicic acid for the absorptiometric determination of silicon in the fusion melt. This compound was used in preference to the more usual ,9-molybdosilicic acid because it was thought that it would allow better precision. However, the dominant errors did not arise in the absorptiometric procedure, and the use of the /3-molybdosilicic acid is more convenient for routine analysis. It was, therefore, decided to use the absorptiometric procedure described by Webber and Wilson.6 The development of this method for total silicon and the tests made of its performance are described below. APPARATUS, REAGENTS AND TECHNIQUE- The procedure given under Method was used for the work described in this section, except when indicated otherwise.Optical-density measurements were made in 4-cm cuvettes at 810 nm with a Hilger Uvispek spectrophotometer,‘distilled water being used to fill the reference cuvet t e. Distilled water from a Manesty still was used for most of the tests, but for certain tests specially purified water was prepared by passing distilled water through a bed of “Powdex” resins (see below). Standard solutions of “reactive” silicon , polymeric silicic acid and dilute suspensions of clay were prepared as described previously.8J p2 For all of the statistical significance tests applied in this work, the 95 per cent.confidence level was used. “POWDEX” RESINS- “Powdex” is the trade name used by the Graver Water Conditioning Company (U.S.A.) in their Graver Powdex process for water purification. This process was designeds for the purification of condensate in power stations, and involves the use of a mixture of finely divided, strongly acidic cation and strongly basic anion-exchange resins (in the hydrogen and hydroxyl forms, respectively) as a pre-coat on a specially designed filter support. This mixture of finely divided resins produces a heavy flocculant type of precipitate that allows rapid filtration. The process thus provides an adsorbant with a large surface area, and yet allows a rapid flow-rate through the pre-coat. Duff and Levenduskys have reported that this process was efficient in removing all forms of silicon and, accordingly, the experiments described below were made to test whether it was suitable for analytical application.PRELIMINARY TESTS- A solution was prepared containing about 1 p.p.m. of “reactive” silica and 1-8p.p.m. of “non-reactive” silica (as bentonite clay). A 200-ml portion of this solution was stirred for 50 minutes with a mixture of 0-5g* of “Powdex” anion-exchange resin and 0.25g of * Weights of resin refer to the “as received” material containing 60 to 60 per cent. of water. EXPERIMENTAL112 WEBBER AND WILSON : ABSORPTIOMETRIC [A.nalyst, Vol. 94 cation-exchange resin in a polythene centrifuge bottle. The solution was then centrifuged for 10 minutes at 2500 r.p.m., and 20-ml aliquots of the supernatant solution were analysed for and total silicon.2 A further 100ml of the original solution were passed through a bed of mixed resins (the weights used being the same as for the previous test) supported on a Millipore filter (pore size, 0 4 p ) in a Perspex filter-stick; the flow-rate was 20ml per minute, and 20-ml aliquots of the filtrate were taken for analysis. The results of the analyses showed that in both tests the total silicon contents of the treated waters were less than 0.02 p.p.m.of silica, and the “reactive” silicon contents less than 0.005 p.p.m. of silica. The use of these resins, therefore, appeared to offer a suitable means of concentrating the silicon from the water, and the main problem would be to determine the silicon in the mixture of resins.The development of a suitable procedure is described below. DETERMINATION OF SILICON IN THE RESIN MIXTURE- E$ect of sodium su&hate-Neutralisation with sulphuric acid of the sodium carbonate used for the fusion leads to the presence of sodium sulphate in the solution for analysis. Morrison and Wilson2 reported that this salt decreased both the sensitivity and precision of the absorptiometric method when using the reduced /3-molybdosilicic acid. However, although our tests confirmed a slight loss of sensitivity, no significant decrease in precision was found when 4 ml of a 10 per cent. solution of sodium carbonate and 8 ml of 1.0 N sulphuric acid were added to standard solutions of “reactive” silicon. Accordingly, the absorptiometric method of Webber and Wilson6 was used in all of the subsequent work.E’ect o f t h e resins-The effect of any residue left after ignition of the mixture of resins was tested by igniting a large bulk of the two resins and fusing the residue with sodium carbonate. After dissolution of the melt in water, the solution was neutralised with sulphuric acid and made up to a known volume. Aliquots of this solution, each containing the equivalent of 0.4 g of anion-exchange resin, 0-08 g of cation-exchange resin, 4 ml of a 10 per cent. solution of sodium carbonate and 4 ml of 2.0 N sulphuric acid, were added to a series of solutions containing standard amounts of silicon. After diluting each solution to 100 ml, the amount of silicon present in each was determined. At the same time, standard silicon solutions containing only sodium carbonate and sulphuric acid were analysed.The results of these determinations, given in Table I, show that the resins had no marked effect on the recovery of silicon, although the optical densities of the blank determinations were increased by the resins. Subsequently, it was found that when the resins were ignited and fused under the conditions recommended in the Method, the resultant solutions were sometimes turbid and smelt of hydrogen sulphide. These effects were accompanied by the formation of a blue colour in the solution immediately the ammonium molybdate - sulphuric acid reagent was added. These observations indicated incomplete oxidation of the resins and, in an attempt to eliminate this effect, hydrogen peroxide was added to the cooled melt in the platinum TABLE I EFFECT OF RESIN RESIDUES ON THE ABSORPTIOMETRIC DETERMINATION OF SILICON Optical density* Concentration of silicon added, Solution containing p.p.m.of silica Pure solution resin residues 0 0.062 f 0.0025 0-336 f 0.0025 O * l t 0.138 f 0.004 0.238 f 0.004 O.6t 0.683 f 0.005 0.692 -& 0.0055 * Mean of four results. t Optical densities corrected for the appropriate blank. crucible; any excess of peroxide was destroyed by gentle warming. This treatment facilitated the subsequent dissolution of the melt and reduced any errors caused by the above effects to less than 0.001 p.p.m. of silica. Other tests indicated that this addition of hydrogen peroxide decreased the slope of the calibration graph by about 4 per cent.The explanationFebruary, 19691 DETERMINATION OF SILICON IN WATER. PART VII 113 of this effect is not known with certainty, but it may be caused by the depressive effectlo of residual hydrogen peroxide on the formation of the molybdosilicic acid. No measurements of residual peroxide were made, and the effect was not studied further because it is allowed for adequately by the method of calibration given under Method. Fusion procedure-The fusion procedure was similar to that used by Morrison and Wilson.% The sodium carbonate was added as a solution, rather than as a solid, so that a coating of solid carbonate would be left on the resins after evaporation; this was thought likely to reduce the possibility of losses of silicon during the ignition and fusion procedure.Addition of a solution of sodium carbonate has the added advantage that it removes the errors introduced if the solid sodium carbonate contains heterogeneously dispersed silicon. Tests of this procedure showed that the recoveries of sodium silicate added to resins in a platinum crucible varied between 98 and 102 per cent. Further tests showed that rapid ignition of the resins in a fierce flame (in preference to the gentle heating prescribed under Method) resulted in the loss of about 2 per cent. of a known amount of silicate added to the resins. The ignition procedure is, therefore, not critical. SAMPLING AND SAMPLE TREATMENT- One was to pass the sample through a small column of the mixed resins, and the other was to shake the sample with the mixed resins, with subsequent recovery of the resins by filtration.Assuming that both methods give 100 per cent. retention of silicon on the resins under ideal conditions, the choice of the method is governed by practical considerations such as sampling technique, contamination and operator time. The latter method was adopted because detailed con- sideration indicated that it is likely to be more robust. The tests carried out to prove its applicability are described under Results. Amoustt of sumfile necessary to obtain the required precisiofi-Morrison and Wilson2 obtained a standard deviation of about 0.015 p.p.m. of silica when 20-ml samples of water were analysed absorptiometncally after fusion with sodium carbonate. Thus, errors from these operations could be made negligible if a much larger initial volume of sample could be analysed.The most likely source of imprecision in the present method appeared to be the heterogeneous distribution of silicon in the ion-exchange resins. Therefore, the magnitude of this error was determined so that the necessary volume of sample for adequate precision could be calculated. For this purpose, 10 samples of resins (each containing 1 and 0.2 g of the anion and cation-exchange resins, respectively) were ignited, fused with sodium carbonate and the melts analysed. The results showed that the standard deviation of the difference between two individual results was about 2pg of silica. Thus, a 1-litre sample of water should give a standard deviation of about 0.002 p.p.m. of silica. This error was considered sufficiently small, and 1-litre samples were, therefore, used in further tests.Removal of silicon from water-A series of tests was carried out to determine the para- meters affecting the removal of “reactive” silicon, polymeric silicic acid and clay from water. These forms of silicon were chosen as being reasonably representative of the forms likely to occur in high-purity waters. Except when stated otherwise, samples were analysed exactly as described under Method. To check the amount of resin required to remove silicon from water, a solution containing about 0.04 p.p.m. of “reactive” silica and 0.15 p.p.m. of silica of both polymeric silicic acid and clay was prepared. Portions of this solution were analysed, and the “reactive” silicon and polymeric silicic acid contents in the filtrates also determined.Three different amounts of resins were used, four determinations being made for each amount; two of these four deter- minations were made by gently shaking the sample with the resins for 20 minutes, the other two being set aside for the same length of time. The mean results obtained are given in Table I1 and show that 1 g of the anion exchanger was a suitable amount, The effect of different weights of the cation exchanger was then investigated by using 1 g of anion exchanger for all of the tests. Duplicate portions of the same solution were analysed, and the mean recoveries of total silicon were 97.2, 98.9 and 99.9 per cent. for 0.1, 0.2 and 0.4 g of cation exchanger, respectively. The weight of cation exchanger was not of crucial importance, and it was decided to use 1 g of anion exchanger and 0.2 g of cation exchanger in subsequent work.Two methods were considered for the collection of silicon on the resins.114 WEBBER AND WILSON : ABSORPTIOMETRIC [Analyst, VOl. 94 TABLE I1 EFFECT OF THE AMOUNT OF ION-EXCHANGE RESINS ON THE REMOVAL OF SILICON FROM SOLUTION “Reactive” Shaking silicon time, removed, Amount of resin minutes per cent. Anion (0-2 g) + cation (0.1 g) 0 23.3 20 2.4 Anion (0.5 g) + cation (0.25 g) 0 51.1 20 89.7 Anion (1-0 g) + cation (0.5 g) 0 71.3 20 97.2 Polymeric silicic acid removed, per cent. 38.9 86.1 73.2 88.5 73.8 98.6 Clay removed,* per cent. 34.8 100.0 45-6 78.2 69.4 97.5 Total silicon removed , per cent. 36.3 82.0 57.6 82.6 70-0 97.2 * These results were obtained by deducting the amounts of “reactive” silicon and polymeric silicic acid removed from the total amount of silicon recovered from the resins.The results in Table I1 indicated the desirability of shaking the sample with the resins. To check this point, eight portions of the same solution as that used in the preceding tests were analysed, four portions being shaken for 6 minutes and the other four for 20 minutes; 10-ml aliquots of the solutions, after the fusion stage, were analysed, and the mean optical densities were 0.503 & 0.018 and 0.497 & 0.013 for 5 and 20 minutes’ shaking time, respec- tively, with an over-all mean recovery of 100 t 3 per cent. It was concluded that a shaking time of 10 minutes was satisfactory. To check the effect of the size of the resin beads, each resin was sieved to give fractions of 72 to 200,200 to 325 and >325 mesh size.Portions of each fraction were used to analyse a solution containing about 0.04 p.p.m. of “reactive” silica and 0.14 p.p.m. of silica as polymeric silicic acid. In all instances, at least 99 per cent. of the “reactive” silicon was removed, and at least 98 per cent. of the polymeric acid was removed for the two smaller mesh sizes. When the 72 to 200-mesh resins were used, about 7 per cent. of the polymeric silicic acid was not retained by the resins. The mesh size of the resins did not, therefore, appear to be critically important. APPARATUS- Evaporating hood-An inverted 8-inch polythene funnel, placed about 0.25 inch above the surface of a hot-plate, forms a suitable hood that will accommodate twelve 30-ml platinum crucibles.Pass a gentle stream of air, filtered through two Whatman No. 542 filter-papers (or suitable alternatives), through the stem of the funnel to reduce condensation on the inside of the funnel. Platinum crzddes-Platinum crucibles (30 ml) are suitable. Clean them thoroughly by repeated fusions of sodium carbonate until satisfactorily precise results are obtained when 4 ml of the 10 per cent. sodium carbonate reagent are fused and the melts analysed as described under Procedure. This treatment did not remove all traces of silicon from all of the crucibles, and it was necessary to keep certain crucibles filled with molten sodium carbonate for up to 16 hours before satisfactory results could be obtained.Store the clean crucibles by inverting them on clean filter-paper in a dust-free atmosphere, and use them only for silicon determinations. They should be handled with platinum-tipped tongs. Polythene bottles for collecting samples-These bottles should be capable of holding about 1250ml of water, and should have plastic screw-top stoppers that do not leak; the necks should be shaped so that the entire contents can be poured out. To enable samples to be collected free from contamination, a modified stopper, with inlet and outlet tubes sealed into it, can be used. These bottles are also used for blank determinations, and are suitable for storing reagents. Clean the bottles by washing them with water and then allowing them to stand overnight filled with water to which 0-5 gt of anion exchanger and 0.1 gt of cation exchanger have been added.t The weights of resin used refer throughout the remainder of this paper to resins that have been dried overnight at 110” C. METHOD Finally, wash them thoroughly with water.February, 19691 DETERMINATION OF SILICON IN WATER. PART VII 115 Polythene bottles, 8-02 ca$acity-These bottles are required for the absorptiometric deter- minations and should be thoroughly washed with water. Their suitability for use should be determined by carrying out blank determinations (absorptiometric stage only) in them. Filtration apparatus-The resins on which the silicon is concentrated are recovered from the sample by collection on a Millipore (or other suitable membrane) filter disc (47-mm diameter, 0.8-p pore size) by filtration under reduced pressure.The filter disc is supported on a porous polythene plate (0.1 inch thick) held in a Perspex holder. A diagram of the apparatus is shown in Fig. 1. 140 I Perspex holder Porous polythene disc Mil I ipore membrane Perspex holder Clamp Buchner flask Fig. 1. Apparatus used for filtration This apparatus should be cleaned initially by immersing it overnight in 5 N sodium hydroxide solution, and then washing it thoroughly with water. It should be stored in a dust-free atmosphere and rinsed with water immediately before use. REAGENTS- All reagents should be of analytical-reagent grade unless otherwise stated. Water-Distilled water from a Manesty still, and stored in polythene, was found suitable, and usually contained less than 0.005 p.p.m.of silica as “reactive” silicon. Sodium carbonate solution-Dissolve 100 g of anhydrous sodium carbonate (micro- analytical grade) in about 800ml of water. Dilute with water to 1 litre in a polythene measuring cylinder, and store in a polythene bottle. Sulphzcric acid, 2 N-Add 56 ml (+Om5 ml) of 98 per cent. sulphuric acid, cautiously, to about 800 ml of water. Allow the solution to cool, and dilute with water to 1 litre in a polythene measuring cylinder; store in a polythene bottle. Hydrogen fieroxide, 100 volumes. Acidi$ed molybdate solution-Dissolve 89 g of ammonium molybdate, (NHJ ,Mo,Oa.4H2O, in about 800 ml of water at room temperature. Add 63 ml of 98 per cent. sulphuric acid, cautiously, to 100ml of water, with stirring, and allow the mixture to cool.Add the acid to the molybdate solution, cool it to room temperature, and dilute to 1 litre with water. This solution has been found to be adequately stable for at least 6 months. Tartaric acid solution, 28 per cent. w/v-This solution has been found to be adequately stable for at least 6 months. Reducing agent solution-Dissolve 2.4 g of sodium sulphite, N+S0,.7H2O, and 0.2 g of l-amino-2-naphthol-4-sulphonic acid (purest grade available) in about 70 ml of water. Add 14g of potassium metabisulphite, shake well until dissolved, and dilute to 100ml. This reagent should be freshly prepared each week.116 WEBBER AND WILSON : ABSORPTIOMETRIC [Analyst, Vol. 94 Standard solutions of silica-Fuse 1.OOO g of pure dry silica with 5 g of anhydrous sodium carbonate in a platinum crucible at red heat.When cool, dissolve in water, and dilute to exactly 1 litre. This solution contains 1OOO p.p.m. of silica. Prepare, by dilution, a solution containing 10 p.p.m. of silica. The solutions containing 1000 p.p.m. were stable, within +O-5 per cent., for at least 2 years in polythene bottles, and those containing 10 p.p.m. Gr at least 1 year. The most suitable silica for this purpose is probably transparent Spectrosil rod (Thermal Syndicate Ltd.), which has metallic impurities of less than 1 p.p.m. and is not appreciably hygroscopic. “Powdex” resins*-Place at least 250-g amounts of the anion and cation exchangers (in the hydroxyl and hydrogen forms, respectively) separately into air-tight polythene containers.Determine their water contents by weighing 1-g samples on to watch-glasses, drying overnight at 110” C, cooling and re-weighing. The efficiency with which “reactive” silicon is retained by the resins is decreased if the anion exchanger absorbs appreciable amounts of carbon dioxide. It is essential, therefore, to store the resin in air-tight containers. PROCEDURE- Before attempting any determinations of silicon in samples, all of the apparatus should be cleaned as described above, and a series of blank determinations should be carried out to ensure that adequate precision is achieved. Sample collection-Weigh out the equivalent of 0.5 gt (&0.005 g) of anion-exchange resin and 0.1 g (+0.005 g) of cation-exchange resin, and transfer directly to a dry, clean, pre-weighed polythene sample bottle. Collect 1 litre (k50 ml) of sample into the bottle, avoiding contamination. During the filtration and fusion stages described below, care must be taken to avoid contamination of the sample by air-borne particles.The immediate area in the laboratory should be wiped clean with a damp cloth before a batch of determinations is started. Sorption and jltration stage-Shake the sample vigorously for 10 minutes (a mechanical shaker is convenient). Allow the resins to settle for at least 5 minutes, and then decant about 800 ml of the sample through the filter. Make a slurry of the resins with the remaining 200 ml of solution, and transfer it quantitatively on to the filter. Fusion stage-Place the filter disc with the resins in a 30-ml platinum crucible.Add 4 ml (40.1 ml) of sodium carbonate solution (a Perspex or polythene pipette should be used), and place the crucible on a hot-plate under the evaporation hood. Warm gently (avoiding spitting) until the contents of the crucible are dry. Place the crucible on a silica triangle and heat gently with a burner to vaporise the resin. If ignition occurs remove the burner. When all of the volatile matter has been removed, increase the heating until the sodium carbonate melts, and continue heating until a clear melt is obtained. Rotate the crucible in the flame so that the melt touches all parts of the inside of the crucible to within about 3 mm of the rim. Replace the crucible in the triangle, cover it with a platinum lid, and continue heating strongly for 1 to 2 minutes.Allow to cool. To the contents of the crucible, add 0-5 ml (+0.05 ml) of 100-volume hydrogen peroxide and, after evolution of gases has ceased, add a further 0.5 ml (k0.05 ml). Warm the solution gently until no more gas is evolved and continue warming for a further 5 minutes, taking care that the solution does not boil. Nearly fill the crucible with water (from a polythene wash- bottle), and heat gently to dissolve the melt. Allow to cool. Add 4 ml (k0.1 ml) of 2 N sulphuric acid to a clean, pre-weighed, 8-02 polythene bottle, and quantitatively transfer the contents of the crucible through a polythene funnel into the bottle; pour out the solution from the crucible in one continuous stream, and rinse the crucible in an inverted position with a jet of water from a polythene wash-bottle.Dilute the contents of the bottle with water to 100 g (k0.2 g). (It was found more convenient to dilute solutions by weight than by volume.) The solution can be left overnight, if necessary, before starting the absorptiometric stage. Re-weigh the bottle and calculate the weight of sample taken. * Obtainable from William Boby & Co. Ltd. t The absolute weight of resins used is not critical and should be calculated to the first decimal place. However, within a batch of resin, the weight of resin used should not vary by more than f l and f5 per cent. for the anion and cation exchangers, respectively.February, 19691 DETERMINATION OF SILICON IN WATER. PART VII 117 Absorptiometric stage-Transfer 25 ml (+Om2 ml) of the above solution to another clean, pre-weighed, 8-02 polythene bottle and dilute with water to 100 g (+Om2 g).If the silicon content of this aliquot is beyond the range of the calibration graph, a smaller aliquot can be taken (see Sources of error below). Add 2-5 ml (k0.1 ml) of acidified molybdate solution and mix; after 10 minutes ( k l minute), add 2.5 ml (k0.1 ml) of tartaric acid solution and mix. After a further 5 minutes (2 1 minute), add 2.0 ml (k0.1 ml) of reducing agent solution and mix. Between 20 and 60 minutes later, measure the optical density of the solution at 810 nm, in 4-cm cuvettes, against distilled water. Subtract the optical density of the blank determination (see below) from that of the sample, and read off the concentration of silicon from the calibration graph.Blank determination-Shake 1.2 litres of water with about 0-8 g of anion exchanger and 0.2 g of cation exchanger in a polythene sample bottle. Filter 1 litre (+50 ml) of this solution through a filter directly into another sample bottle containing 0.5 g (+0.002 g) of anion exchanger and 0.1 g (+0.005 g) of cation. exchanger. (It is assumed that the filtrate is silicon-free water.) Treat this water as for a sample. Preparation of calibration graph-To each of a series of polythene sample bottles, add 0.5 g ( k 0.005 g) of anion exchanger, 0.1 g ( k 0-005 g) of cation exchanger and 1 litre ( 20 ml) of water. To these bottles add 0 , 5 and 10 ml of a standard silicon solution containing 10 p.p.m. of silica; these volumes correspond to 0, 0.05 and 0.1 p.p.m.of silica, respectively. Repeat the procedure given for samples, and plot a graph of the optical densities (corrected for the blank) against concentration of silicon added. Repeat these determinations until the calibration graph is defined with the required precision. SOURCES OF ERROR- Turbidity in Jinal solutions-As described under Experimental, the addition of hydrogen peroxide reduced errors caused by turbidity to negligible proportions. It is conceivable that the magnitude of these errors depends on the precise technique used during the ignition of the resins. It is desirable, therefore, to check this effect occasionally. This can be achieved by analysing an additional 25-ml aliquot of the sample and blank exactly as described above, but adding the tartaric acid before the molybdate and reducing-agent solutions.The optical densities of these solutions will be due to any turbidity or colour, or both, resulting from incomplete destruction of the resins. If these optical densities are appreciable, they should be subtracted from the normal determinations. Size of aliquot-The calibration graph is prepared by using 25-ml aliquots of the solution obtained after the fusion stage of the procedure. If an aliquot of different volume is taken, the concentration of sodium sulphate in the final solution will differ from that in the solutions used to define the calibration graph. The effect of sodium sulphate is small, e.g., when no sodium sulphate is present the slope is about 2 per cent. greater than for the conditions recom- mended above.Correction for this effect should be made if this error is unacceptable. Other waters-The method has not been tested for waters containing high concentrations of dissolved solids, e.g., raw waters, but it seems probable that there would be incomplete sorption of silicon on the resins in the presence of large amounts of other anions. Carbon dioxide-When resins that had been exposed to the atmosphere for many months were used, it was found that “reactive” silicon was not quantitatively retained on them. This effect is thought to be caused by the sorption of carbon dioxide by the anion exchanger. However, it was shown that polymeric silicic acid and clay were retained, and it was possible to determine the “reactive” silicon in the filtrate from the resins. It is recommended, there- fore, that the filtrates from the resins should be checked occasionally for the presence of “reactive” silicon, and new batches of resin should be used whenever appreciable concentra- tions of silicon are found in the filtrate.RESULTS PRECISION- On each of five occasions, the calibration procedure was carried out with concentrations of 0, 0-02 and 0.1 p.p.m. of silica, each in triplicate. “Reactive” silicon was used in these tests, as it was thought that this form of silicon would be the most difficult of the three forms investigated to remove from solution. A summary of the results is given in Table 111.118 WEBBER AND WILSON : ABSORPTIOMETRIC [Andyst, Vol. 94 TABLE I11 PRECISION OF ANALYTICAL RESULTS Optical Standard deviation,* p.p.m. of silica Sample density of silica batch batch Totalt Water + 0.02 p.p.m.of silica . . 0.182 0.029, 0.0020 N.S. 0.0026 Water + 0.10 p.p.m. of silica . . 0.436 0.031, 0.0028 N.S. 0.0036 Absorptiometric standard density per A ’L Optical 0.01 p.p.m. Within Between Water .. .. .. .. 0.123, - 0.0020 - - (0.5 p.p.m. of silica): . . 0.666 0.033, 0.0018 N.S. 0.0018 Each batch of within and between-batch standard deviations has 10 and 4 degrees of freedom, respectively. N.S. means not statistically significant a t the 6 per cent, probability level. t The total standard deviation is the estimate of the standard deviation of any one result in any one batch. 1 One-hundred millilitres of this solution were analysed by the absorptiometric stage only.The optical density in the second column has been corrected for the appropriate blank, and the value in the third column calculated on the basis of the equivalent concentration of silicon in a l-litre sample analysed by the full procedure. On the second and fifth occasions, the filtrates were analysed for “reactive” silicon. In the filtrates from the solutions, to which either no or 0.02 p.p.m. of silica had been added, less than 0.0003 p.p.m. of silica were detected; in the solutions to which 0.1 p.p.m. of silica had been added, 0401 The total standard deviation for the 0.5 p.p.m. of silica standard (to which 1 ml of 10 per cent. sodium carbonate solution and 1 ml of 2 N sulphuric acid were added) agrees well with the value of 0.0016 p.p.m.of silica previously reported by Webber and Wilson.g 0.0003 p.p.m. of silica were detected. EFFECT OF OTHER SUBSTANCES- Three solutions containing different substances were prepared and analysed in triplicate exactly as described under Method. The tests were repeated at another concentration of “reactive” silicon. The substances added were as follows (all concentrations refer to a sample of 1 litre). Solution 1 - 0 . 1 p.p.m. each of iron(III), copper(I1) and nickel(II), 1-0 p.p.m. each of ammonia and hydrazine, 0.5 p.p.m. of sulphate and 0.2 p.p.m. of chloride. Solution 2-0-1 p.p.m. each of tungsten(V1) , aluminium(II1) , molybdenum(V1) , cobalt(I1) , zinc(I1) , vanadium(V) , manganese(I1) , chromium(II1) , magnesium(II), tin(I1) , titanium(1V) , fluoride and phosphate, 0.2 p.p.m.of sodium, 0.4 p.p.m. of potassium, 0.05 p.p.m. of ammonia, 0.06 p.p.m. of chloride and 2 p.p.m. of sulphate. Solutiovt 3-1.0 p.p.m. of cyclohexylamine, 2.0 p.p.m. of morpholine, 0.1 p.p.m. each of octadecylamine and alkylaryl sulphonate and 1.0 p.p.m. of ammonia. The mean results of these tests are given in Table IV; they indicate that at the con- centrations likely to be present in feed and make-up waters, these substances do not cause any appreciable bias. For solution 2, the final solution after fusion was slightly turbid, and gave an optical density about 0-016 greater than the blank value; this effect was eliminated by correcting for the turbidity as described under Sources of error. TABLE IV EFFECT OF OTHER SUBSTANCES Apparent* silicon content, p.p.m.of silica, a t silicon concentrations of- r A \ Substances added 0.000 p.p.m. of silica 0.030 p.p.m. of silica Solution 1 .. .. -0.001, 0.030 Solution 2 .. .. 0.001, 0.036, Solution 3 . . .. 0-003, 0.033, &0.004 p.p.m. of silica. *The 96 per cent. confidence limits for these determinations areFebruary, 19691 DETERMINATION OF SILICON I N WATER. PART VII 119 DISCUSSION ION-EXCHANGE RESINS- We used only “Powdex” resins in this work as they were the only resins of such fine mesh size known to us that were readily available in the hydrogen and hydroxyl forms. After this work was finished, we learned that the corresponding Dowex resins are also available in these ionic forms with sizes of 200 to 400 mesh. We would expect that any other manu- facturers’ resins would also be satisfactory, provided the size and ionic form were suitable.The silicon content of the “Powdex” resins is higher than is desirable, and the results obtained indicate that the amounts of resin used in the analysis of a sample contain about 40 pg of silica. We understand that Dowex resins with a silicon content of less than 1 p.p.m. of silica are available; the use of resins of this purity would markedly decrease the optical density of blank determinations and also improve the precision attainable. PRECISION- The results in Table 111 show that the total standard deviation of analytical results was about 0.0025 p.p.m. of silica at a concentration of 0.02 p.p.m. of silica; the standard deviation of 0.1 p.p.m. was slightly, but not significantly, greater.This precision is considered adequate, and represents a considerable improvement over that reported by Morrison and Wilson.2 The criterion of detectionll (taken as 2.326 times the within-batch standard deviation of the blank determinations) was about 0.005 p.p.m. of silica (95 per cent. confidence level). The within-batch standard deviations of the results are largely independent of the concentrations of silicon. The most likely cause of this imprecision is the heterogeneous distribution of silicon in the resins used, and it is, therefore, essential to test each batch of resin used to ensure that adequate precision is obtained. The most fruitful approach for improving precision appears to be by the use of resins with a much smaller silicon content than the resins used by the authors.In the present method a much greater volume of sample is used than that suggested for the previous method of Morrison and Wilson2 and, consequently, the precision of results with the present method is much less sensitive to contamination. BIAS- The results obtained indicate that 100 per cent. recovery of silicon was not obtained; the mean recoveries at concentrations of 0.02 and 0.1 p.p.m. of silica were 88 10 per cent. and 94 3 per cent., respectively. These recoveries are based on the optical density of a standard “reactive” silicon determination in which the concentration and fusion stages of the procedure were omitted. The results given earlier indicated that the ignition and fusion stages caused the slope of the calibration graph to decrease by about 4 per cent.The mean recoveries are consistent with an expected value of 96 per cent., and no further study of the apparent bias was made. It is concluded that the effect can be allowed for adequately by preparing a calibration graph as recommended under Method. TIME REQUIRED FOR ANALYSIS- The method is time consuming, but a batch of twelve analyses, including a blank and control standard can be carried out in about 8 hours, of which 5 hours are operator time. We thank Mr. G. S. Solt for making samples of “Powdex” resin available to us. We are also grateful to Mr. J, A. Tetlow for information, relating to the danger of incomplete destruction of hydrogen peroxide, obtained during his tests with the method. This paper is published by permission of the Central Electricity Generating Board. Baker and Farrant12 have recently published details of an adaptation of Morrison and Wilson’s method.2 They have taken great precautions to reduce random errors caused by contamination (for example, by working in a specially designed laboratory), and obtained appreciably better precision than that reported by Morrison and Wilson.2 Baker and Farrant’s paper is of value in showing the precision that can be achieved with their approach. However, laboratories meeting their specification will often not be readily available; the main advantage of the method in the present paper is that it is not as easily affected by con- tamination.120 WEBBER AND WILSON REFERENCES Morrison, I. R., and Wilson, A. L., Analyst, 1969, 94, 64. Robinson, H. E., G r i m , E. J., and Brown, C., Amer. S O ~ . Mech. Efigvs, 1951, Paper 61-A-93. Kosfeld, E. G., Tech. tfberwuach., Essen, 1967, 8, 133. Kostrikin, Yu. M., Shtern, 0. M., and Dzysyuk, A. A., Teplokevgetika, 1966, 13, 93. Webber, H. M., and Wilson, A. L., Analyst, 1964, 89, 632. Wickbold, R., 2. analyt. Chem., 1969, 171, 81. Morrison, I. R., and Wilson, A. L., Analyst, 1963, 88, 100. Duff, J. H.,-and Levendusky, J. A., PYOC. Amev. Peew Conf., 1962,24, 739. Feigl, F., “Spot Tests in Inorganic Analysis,” Translated by Oesper, R. E., Fifth Edition, Elsevier Publishing Co., Amsterdam: Cleaver-Hume Press Ltd., London; D. Van Nostrand Co. Inc., New York and Toronto, 1968, p. 333. Roos, J. B., Analyst, 1962, 87, 832. Baker, P. M., and Farrant, B. R., Ibid., 1968, 93, 732. NOTE-References 1, 2, 6 and 8 are to Parts VI, 111, IV and I1 of this series, respectively. , , Ibid., 1963, 88, 4.46. -- Received Febvuavy 21st, 1968 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

ISSN:0003-2654    

DOI:10.1039/AN9699400110

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, February, 1969, Vol. 94, $9. 121-125 121 An Improved Technique for Transferring Fractions from a Gas lChromatograph to a Mass Spectrometer BY W. D. WOOLLEY (Ministry of Technology and Fire O$ces’ Committee Joint Fire Research Organisation, Fire Research Station, Melvose Avenue, Boreham Wood, Herts.) A simple technique is outlined for the collection, storage and mass- spectrometric analysis of small amounts of volatile components separated by a gas - liquid chromatograph. Mass spectra obtained with an A.E.I. MSlOc2 mass spectrometer are shown, indicating a collection efficiency of about 95 per cent. A useful modification to the inlet system of the mass spectro- meter for the analysis of small samples is described, giving a sensitivity increase of up to thirty times the standard inlet sensitivity.This is demon- strated by a mass spectrum of 0.005 pl (3-5 pg) of hexane collected from a chromatograph. A maximum working sensitivity for the system is about 0.1 pg. THE recent introduction of relatively inexpensive low resolution mass spectrometers has considerably widened the use of the mass spectrometer as a general analytical instrument. Unfortunately, these instruments are usually slow-scanning and, without sophisticated inter- rupted elution techniques,l cannot be coupled directly to a gas chromatograph. The analytical system outlined in this paper was developed for the indirect mass spectro- metry of chromatographic fractions in a study of the possible toxic combustion products of building materials, especially plastics, involved in fires.A high analytical sensitivity was required for the analysis of minor components, particularly volatile oxygenated species, from burning plastics. In preliminary work, chromatographic fractions were collected in short tubes filled with column packing attached to the exit port of the chromatograph. Samples could be collected satisfactorily in this way but the over-all collection and re-elution into the mass-spectrometer inlet proved to be rather laborious. The heated line and refrigerated trap technique described in this report appears to be extremely simple to operate, and applicable to a wide range of materials. There is usually visual indication of trapped material in the glass trap during collection, and the traps are coupled directly to the mass spectrometer for analysis.Preliminary experiments in the analysis of the combustion products of poly(viny1 chloride) showed that many of the materials separated by gas chromatography were close to the detection limit of the A.E.I.* MSlOc2 mass spectrometer. Details are given of a modified inlet unit for the mass spectrometer such that the normal batch injection system is coupled to the analyser tube by a new and much faster leak. Thus a substantial increase in sensitivity has been obtained without altering the normal performance of the instrument and without using relatively complex direct injection technique^.^^^ * Associated Electrical Industries Ltd., now General Electric Company-Associated Electrical Industries 0 SAC; Crown Copyright Reserved. Ltd., Manchester.122 WOOLLEY AN IMPROVED TECHNIQUE FOR TRANSFERRING FRACTIONS [ATZdJJSt, VOl.94 EXPERIMENTAL COLLECTION SYSTEM- The collection unit is shown diagrammatically in Fig. 1. A length of thin-walled stainless- steel tubing with 1.25-mm o.d., supplied by Research and Industrial Instruments Co., is attached to the exit port of the gas chromatograph to direct the effluent gas into a glass collection trap, which is cooled with a refrigerant. Another stainless-steel tube running parallel to the collection line in the trap is silver-soldered to the tip, as shown, acts as an electrical conductor and provides, via hole C, an exit path for the carrier gas from the trap. The tube is heated electrically by applying a low voltage power supply across the points A and B.A and B = Stainless-steel tubes C = Hole in tube D = Glass collection trap E = Liquid nitrogen F = Silicone rubber cap G = Ceramic insulator H = Araldite J = Exit port of gas chromatograph K = Collection line L = Clamp Fig. 1. Details of the collection unit The two stainless-steel tubes are electrically insulated at the head of the trap by a 50-mm length of ceramic tubing of about 36mm 0.d. inserted over the main collecting tube, as shown. The unit is held together and fixed into a small clamp with Araldite. The Pyrex collecting tubes are held in place by a silicone rubber cap, cut out in the centre to fit over the ceramic tube. The collection tubes are about 130-mm long with 7-5-mm o.d., enlarging to 10-mm 0.d. near the collection zone to ensure that the collection tip does not touch the cold sides of the trap.For the collection of volatile material from the chromatograph, the lower part of the trap is cooled with liquid nitrogen. It is usually adequate to raise the liquid nitrogen container until the liquid surface is level with the tip of the collection tube. Further immersion into the liquid nitrogen may cool the collection line and promote condensation in the tip. After collection the trap is removed, quickly sealed with a silicone rubber cap and stored in liquid nitrogen. Samples collected in this way have been stored for several days without any apparent loss. The collection line has a maximum working temperature of about 300" C. Relatively high boiling materials are collected by using less severe refrigerants, such as a solid carbon dioxide bath at -78" C, or an ice-bath.February, 19691 123 For mass-spectrometric analysis the sample tubes are coupled directly to the MSlOc2 inlet, as shown in Fig. 2.For preliminary work the glass traps were simply inserted into a short length of PVC tubing attached to a brass adaptor on the inlet pipe, as shown in Fig. 2 (A). This gave an adequate vacuum seal but was later replaced by fixing the trap to a modified adaptor with a suitable O-ring, as in Fig. 2 (B). This latter method allowed the whole inlet system to be operated at elevated temperatures. FROM A GAS CHROMATOGRAPH TO A MASS SPECTROMETER t MSlOc2 inlet. t PV tubi Sample tubes r - <O ring A 0 Fig. 2. Collection tubes coupled to the mass spectrometer Fig. 3 (a) shows a mass spectrum of ethyl acetate collected from a 1-pl injection into the chromatograph, fitted for test purposes with a 1 : 1 collector - detector split.For com- parison purposes a mass spectrum of 0-5 p1 of ethyl acetate injected directly into a cold collec- tion tube with a syringe is shown in Fig. 3 (b). The two spectra indicate a collection efficiency of almost 95 per cent. The spectra were recorded on different days, and this may account for the slight variations in the general cracking pattern. The peaks at 44+ in the spectra arise from carbon dioxide frozen from the atmosphere into the cold tubes when the caps are removed before fitting to the mass spectrometer, but if the tubes are opened and fitted quickly to the inlet this peak is virtually eliminated.However, on certain occasions the 44+ peak provided a useful mass marker in the spectrum. .-- I mle m/e Fig. 3. Mass spectra of 0.6 pl of ethyl acetate. In (a) the sample is collected from a gas chromato- graph and is compared with a direct mass spectrum in (b). Ion intensities are given as percentages of full-scale recorder output on amplifier range 10 MODIFIED MASS-SPECTROMETER INLET- The A.E.I. MSlOc2 mass spectrometer contains a stainless-steel inlet block fitted with five valves for the pumping, handling and injection of samples into the main analyser tube. For the analysis of small amounts of material, the sample is introduced into the inlet block only and then into the analyser tube via a porous plug leak. The reservoir and pressure gauge cannot be used because of their large volumes.124 [Amlyst, Vol.94 The mass spectrum in Fig. 3 ( b ) indicates the general order of sensitivity of the mass spectrometer for materials injected in this way. The spectrum was recorded on amplifier range 10. A further magnification of ten times is available (range l), but a clean background is then more difficult to obtain and the pumping out time between samples increases. Various methods were investigated for increasing the sensitivity of the instrument. An inlet pipe inside the analyser tube to feed the sample directly into the ion source gave only a small increase in sensitivity. A reduction in the pumping speed of the main tube gave a further increase, but the pumping speed was soon returned to the original 2 litres per second because of the increased pumping out time between analyses. To maintain a constant pressure in the mass-spectrometer tube during an analysis, the inlet leak is designed so that only a small fraction of the sample leaks into the mass spectro- meter.Experiments showed that even a small sample would remain in the inlet block for several hours during an analysis, and it was considered that a sensitivity increase could be obtained by fitting a new and faster leak. To fix the new leak to the mass spectrometer without altering the normal performance of the instrument, the inlet pressure gauge was removed, and a new inlet line and leak bolted from the vacant gauge position to the spare port on the analyser tube. This gave an inlet system that could be operated as before, but now with the choice of a standard or fast leak.WOOLLEY: AN IMPROVED TECHNIQUE FOR TRANSFERRING FRACTIONS Swagelok union Stud and capillary leak --- To analyser section v '0' ring seals Inlet pipe 1 Fig. 4. Leak unit The inlet pipe consists of a length of thick-walled stainless-steel tube, 6-35 mm o.d., shaped to match the existing standard inlet line. The two inlet pipes are strapped together and heated by the same heating tape for operation at temperatures up to 150" C. The leak itself is a 20-mm length of stainless-steel capillary tubing, 0.15-mm i.d. and 1.58-mm o.d., fitted into the inlet line near the analyser tube, as shown in Fig. 4. The inlet pipe is cut and rejoined with a modified stainless-steel Swagelok union by using rubber O-rings in place of the normal metal ferrules.The capillary leak is soldered with a silver - 15 per cent. man- ganese soldering alloy, obtained from Johnson Matthey Metals Ltd., into a special stud, which, in turn, screws into the inlet pipe; the end of the inlet pipe and the face of the stud Fig. 6. Mass spectra of 0-005 p1 of hexane collected from a gas chromatograph recorded on (a), the standard leak, and (b), the fast leak. Ion intensities are given as percentages of full-scale recorder output on amplifier range 10February, 19691 125 are machined so no gasket is required. The leak unit is versatile because the leak can be readily removed simply by unbolting the Swagelok union. The inlet pipe is fixed to the pressure gauge port with an O-ring seal and the other end soldered into the analyser section.Fig. 5 (a) shows a mass spectrum, obtained when the standard leak was used, of hexane collected from a 1-p1 injection into the chromatograph of a solution of 1 per cent. hexane in toluene. When the 1 : 1 split is used this corresponds to a collection of 0.005 p1 (3.5 pg) of hexane. After analysis the fast-leak valve was opened and the new spectrum, shown in Fig. 5 (b), recorded. As can be seen the ion intensities are increased considerably and, based on more accurate measurements, a magnification of about thirty times is obtained with the new leak. However, at this rate of injection into the analyser tube, the ion currents show a decrease of about 10 per cent. in the 15 minutes required for a complete scan, but to overcome this difficulty it is usual, after a scan with a fast leak, to repeat the scan immediately. From a comparison of the first major peaks in each scan the rate of decrease is readily found. The mass spectrum in Fig. 5 (b) has been corrected to compensate for this error. As yet, little work has been done, with the collection system outlined in this report, on samples smaller than about 1 pg. However, by using the most sensitive range of the mass spectrometer (range 1) and a more favourable ratio of chromatographic collection to detector split (e.g., 20: l), a 0-l-pg sample of hexane injected into the chromatograph should give an acceptable mass spectrum. This paper is published by permission of the Director of the Fire Research Station of the Ministry of Technology and Fire Offices’ Committee. FROM A GAS CHROMATOGRAPH TO A MASS SPECTROMETER REFERENCES 1. 2. 3. Scott, R. P. W., in Littlewood, A. B., Editor, “Gas Chromatography 1966,” Institute of Petroleum, Ryhage, R., Ark. Kemi, 1962, 20, 185. McFadden, W. M., Separation Sci., 1966, 1, 723. London, 1967, p. 318. Received July 23rd, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400121

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

126 Analyst, February, 1969, Vol. 94, $$. 126-129 The Use of Thin-layer Chromatographic Techniques for the Determination of Breakdown Products of Additives to Plating Solutions BY W.-E. RUPPRECHT* ( Wilmot Breeden Ltd., A mington Road, Birmingham 26) The combination of conventional thin-layer chromatographic techniques with a novel form of column chromatography has enabled seventeen deriva- tives of coumarin in used nickel-plating solutions containing this additive to be isolated. Ten derivatives were identified, as mono-, di- and trihydroxy- coumarins and dihydrocoumarin, melilotic acid, o-coumaric acid and umbellic acid. Rp values and colour reactions of twenty typical coumarin derivatives are also reported. ORGANIC additives are frequently added to electroplating solutions to modify the properties of the electro-deposit .CH A common additive to nickel-plating solutions is coumarin. This is added primarily to give an even deposit, but it is also used to promote ductility, which is particularly required in motor-car components (bumpers, etc.). During electroplating, electrode reactions cause the formation of other compounds from coumarin. As these build up in solution, they counteract the beneficial effect of the additive. Identification of the breakdown products might lead to knowledge about the mechanism of their formation and possibly to ways of their prevention. Thin-layer chromatography was considered a more rapid and sensitive technique for identification than paper chromatography, and it had the added advantage that acidic plating solutions could be examined directly and compared with concentrated extracts of plating solutions, which were obtained from plating solutions adjusted to pH 1 by extraction with chlorof o m .Direct chromatography of the plating solutions and the extracts (thick, oily substances) proved difficult, because of masking effects and “tailing” by the additive itself and one breakdown product, identified as melilotic acid (o-hydroxyphenylpropionic acid). [rH2-cH2- Column chromatography was tried to improve separation, but a solid, open column, cast from thin-layer adsorbent slurry, which could be developed like an ordinary thin-layer chromatogram, provided a more convenient means. This was followed by a further separation of the minor constituents on thin-layer chromatographic plates.EXPERIMENTAL PREPARATION OF COLUMNS- Cellophane dialysis tubing (60 cm long x 20 mm 0.d.) was sealed at one end by a knot and filled with silica gel G or aluminium oxide G slurry, as made for normal thin-layer chromatographic plates, The tube was then suspended vertically and the bottom per- forated by a few holes (to allow excess water to drain off). After 20 minutes, the tube was *Present Address : The Gas Council, Midlands Research Station, Wharf Lane, Solihull, Warwickshire. 0 SAC and the author.RUPPRECHT 127 cut, with a razor blade, into sections of 20-cm length, which were then placed in an oven for 25 minutes at 110" C, when the cellophane could be removed. The columns thus obtained were activated for 60 minutes at 110" C.PREPARATION AND STORAGE OF PLATES- Good quality glass plates were coated with a 250-p thick layer of water - silica gel G (2 + 1) or water - aluminium oxide G (4 + 3) slurry, with a home-made spreader. After being left to set for 15 minutes, the plates were activated for 30 minutes at 130" C and subsequently stored in an empty desiccator under vacuum. SAMPLE APPLICATION- Amounts of 1 to 10 p1 of solution were applied at 2 cm from the bottom edge of the plates by repeated application of 1-p1 drops. For columns, 200 to 500 pl of solution were applied to the circumference at 2 cm from the bottom edge by allowing the solution, contained in a short piece of glass tubing, to flow through a capillary on to the rotating column. DEVELOPMENT- Plates were developed in a tank (23 x 23 x 8 cm) fitted with a ground-glass lid and lined with filter-paper for uniform saturation.Columns were developed either in cork-stoppered, flat-based test-tubes or, alternatively, in a shallow Petri dish filled with solvent and covered with a beaker to provide an air-tight seal. Plates were developed 15 cm and columns 10 cm. SOLVENTS- The three solvents, toluene - ethyl acetate - formic acid (5 + 4 + 1); hexane - ethyl acetate (3 + 1) ; and toluene - ethyl formate - formic acid (5 + 4 + 1) , were tried. Although the second and third gave good separation of monohydroxy derivatives of coumarin on silica gel G, the third gave improved separation on aluminium oxide G and was used throughout the later work; results reported are confined to the third solvent.DETECTION- Columns were examined under ultraviolet light (253 nm) only; detected bands were cut from the column, extracted with purified ethanol, and the concentrated extract applied to a normal thin-layer chromatographic plate. Plates were examined under ultraviolet light before and after spraying with either diazo- tised sulphanilic acid in 5 N sodium hydroxide solution, diazotised o-dianisidine in 5 N sodium hydroxide solution, p-nitrobenzenediazonium fluoroborate in water or N sodium hydroxide solution. COLOUR PHOTOGRAPHY- Plates, 20 x 15 cm, were recorded on 35-mm Agfacolor CT18 film under ultraviolet light (253 nm) ; the exposure was 10 minutes at f 11. Two standard photographic ultraviolet filters were used to exclude reflected ultraviolet light. RESULTS AND DISCUSSION Table I gives RF values and colour reactions of some synthesised coumarin derivatives and related compounds. The relationship between RF value and structure is clear, especially the decrease in R, value as the substituent moves from the 3- to the 5- and 7-positions, or from the 4- to the 6- and 8-positions, and when a second and third hydroxyl group is introduced.In plating solutions only 7-hydroxycoumarin and o-coumaric acid were detected, in addition to coumarin and melilotic acid. The chloroform extractions provided the breakdown products in the plating solution in a much more concentrated form and hence the number of identified compounds is much greater; it increases from two, in an extract from a relatively new solution, to seventeen, in an extract from a spent solution.Melilotic acid, o-coumaric acid, 4-hydroxycoumarin , 6-hydroxycoumarin, 7-hydroxycoumarin, 6,7-dihydroxycoumarin. 4,6,7-trihydroxycoumarinDerivative 3-Hydroxycoumarin . . 4-Hydroxycoumarin . . 5-Hydroxycoumarin . . 6-Hydroxycoumarin . . 7-Hydroxycoumarin . . 8-Hydroxycoumarin . . Coumarin .. .. Dihydrocoumarin . . Melilotic acid . . .. o-Coumaric acid .. Umbellic acid . . .. 3,6-Dihydroxycoumarin 4,7-Dihydroxycoumarin 5,6-Dihydroxycoumarin 5,7-Dihydroxycoumarin 6,7-Dihydroxycoumarin 7,8-Dihydroxycoumarin 4,5,7-Trihydroxycoumarin 4,6,7-Trihydroxycoumari 4,7,8-Trihydroxycoumarin 2 3 5 6 7 8 9 10 13 14 16 17 18 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. TABLE I COUMARIN DERIVATIVES SEPARATED BY THIN-LAYER CHROMATOGRAPHY SoIvent, toluene - ethyl formate - formic acid; layer thickness, 250p RF value * on on Al,O, silica G gel G 0.85 0.55 0.71 0.40 0.78 0.42 0.73 0.37 0.75 0.38 0.66 0.35 0.89 0.64 0.91 0.61 0.75 0.46 0.72 0.48 0.59 0-35 0.55 nd 0.46 nd 0.5 1 nd 0.63 nd 0.47 nd 0.59 nd 0.34 nd 0-46 nd 0-44 nd Diazotised Diazotised Ultraviolet light sulphanilic o-dianisidine (253 nm) acid spray spray * * * no sodium ultra- ultra- spray hydroxide visible violet visible violet aqueous 26 26 7 34 33 65 66 65 16 32 34 26 44 33 33 44 26 32 26 - 26 26 7 34 33 65 40 - - 47 44 7 26 7 66 39 44 26 32 26 71 nd 58 8 8 6 7 8 7 nd 5 nd nd 48 nd 7 71 57 - - 40 26 nd 72 33 65 40 13 61 47 51 nd 26 nd nd 40 nd 65 71 70 5 56 26 nd 13 17 64 18 26 nd 18 nd 65 56 67 23 57 66 I - nd = not determined.- - - not detectable.44 26 7 3 27 nd 40 14 40 33 nd 26 nd 65 44 nd 25 71 70 - Colour code (appearance as compared with a Denvent colour pencil chart) P-Nitrobenzene- diazonium fluoroborate spray * ultra- visible violet 69 10 2 44 nd nd 58 65 59 33 nd nd - 65 3 65 67 66 57 65 9 23 nd nd 68 44 nd nd nd nd 57 44 nd nd 62 65 18 27 66 23 Lemon Cadmium Gold Straw Yellow Deep Cadmium Naples Yellow Middle Chrome Deep Chrome Orange Chrome Pale Vermilion Deep Vermilion Flesh Pink Pink Madder Lake Rose Pink 23 26 26 27 32 33 34 39 40 44 47 48 Imperial Purple Dark Violet Light Violet Blue Violet Lake Spectrum Blue Light Blue Sky Blue Turquoise Blue Turquoise Green Water Green Grass Green May Green 51 56 57 68 59 61 62 64 65 70 71 72 Olive Green Raw Umber Brown Ochre Raw Sienna Golden Brown Copper Beech Burnt Sienna Terra-cotta Burnt Carmine French Grey Silver Grey White Found in CHCl, extract Foundin of plating plating solution solution X X X X X X X X X X X X X XFebruary, 19691 TECHNIQUES FOR THE DETERMINATION OF BREAKDOWN PRODUCTS 129 and umbellic acid were identified positively, and 5-hydroxycoumarin tentatively.The RF values and colour reactions of other compounds could not be related to any available derivative. Melilotic acid was always detectable in all samples, including extracts. Both o-coumaric acid and 7-hydroxycoumarin appeared at a fairly early stage in the life of the plating solution, and their concentration increased gradually throughout the life of the plating solution. Many reaction paths are possible for the various compounds produced. Melilotic acid appears to be formed by cathodic hydrogenation of coumarin to S,Pdihydrocoumarin, followed by rapid hydrolysis at pH 4. For the formation of o-coumaric acid, Ashurstl has suggested that the reaction might be analogous to the oxidation of cyclohexane. 6-Hydroxy- coumarin and 6,7-dihydroxycoumarin could be formed by the well known persulphate oxidation.2 The persulphate formation in the plating solution according to the reaction 2SO42- - 2e- --+ S20,2- is quite feasible. However, a random attack of the coumarin molecule by active oxygen is the most likely explanation. The author thanks Mr. S. D. Cashmore for the interest shown and help given during this study, and the Directors of Messrs. Wilmot Breeden Ltd., for permission to publish this paper. REFERENCES 1. 2. Ashurst, K. G., Trans. Inst. Metal Finish., 1963, 40, 74. Bargellini, G., and Monti, L., Gazz. Chim. Ital., 1916, 45, 96. Received April llth, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400126

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

130 Analyst, February, 1969, Vol. 94, fip. 130-135 Gas-chromatographic Determination of Acetyl and Trirnethylsilyl Derivatives of Alkyl Carbarnates and their N-Hydroxy Derivatives BY R. NERY (Chester Beatty Research Institute, Institute of Cancer Research, Royal Cancer Hospital, Fulham Road, LondoH, S . W.3) Microgram amounts of mixtures of alkyl carbamates, and of urethane and N-hydroxyurethane, as their trimethylsilyl derivatives, and similar mixtures of alkyl N-hydroxycarbamates, as their trimethylsilyl and acetyl derivatives, have been analysed by gas chromatography on SE30 columns. With a programmed temperature rise, the elution temperatures varied linearly with the number of carbon atoms in the alkyl side-chain of a homologous series; the isobutyl analogues were eluted at lower temperatures than the corresponding butyl analogues.When the corresponding isobutyl derivatives were used as internal standards, the ratios of peak heights (test to standard) vaned linearly with the concentration of urethane and N-hydroxyurethane. THE alkyl carbamates and their N-hydroxy derivatives are of interest in carcinogenesis and in many other chemical and biological studies.ls2J These substances have been determined by various paper-chromatographic4 and colorimetric procedures,6 8 and several carbamates have been qualitatively analysed by gas chromatography.’ This paper describes (i) the quantitative analysis by gas chromatography of urethane and N-hydroxyurethane, as their acetyl or trimethylsilyl derivatives, and (ii) the qualitative analysis of mixtures of alkyl carbamates or their N-hydroxy analogues, containing from one to six carbon atoms in the alkyl group in a homologous series, and one branched chain, the isobutyl group.METHOD APPARATUS- The analyses were performed on a Perkin-Elmer, Model 800, dual column, gas chromato- graph, incorporating a dual flame-ionisation detector and a Honeywell “Electronik” con- tinuous balance recorder with a range of from 0.25 to 2.5 mV. The dual chromatographic columns consisted of 1-m stainless-steel, coiled tubes of & inch o.d., packed with 1-5 per cent. silicone gum rubber (SE30) on a solid support of 80 to 100-mesh, HMDS-treated Chromosorb W. OPERATING CONDITIONS- The conditions used were : hydrogen pressure, 15 lb per inch2; air pressure, 30 Ib per inch2; nitrogen flow-rate through both columns, 30 ml per minute; and injector block temperature, about 160” C (dial setting 4).The rate of temperature rise in both columns in all of the experiments was 5” C per minute. The columns were used after equilibration for 24 hours at an oven temperature of ZOOOC, and all recordings were made at the basic chart speed of 15 inches per hour. 0 SAC and the author.NERY 131 REAGENTS- Py ridine . Triethylamine. Diethyl ether. Trimethy lchlorosilane . Hexamethyldisilazane. All were dried and distilled before use. MATERIALS- Isobutyl and pentyl carbamates were prepared from the corresponding chloroformates and ammonia solution. All were recrystallised to constant melting-points and dried overnight in an evacuated desiccator over phosphorus pentoxide before use.The alkyl N-hydroxycarbamates and some of their 0-acetyl and NO-diacetyl derivatives were prepared as previously d e s ~ r i b e d . ~ ~ ~ TrimethyZsilyZ-N-hydroxyurethane-A 10-ml portion of trimethylchlorosilane was added, dropwise, during 15 minutes, to a stirred mixture of 5 g of N-hydroxyurethane, 20ml of hexamethyldisilazane and 50 ml of pyridine. The mixture was heated at 75" C for 2 hours, diluted with 100 ml of ether, stored at 4" C for 16 hours, filtered and the residue washed with 50 ml of ether. The ethereal solutions were combined, washed with two 20-ml portions of water, dried over anhydrous sodium sulphate and distilled, giving 8 g (95 per cent. yield) of trimethylsilyl-N-hydroxyurethane as a colourless oil with a fruity odour (b.p. 45" to 49" C at 0.04-mm pressure of mercury).The composition of the oil is given below. Element . . .. .. c H N Found, per cent. . . . . 41-28 9-12 7.50 C,H,5N0,Si requires . . 40.66 8.53 7.90 Methyl, ethyl, propyl and butyl carbamates were of commercial origin. The compound was miscible with all of the common organic solvents and immiscible with water; it gave an immediate purple colour with 1 per cent. w/v iron(II1) chloride solution, and immediately reduced ammoniacal silver nitrate solution. These properties are consistent with those expected from an N-substituted derivative, but do not provide conclusive evidence of structure, which was not further investigated. Bis-(trimethylsilyl)-N-hydroxyz.wethane-A mixture of 9.6 g of N-hydroxyurethane and 100ml of pyridine was introduced into a 1-litre, round-bottomed flask, equipped with a mercury-sealed stirrer, a dropping funnel and a reflux condenser with exit guarded by a calcium chloride drying tube.The apparatus was flushed with dry nitrogen, 25ml of tri- ethylamine introduced and 30 ml of trimethylchlorosilane added, dropwise, with stirring, during 30 minutes. The mixture was heated under reflux for 3 hours, allowed to cool, diluted with 300 ml of ether, filtered after being allowed to stand for 16 hours at 4" C and the residue washed with 100 ml of ether. The ethereal solutions were combined and distilled, giving 18 g (79 per cent. yield) of bis-(trimethylsily1)-N-hydroxyurethane as a colourless liquid (b.p. 42" to 44" C at 0-1-mm pressure of mercury) with a fruity odour.The composition of the oil is given below. Element . . .. .. c H N Found, per cent. . . . . 43-78 9.44 6.14 CoH,3N03Si, requires . . 43.33 9.29 5.62 The compound had properties similar to those described for the mono-trimethylsilyl derivative, except that the coloration with iron(II1) chloride solution and the reduction of ammoniacal silver nitrate solution occurred more slowly. PROCEDURES- 1. Preparation and chromatography of trimethylsilyl derivatives-The carbamates and N-hydroxycarbamates were dissolved in 0.5 ml of pyridine, and 0.2 ml of hexamethyldisilazane, 0.1 ml of trimethylchlorosilane and 0.2 ml of triethylamine then added. After being allowed to stand for 4 hours at 50" C, the mixtures were centrifuged, and 1 pl of the supernatant liquor was analysed as described under Operating conditions.2. Preparation and chromatography of acetyl derivatives-The N-hydroxycarbamates were dissolved in 0.5 ml of pyridine, to which were then added 0-2 ml of acetic anhydride and 0.3 ml of triethylamine. After being allowed to stand for 4 hours at 23" C, a 1-p1 portion was analysed as described under Operating conditions.132 NERY : GAS-CHROMATOGRAPHIC DETERMINATION OF ACETYL AND [ArtdySt, VOl. 94 The concentrations of the test substances in the final mixtures obtained by Procedures 1 and 2 were such that 1 pl contained the amounts determined, as shown in Figs. 1 to 6. RESULTS AND DISCUSSION Zielinski and Fishbein' described the qualitative analysis of several alkyl carbamates by gas chromatography on two polar columns [Carbowax 20M, a poly(ethy1ene glycol) resin, and Versamid 900, a polyamide resin] and on one non-polar column (SE30, a methylsilicone resin).In the present study, 1 pl of a 0.1 M solution of urethane in ether was not detectable under the conditions described (see legend, Fig. 1); 1 p1 of a 1.0 M solution, under similar conditions, gave a broad elution peak, with a mean retention time of 3.8 minutes, and a sharp peak, with a retention time of 1.0 minute when the temperature programme was changed to 100" to 120" C, other conditions being unchanged. This agrees with the 1-2 minutes reported bv Zielinski and Fi~hbein.~ who used isothermal conditions (100" (3 and a 10 per \ , Chromosorb W. cent. w/v SE30 coating on 60 to Retention time, minutes Number of C atoms in n-alkyl side-chain Fig.1. Gas chromatogram of alkyl carba- Fig. 2. Relationship between alkyl chain mates (Procedure 1). Temperature programme : length and elution temperature in the gas 35" to 100" C; and attenuation x 200. Amount chromatography of n-alkyl carbamates and of each carbamate analysed, 5 x lod8 moles. N-hydroxycarbamates by Procedures 1 and 2. Alkyl groups: (a), methyl; (b), ethyl: (c), propyl; 0, n-Alkyl carbamates (Procedure 1). Conditions ( d ) , isobutyl; (e), butyl; and (f), pentyl as for Fig. 1. A, n-Alkyl N-hydroxycarbamates (Procedure 1). Temperature programme: 36" to 130°C; and attenuation x 200. 0, n-Alkyl N-hydroxycarbamates (Procedure 2). Tempera- ture programme: 50" to 140" C; and attenuation x 20 The generally more volatile and less polar derivatives formed when active hydrogen atoms are substituted by trimethysilyl groups have been extensively used in the gas-chromatographic determination of many substances, including pesticidal carbamates,lO ureas,lOs11 steroids,12,13 fatty acids,l4 amino-acids,15s16 biological amines17s1* and ~arb0hydrates.l~ The elution tem- peratures of the trimethylsilyl derivatives of the homologous series of alkyl carbamates (from methyl to pentyl), and of the trimethylsilyl and acetyl derivatives of the alkyl N-hydroxy- carbamates (from methyl to hexyl), under the conditions described, were directly proportional to the number of carbon atoms in the alkyl side-chains, except for the methyl derivatives, which gave somewhat longer retention times (Fig.2); the elution temperatures were also approxi- mately proportional to the respective boiling-points of the parent compounds (Table I), Comparison of the butyl and corresponding isobutyl derivatives shows that chain branching reduced the retention time (Figs. 1, 3 and 4) in spite of the higher boiling-point of isobutyl carbamate (Table I). The same effect of chain branching on retention time has been observedFebruary, 19691 TRIMETHYLSILYL DERIVATIVES OF ALKYL CARBAMATES 133 TABLE I GAS CHROMATOGRAPHY OF ALKYL CARBAMATES AND ALKYL N-HYDROXYCARBAMATES BY PROCEDURES 1 AND 2 Compound A. ROCONH, R = Methyl Ethyl Propyl Isobu tyl Bu tyl Pentyl R = Methyl Ethyl Propyl Isobutyl Butyl Pentyl Hexyl B. ROCONHOH Boiling-point, "C/mm pressure of mercury 177/760 184/760 195/760 206/760 203 / 760 56$ 50 to 51: 86 to 88/0.6 90 to 92/0.6 41$ 100 to 102/0-8 115 to 118/0*04 42 3 Reference t 23 23 23 23 23 7 8 8 8 9 8 9 9 Relative elution* Trimethylsilyl derivative 0.60 1.0 1.73 2.20 2.53 3.33 0.79 1-0 1-29 1.45 1-58 1.84 2.13 1 Acetyl derivative N.D.N.D. N.D. N.D. N.D. N.D. 0.77 1.0 1.33 1.53 1.70 2.07 2.43 * Relative to the retention time of the corresponding derivative of the ethyl analogue. For the trimethylsilyl derivatives of urethane and N-hydroxyurethane, and for the acetyl derivative of N-hydroxyurethane, the retention times under the conditions described were 3.0, 7.6 and 6-0 minutes, respectively. t Refers to sources of values given in second column. $ Melting-point, "C. N.D.-Not determined. when 0-alkyl carbamates' and their N-alkyl derivatives20 were determined directly on polar and non-polar columns. When the corresponding isobutyl derivatives were used as internal standards, urethane (by Procedure 1) and N-hydroxyurethane (by Procedures 1 and 2) gave peak heights that were directly proportional to the concentration of the corresponding test substance (Fig.5). Fig. 6 shows the elution of a mixture of urethane and N-hydroxyurethane, each at a concentration of 0.1 M in pyridine and determined by Procedure 1. d I I I I 16 12 8 4 Retention time. minutes d 116 I I I I 1 16 I2 8 4 Fig. 3. Gas chromatogram of alkyl N - hydroxycarbamates (Procedure 1). Temperature programme: 35" to 130" C and attenuation x 200. Amount of each hydroxycarbamate Fig. 4. Gas chromatogram of alkyl analysed, 4-25 x 10-8 moles.Alkyl groups: N-hydroxycarbamates (Procedure 2). Tempera- (a) to (f) as for Fig. 1; and (g), hexyl ture programme: 60" to 140" C; and attenuation x 20. Amount of each hydroxycarbamate analysed. 1.26 x moles. Alkyl groups as for Fig. 3 Retention time, minutes134 NERY : GAS-CHROMATOGRAPHIC 3, 1 Amount of test substance x IO* moles DETERMINATION OF ACETYL AND [ArtdySt, VOl. 94 Retention time, minutes Fig. 5. Calibration graphs for urethane (Procedure 1) and N-hydroxyurethane (Pro- Fig. 6. Gas chromatogram of a mixture of cedures 1 and 2). 0, Urethane. Temperature urethane and N-hydroxyurethane (Procedure 1). programme : 36" to 95" C; and attenuation x 200. Temperature programme: 36" to 90" C; and Amount of isobutyl carbamate in each assay, attenuation x 200.Amount of each substance 6 x 10-8 moles. A, N-Hydroxyurethane (Pro- analysed, 5 x 10-8 moles. (a), Urethane; and cedure 1). Conditions as for urethane, except (b) , N-hydroxyurethane that isobutyl N-hydroxycarbamate was used as internal standard. 0, N-Hydroxyurethane (Pro- cedure 2). Temperature programme: 60" to 110" C; and attenuation x 20. Amount of isobutyl N-hydroxycarbamate in each assay, 1.25 x 10-8 moles The structures of the products formed after trimethylsilylation or acetylation of N-hydroxyurethane, by Procedures 1 and 2, were ascertained by comparison of their retention times with those of the synthetic mono- and disubstituted derivatives (see Materials). Trimethylsilyl- and O-acetyl-N-hydroxyurethanes were first formed by Procedures 1 and 2, respectively, with retention times of 8.0 and 6-2 minutes; after being allowed to stand for 4 hours, these were quantitatively converted into the corresponding disubstituted derivatives, retention times for the di-(trimethylsilyl) and diacetyl derivatives being 7.6 and 6-0 minutes, respectively (Figs.3 and 4, and Table I). These results show that, for N-hydroxyurethanes, the species that are analysed after 4 hours (by Procedures 1 and 2) are the di- and not the monosubstituted derivatives. The other alkyl N-hydroxycarbamates probably formed analogous derivatives, as no change in the retention times, obtained after 4 hours, occurred after longer reaction times (up to 8 hours). No attempt was made to establish the structures of the trimethylsilylcarbamates, but they were probably the NN-bis(trimethylsily1) deriva- tives, as acetamide reacts analogously,21 and the retention times of the derivatives formed were not changed after up to 18 hours.The results shown (Figs. 1 to 6 and Table I) were reproducible on repeat determinations, providing the conditions described were unchanged. Each experiment was performed at least five times, and standard deviations varied between & 0.005 and 0.021 ; larger variations sometimes occurred, but these were due to malfunctioning of the instrument, e.g., to partial clogging of the injector port or columns. The results were obtained by starting with known amounts of the various unsubstituted carbamates or N-hydroxycarbamates, forming the rele- vant derivatives (see Procedures 1 and 2), and analysing 1 pl of the corresponding supernatant solutions.For N-hydroxyurethane alone, the quantitative relationships shown in Fig. 5 were also obtained by direct analysis of known amounts of the pre-formed NO-disubstituted acetyl and trimethylsilyl derivatives. No attempt was made to establish a linear peak height - concentration relationship, except for urethane and N-hydroxyurethane (Fig. 5). This work resulted from an attempt to study, by gas chromatography, the metabolic interconversion22 of urethane and N-hydroxyurethane by rodent tissues. Organic solvent extracts of de-proteinised rat or mouse liver homogenates, when analysed by Procedures 1 and 2, contained interfering substances that gave ambiguous results when the homogenates v ( : ( incubated with urethane or N-hydroxyurethane. Attempts to eliminate the interferenceFebruary, 19691 TRIMETHYLSILYL DERIVATIVES OF ALKYL CARBAMATES 135 by using ether, benzene, chloroform or methylene chloride as the solvent for extraction, or trichloroacetic acid, ammonium sulphate, zinc sulphate - sodium hydroxide or ethanol as the protein precipitant, were unsuccessful.The author thanks Professor E. Boyland and Dr. D. H. de Kock for their interest. This investigation has been supported by grants to the Chester Beatty Research Institute (Insti- tute of Cancer Research, Royal Cancer Hospital) from the Medical Research Council and the British Empire Cancer Campaign for Research, and by the Public Health Service Grant No. CA-03188-09 from the National Cancer Institute, US.Public Health Service. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. REFERENCES Haddow, A., in “Professor Khanolkar Felicitation Volume,” Bombay University Press, Bombay, Tarnowski, G. S., Kreis, W., Schmid, F. A., Cappuccino, J. G., and Burchenal, J. H., Cancev Rcs., Adam, P., and Baron, F. A., Chem. Rev., 1965, 65, 667. Moss, M. S., and Jackson, J . V., J . Pharm. Phavmac., 1961, 13, 361. Marquart, R. P., and Luce, E. N., J . Agric. Fd Chem., 1963, 11. 77. Nery, R., Analyst, 1966, 91, 388. Zielinski, W. L., and Fishbein, L., J . Gas Chvomat., 1965, 3, 142. Boyland, E., and Nery, R., Analyst, 1964, 89, 620. -- , J . Chem. SOC. ( C ) , 1966, 346. Fishbein, L., and Zielinski, W. L., J . Chromat., 1965, 20, 9. Wannagat, U., Buerger, H., Krueger, C., and Pump, J., 2. anorg. allg. Chem., 1963, 321,208. Wells, W. W., and Makita, M., Analyt. Biochem., 1962, 4, 204. Luukkainen, T., and Adlercreutz, H., Biochim. Biophys. Ada, 1963, 70, 700. Wood, R. R., Raju, P. K., and Reiser, R., J. Amer. Oil Chem. Sot., 1966, 42, 161. Ruhlman, K., and Giesecke, W., Angew. Chem., 1961, 73, 113. Birkofer, L., Ritter, A., and Neuhausen, P., Justus Liebigs Annln Chem., 1962, 659, 190. Brooks, C. J. W., and Horning, E. C., Analyt. Chem., 1964, 36, 1540. Sen, N. P., and McGeer, P. N., Biochem. Biophys. Res. Commun., 1963, 13, 390. Wells, W. W., Chin, T., and Weber, B., Clinica Cham. Acta. 1964, 10, 362. Zielinski, W. L., and Fishbein, L., J . Gas Chromat., 1965, 3, 333. Klebe, 2. F., Finkbeiner, H., and White, D. M., J. Amer. Chem. SOC., 1966, 88, 3390. Boyland, E., and Nery, R., Biochem;, J.. 1965, 94, 198. “Dictionary of Organic Compounds, 1963, p. 168. 1966,26, 1279. Volume 1, Fourth Edition, Eyre and Spottiswoode Ltd., First received August 16th, 1967 Amended June 7th, 1968 London, 1966, p. 661.

ISSN:0003-2654    

DOI:10.1039/AN9699400130

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

136 Analyst, February, 1969, Vol. 94, fifi. 136-142 A Method for the Analysis of Cereals and Groundnuts for Three M.ycotoxins BY L. J. VORSTER (National Nutrition Research Institute of the South African Council for Scientijc and Industrial Research, P.O. Box 395, Pretoria, Republic of South Africa) A method is proposed for the analysis of samples for three mycotoxins, aflatoxin, ochratoxin and sterigmatocystin, by suitable treatment of a single sample extract. Based on the subjective evaluation of thin-layer chromato- grams of the extract, results can be reproduced with an accuracy of f20 per cent. The method is considered to be satisfactory for the purposes of a field survey when the determination of the approximate level of mycotoxin con- tamination of cereals and groundnuts in the shortest possible time is of prime importance.Problems encountered with samples that have high oil contents or that are darkly pigmented are dealt with by appropriate modifications of the method. THE discovery of the aflatoxins has led to much research in the field of mould contamination of food and feeds. It has already been established that many moulds elaborate metabolites that are toxic or carcinogenic, or both, to laboratory animals. While direct evidence linking aflatoxin, ochratoxin and sterigmatocystin with diseases in man is lacking, the possibility cannot be excluded that the presence of these materials in food constitutes a grave threat to human health. It has, therefore, become essential for the safety of both man and animals that the extent to which our food and feed crops are contaminated by mycotoxins should be determined. Assaying for mycotoxins, for which analytical procedures have been developed, is expen- sive and time consuming, especially in respect of the preparation from the samples of extracts containing the toxins.It would, therefore, be of great advantage if a method could be developed to assay for a number of mycotoxins in the same sample extract. The following paper deals with work carried out to determine whether ailatoxin, ochra- toxin and sterigmatocystin contents could be determined on one extract of a sample. As a result, a method is proposed that gives satisfactory results when applied to samples of maize and sorghum. Problems caused by high fat content of the extract (as is the case with ground- nuts) and by fluorescent pigments (extracted from some varieties of sorghum) were overcome by modifying the basic method.PREPARATION OF SAMPLES AND STANDARD SOLUTIONS FOR PRELIMINARY ANALYSIS PREPARATION OF MYCOTOXIN-CONTAINING SAMPLES- The main South African crops liable to contamination by mycotoxins are maize, ground- nuts and sorghum. It was, therefore, decided that each of the three mycotoxins being investigated should be produced separately on media prepared from each of the three food crops referred to. After determining the mycotoxin contents of these materials, samples containing varying but known concentrations of all three mycotoxins were prepared by mixing the media in suitable proportions. These samples were used as raw material for the proposed investigation.0 SAC and the author.VORSTER 137 Separate samples (250 to 300 g) of the three media, crushed and adjusted to a moisture content of 30 per cent., were sterilised in 5-litre Erlenmeyer flasks and inoculated with the appropriate mould spores (Aspergillus j a w s , A. ochraceous and A. rtidulans) . The cultures were incubated at 28" C ( rf: 3" C) in a darkened room for 11 days in the case of A . ochraceous, and for 17 to 18 days for the other moulds. The cultures were then dried in a forced-draught oven at 40" C before being ground in a hammer mill. Each medium was thus cultured to produce, separately, each of the three mycotoxins under investigation. The aflatoxin contents of the respective samples were determined according to the procedure prescribed by the IUPAC Sub-Commission on trace substances.1 The method of Steyn and van der Menve,2 with modifications suggested by Scott and HandJs was used to determine ochratoxin contents, and the procedure proposed by Vorster and Purchase4 was used for sterigmatocystin.Each culture was checked qualitatively to ensure that it was not contaminated by the other mycotoxins under investigation. PREPARATION OF QUANTITATIVE AND QUALITATIVE STANDARD SOLUTIONS- Samples of several toxins known to be produced by the three species of fungi under consideration, with the notable exceptions of aspertoxin6 and nidulo1,B were obtained either in pure form or as mixtures for use in reference solutions. Ajatoxiuts-Pure aflatoxins B, and GI were dissolved separately in analytical-reagent grade benzene' to give estimated concentrations of 10 pg ml-l.The procedure described in Appendix A of the IUPAC bulletin1 was followed to determine the exact concentration of the respective toxins in the two solutions. A qualitative standard solution was prepared by dissolving sufficient amounts of the four major, and all of the available minor, aflatoxins in benzene so that lop1 of the solution would produce readily visible spots of each toxin when separated by thin-layer chromatography. Ochratoxiuts-Chloroform solutions of ochratoxins A and C (ethyl ester of A) containing an estimated 40pgml-1 were prepared (allowing for benzene content of crystalline A) and standardised by the procedure described for the aflatoxins, substituting the following values8 in the calculations- Ochratoxin A: E ~ ~ ~ , ~ ~ ~ = 2400; molecular weight = 402.5 Ochratoxin C: E~~~~~~~ = 7000; molecular weight = 430.5 As ochratoxin B and its methyl and ethyl esters have been reported to be non-toxicJ9 it was considered that their determination would be unnecessary.Consequently, no standard solutions of them were prepared. Stwigmatocystin-Pure sterigmatocystin was dissolved in analytical-reagent grade chloroform to give an estimated concentration of 200 ,ug ml-1. The optical density a t 327 nm was measured and the concentration calculated ( ~ : $ i ~ ~ ~ ~ ~ ~ = 16,220 and molecular weight = 324). DETERMINATION OF THE LIMIT OF VISUAL DETECTABILITY The subjective evaluation of chromatograms is considered to be more accurate when fluorescent spots at, or just above, the limit of visual detectability are compared with spots of about the same intensity derived from a standard solution.Consequently, the limit of visual detectability for each of the mycotoxins in question was determined. This was done by spotting different amounts of each standard solution on an activated silica chromatoplate, starting with aliquots that were known to produce readily visible spots and decreasing the amount, stepwise, until a point was reached at which the developed chromatogram would no longer contain a visible fluorescent spot. Each plate was then developed with an eluant suitable for the particular toxin. After evaporation of the solvent, the plate was examined under long wave (peak eniission 360 nm) ultraviolet light or a combination of long and short wave (peak emission 254 nm) ultraviolet light.The smallest detectable amount was thus determined and noted as being the limit of visual detectability for each particular toxin. For sterigmatocystin this value could be reduced to about one quarter by lightly spraying the chromatogram with 20 per cent. potassium hydroxide solution. The dull red fluorescence138 VORSTER: A METHOD FOR THE ANALYSIS OF [Analyst, Vol. 94 of the toxin is thereby changed to greenish yellow, which is easier to detect (Note 1). The following limits of visual detectability values were determined under 5 x 20-watt Philips fluorescent tubes (TL20W/08), positioned 30 cm above the developed chromatoplate- Aflatoxin B, .. .. . . 0.0004 pg Aflatoxin GI .... . . 0.0003 pg Ochratoxins A and C . . . . 0.002 pg Sterigmatocystin . . .. . . 0.04 pg (brick red) 0.01 pg (yellow) NOTE 1- The technique of spraying a developed chromatoplate with potassium hydroxide solution is found to be very useful during the initial evaluation of chromatograms of a crude extract. The fluorescence of ochratoxins is changed from green to light blue without decrease in intensity, while the dull fluores- cence of sterigmatocystin becomes more readily visible and that of interfering pigments is often reduced considerably. The fluorescence intensity of the aflatoxins is, however, decreased by this treatment. These colour changes have been shown to be reversible. CHOICE OF SOLVENT FOR EXTRACTION As a starting point in the development of the method the solvents known to be most suitable for aflatoxin extraction were examined, vix., chloroform (dampened sample), chloro- form - methanol and aqueous acetone.All three solvents were used to extract separate samples of the three media, which contained only ochratoxin or only sterigmatocystin. Each sample was macerated with the solvent in an explosion-proof blender for 3 minutes, followed by filtration and re-extraction with fresh solvent for 1 minute. The extraction efficiencies of the solvents were compared by thin-layer chromatography of the concentrated extracts. It was shown that the most effective extraction (more than 90 per cent.) occurred when chloroform - methanol (8 + 2) was used. It was, therefore, decided to use this solvent, with the addition of a small proportion of hexane to facilitate the removal of lipids with which the mycotoxins seem to be associated physically, for the investigation of mixed toxin extraction.Samples of each of the three media were compounded so that three composite samples were obtained with the following mycotoxin contents, expressed as pg kg-1 of final mixture- Sample No. A r \ Mycotoxin 1 2 3 Aflatoxin . . .. . . high (600) medium (50) low (10) Ochratoxin . . .. . . low (60) high (2000) medium (600) Sterigmatocystin . . . . medium (600) low (200) high (2000) These samples were analysed according to the following method and modifications. METHOD REAGENTS- All solvents should be of recognised analytical-reagent grade. Chloroform. Methanol. Hexane. Benzene.Light petroleum, boiling range 30" to 60" C. Diethyl ether, anhydrous. A cetic acid, glacial. Formic acid. Toluene. Ethyl acetate. n-Propanol. Trichloroet hylene. Sodium sulfhate, anhydrous powder. Potassium hydroxide, pellets. Silica gel-Suitable for column chromatography, 0.05 to 0.2 mm; and for chromatography, Macherey-Nagel GHR or Camag D-5. thin-layerFebruary, 19691 CEREALS AND GROUNDNUTS FOR THREE MYCOTOXINS 139 APPARATUS- Chromatographic columns-These were 22 x 300 mm, and fitted with Teflon stopcocks. Filter-paper, Whatman No. 12, 186 cm. Rotary evaporator. Blender, explosion-+roo f. Centrifuge, with 4 x 250 head. Thin-layer chromatographic apparatus. PREPARATION OF THIN-LAYER PLATES- Prepare several thin-layer chromatographic plates (10 x 20 cm) as described in the Official Method of the A.O.A.C.,1° with silica gel.Activate the plates for 90 minutes in an oven at 105" C after air-drying them for 30 minutes in a dust-free atmosphere. EXTRACTION OF SAMPLES- Extract 50g of finely ground sample with 200ml of solvent (chloroform-methanol- hexane, 8 + 2 + 1, v/v) for 3 minutes at high speed in an explosion-proof blender. Transfer the resulting suspension to a 250-ml centrifuge tube and centrifuge at 2000 r.p.m. for 5 minutes. Filter the supernatant liquid through fluted filter-paper. Break up the sediment in the tube with a glass rod and rinse into the blender flask with 100 ml of fresh solvent. Blend for 1 minute, repeat the centrifugation and filtration and combine the filtrates. Concentrate the clarified extract to 5 ml under slightly reduced pressure.PRELIMINARY THIN-LAYER CHROMATOGRAPHY- Spot lop1 of the concentrated extract and lop1 of each of the qualitative standard solutions on an imaginary line, 3 cm from the bottom of an activated chromatoplate. Develop the plate in an unequilibrated tank, containing toluene - ethyl acetate - 90 per cent. formic acid (5 + 4 + 1, v/v), to about 12 cm above the origin. Remove the plate from the tank, allow the solvent to evaporate and examine it under an ultraviolet source. The mycotoxins contained in the three reference standards should all be clearly resolved. Observe whether there are any fluorescent spots at the comparable RF values in the sample chromatogram. Spray the chromatoplate with 20 per cent. potassium hydroxide solution and immediately examine it again under ultraviolet light.Any samples that definitely do not contain detectable amounts of the mycotoxins can be eliminated at this stage, and the results can be reported as containing less than 4 pg kg-l of aflatoxin B,; less than 3 pg kg-l of aflatoxin B,; less than 20 pg kg-l of ochra- toxin A; less than 20 pg kg-l of ochratoxin C; and less than 100 pg kg-1 of sterigmatocystin. If a positive or doubtful result is obtained, the extract is processed as detailed below. COLUMN CLEAN-UP OF EXTRACT- Place a ball of glass-wool or non-absorbent cotton-wool loosely into position at the constriction of a chromatographic column, which is half filled with light petroleum. Drain off a portion of the light petroleum to ensure that all air bubbles are removed.Add about 5 g of anhydrous sodium sulphate to form an even base for 10 g of silica gel, which is subse- quently poured slowly into the column. Allow the silica to settle evenly by drawing off the light petroleum until the level is about 5 cm above the silica, and add about 10 g of anhydrous sodium sulphate. Draw off the light petroleum to just above the top of the sodium sulphate and transfer the concentrated sample extract quantitatively to the column, with the smallest volume of light petroleum required to effect the transfer. Drain the extract into the column and elute with a mixture of 75 ml of light petroleum and 25 ml of anhydrous diethyl ether. Adjust the flow-rate to about 15ml minute-'. Collect the eluate as Fraction I. Continue the elution of the column with 100d of chloroform- methanol (97 + 3, v/v), collecting the eluate as Fraction 11. Finally, elute with 100 ml of benzene - acetic acid (9 + 1, v/v).Allow the column to drain completely and collect the acidic eluate as Fraction 111. Evaporate the three fractions separately to dryness under reduced pressure in a rotary evaporator. Fraction I11 will be found to require increased vacuum, and a few millilitres of toluene added140 to the acetic acid residue will improve the distillation rate. Cool the flasks and immediately dissolve the residues in 5 ml of benzene. Stopper the flasks and store them in a dark cupboard while preparing for thin-layer chromatography. VORSTER: A METHOD FOR THE ANALYSIS OF [Analyst, VOl. 94 THIN-LAYER CHROMATOGRAPHY OF THE EXTRACT- The extent of dilution required to produce satisfactory results from the thin-layer chromatography of a positive sample extract can often be assessed as a result of the pre- liminary thin-layer chromatography.Much time can thus be saved by suitable dilution of the various fractions with benzene, before attempting the qtlantitative determination of the toxins. Follow the generally accepted procedure by spotting a series of aliquots from the sample extract and from the relevant standard solution calculated to produce fluorescent spots with intensities at, or just above, the limit of visual detectability. Develop the plate in any one of the following eluants (all proportions are v/v). For sterigrnatocystin (Fraction 1)- (i) chloroform - methanol, 98 + 2; (ii) toluene - ethyl acetate - formic acid (90 per cent.), 5 + 4 + 1; (iii) trichloroethylene - n-propanol - acetic acid, 90 + 9 + 1 ; (iv) benzene - n-propanol - acetic acid, 86 + 10 + 4.For the ajatoxins (Fraction II)- (i) chloroform - methanol, 96 + 4; (ii) chloroform - acetone, 9 + 1. (i) benzene - acetic acid, 9 + 1; (ii) toluene - ethyl acetate - formic acid (90 per cent.), 5 + 4 + 1. For the ochratoxins (Fraction III)-- Remove the plate from the tank, allow the solvent to evaporate and examine the fluorescent pattern on the plate under long wave ultraviolet light. Compare the fluorescent intensities caused by the toxin in the sample extract with those of the standard spots. If the sample appears to be negative for sterigmatocystin, spray the plate lightly with 20 per cent.potassium hydroxide solution. Re-examine the plate immediately. The sterigmato- cystin standard will appear as an intensely fluorescent yellow spot and the sample chromato- gram may also show up the presence of sterigmatocystin. The extracts from highly con- taminated samples may have to be diluted and re-chromatographed before the spots can be matched with the standards. Calculate the concentration of the toxins in the sample as follows- Ss x Cs x FV x x w pg of toxin per kg of sample = where Ss is the volume of standard solution spotted to give fluorescence equal to Xpl of sample extract, pl; Cs is the concentration of standard solution, pg ml-1; F V is the final volume of sample extract, pl; X is the volume of sample extract spotted to give fluorescence equal to Ss, 1-11; and W is the weight of sample extracted, g.PROBLEMS ENCOUNTERED IN APPLYING THE METHOD TO THE ANALYSIS OF GROUNDNUTS AND CERTAIN VARIETIES OF SORGHUM A significant percentage of the oil content of groundnuts is extracted by the solvent used for mycotoxin extraction. This oil causes problems when assaying for sterigmatocystin and the ochratoxins. It is impracticable to separate oil from sterigmatocystin by column chromatography. Davies , Kirkaldy and Roberts" proposed the purification of sterigmato- cystin from liquid culture extracts by separation on a column of heavy magnesium oxideFebruary, 19691 CEREALS AND GROUNDNUTS FOR THREE MYCOTOXINS 141 by elution with chloroform. However, no clear separation of oil and sterigmatocystin can be obtained by this procedure, and inconveniently large volumes of solvents are required to recover all of the sterigmatocystin. The use of 1 g of magnesium oxide above the silica in the column was tried, but complete recovery was not achieved.The effects of thin-layer chromatography of the extract, by using layers of magnesium oxide or alumina with various solvent systems, were determined, but again, no clear separation was obtained. In addition, it was found impossible to recover known amounts of ochratoxin from oily extracts with a silica column. It was subsequently determined that all fractions of the eluting solvents contained the mycotoxin. This is remarkable as, in the absence of oil, ochratoxin is known to migrate only in an acidic eluant.As a result of the problems encountered when attempting purification of the toxin extract by column chromatography, this approach was abandoned when appreciable propor- tions of fats were present, and it was reluctantly decided to resort to the tedious method of partitioning between various solvents. The dark-coloured pigments extracted from interfere with the detection of sterigmatocystin This was overcome by precipitating the pigments chromatography. MODIFICATIONS OF THE METHOD TO SAMPLES WITH HIGH F 4 T CONTENT- some varieties of sorghum were found to when the extract was chrornatographed. with diethyl ether before applying column DEAL WITH THESE PROBLEMS Prepare an extract as described in the method. Shake the clarified extract with 50 ml of 0.1 M sodium hydrogen carbonate in a separating funnel.Drain off the lower phase and re-extract it with a fresh 50-ml portion of sodium hydrogen carbonate solution. Drain off the chloroform layer ( A ) and retain it for further processing. Combine the aqueous phases, acidify with 2 M hydrochloric acid and extract three times in a separating funnel with 40-ml portions of chloroform. Filter the combined chloroform extracts through a bed of anhydrous sodium sulphate and wash the latter with 20ml of chloroform. Evaporate the filtrate to dryness and dissolve the residue in benzene. Transfer the solution quantitatively to a small vial and make up the volume to 5 ml (Fraction I). The above procedure should be carried out in subdued light and as quickly as possible as the ochratoxins are very susceptible to photolysis in aqueous solutions.Evaporate the chloroform solution ( A ) under reduced pressure and dissolve the oily residue in 100ml of 85 per cent. methanol. Transfer the solution to a separating funnel and rinse the flask with 50ml of hexane, adding the washings to the funnel. Shake the funnel thoroughly and, after separation of the layers, re-extract the methanol layer with a fresh portion of 50 ml of hexane. Retain the methanol layer while extracting the combined hexane portions with an equal volume of 85 per cent. methanol. This is done to recover any sterigmatocystin that may have dissolved in the hexane. Discard the hexane and add the methanol layer to the retained portion. Add water to the methanol phase to adjust the methanol content to about 50 per cent.Extract this solution three times with 40-ml portions of chloroform. Filter the combined extracts through anhydrous sodium sulphate, wash the filter bed with about 20 ml of fresh chloroform and evaporate the combined filtrates to 5 ml (Fraction 11). Proceed with thin-layer chromatography as described in the method, with Fraction I for the determination of ochratoxin A and Fraction I1 for that of sterigmatocystin and the aflatoxins. SAMPLES WITH DARK-COLOURED PIGMENTS- Concentrate the clarified extract to about 25 ml. Add 2 g of Hyflo Supercel, or similar filter aid, and 25ml of anhydrous diethyl ether. Continue the evaporation of the solvent until distillation ceases, and transfer the residual suspension to the column with a small volume of light petroleum. Proceed with the elution of the column as described.The first fraction, which contains the sterigmatocystin, should be free from dark-coloured pigments. The second fraction, containing the aflatoxins, may also contain some pigment, but this will not interfere with the detection of the aflatoxins.142 VORSTER RESULTS Table I shows the results of the analysis of the compounded samples according to the proposed method. TABLE I RESULTS OF MYCOTOXIN ANALYSIS OF COMPOUNDED SAMPLES Sample No. 1 2 3 1 2 3 1 2 3 Gvoundnuts- Sorghum- Aflatoxin content, Pg k c l - Calculated Found 600 500 50 45 10 10 500 480 60 38 10 8 500 500 50 40 10 10 Ochratoxin content, Pg kg-’ Calculated 50 2000 500 60 2000 500 50 2000 500 Found 46 1860 500 40 1760 460 42 1800 480 Sterigmatocystin content, Pg kg-l r Calculated Found 600 600 200 200 2000 1860 500 460 200 140 2000 1600 500 460 200 160 2000 1700 I thank Mr. M. Steyn for the mycotoxin standards, and Mrs. H. E. Pretorius for technical assistance. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. REFERENCES IUPAC Information Bulletin No. 31, March, 1968, Butterworth Scientific Publications, London, Steyn, P. S., and van der Merwe, I<. J., Nature, 1966, 418. Scott, P. M.. and Hand, T. B., J. Ass. 08. Andyt. Chem., 1967, 50, 366. Vorster, L. J., and Purchase, I. F. H., Analyst, 1968, 93, 694. Rodricks, J . V., Lustig, E., Campbell, A. D., and Stoloff, L., Tetrahedron Lett., 1968, 25, 2975. Aucamp, P. J., and Holzapfel, C. W., J. S. Afr. Chern. Inst., 1968, 21, 26. Stoloff, L., Beckwith, A. C., and Cushmac, M. E., J. Ass. 08. Analyt. Chem., 1968 51, 65. Van der Merwe, K. J., Steyn, P. S., and Fourie, L., J. Chem. SOC., 1966, 7083. S t e p , P. S., and Holzapfel, C. W., J. S. Afr. Chem. Inst., 1967, 20, 186. “Changesin Methods-26 Nuts and Nut Products,” J. Ass. Ofl. Agric. Chem., 1966,94, 230. Davies, J . E., Kirkaldy, D., and Roberts, J. C., J. Chem. SOC., 1960, 2173. p . 36. Received August 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400136

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Afialyst, February, 1969, Vol. 94, pp. 143-147 143 Determination of Mercury Residues in Potatoes, Grain and Animal Tissues Using Perchloric Acid Digestion BY N. A. SMART AND A. R. C . HILL (Plant Pathology Laboratory, Ministry of Agriculture, Fisheries and Food, Harpenden, Hertfordshire) A rapid method for determining mercury residues in potatoes, grain and animal tissues is presented. Digestion with nitric, sulphuric and per- chloric acids is used in place of the oxidation with nitric and sulphuric acids and hydrogen peroxide, which is the method recommended by the Joint Mercury Residues Panel. Recovery of added organomercurial is about 90 per cent. Six analyses can be carried out in 1 day by one worker. METHODS for determining mercury residues in vegetable and animal materials have recently been reviewed.l Four methods, which involve wet oxidation with nitric and sulphuric acids, and colorimetric determination of mercury with dithizone, have been recommended as official or standard method^.^^^,^^^ They differ in such details as the type of apparatus in which the wet oxidation is carried out and the purification of the initial dithizone extract.The methods are all lengthy; that of the Joint Mercury Residues Panel2 enables two to four determinations to be made in 1 day, depending on the availability of preceding or succeeding days to start or finish determinations. We have had considerable experience with this method, and felt that a shortening of the wet-oxidation stage would be advantageous. The use of perchloric acid should lead to a more complete oxidation than that of nitric and sulphuric acids done.We have, therefore, investigated the use of mixed nitric, sulphuric and perchloric acids in wet oxidation, in conjunction with the dithizone extraction, reversion and determination procedure. Gorsuch6 first found that mixed nitric, sulphuric and perchloric acids could be used in determining traces of mercury, although he did not propose any particular method for specific foodstuffs. His work showed that mercury compounds are more volatile when treated with perchloric acid than when digested with nitric and sulphuric acids alone, but that if efficient condensers are used, mercury should not be lost from the system. In 1960, the Analytical Methods Committee of the Society for Analytical Chemistry proposed the use of nitric, sulphuric and perchloric acids in wet oxidation of foodstuffs con- taining more than about 5 p.p.m.of mercury, observing that losses occurred when the method was used for lower levels of mercury.' This method is not suitable, therefore, for most residue determinations. Hordynska, Legatowa and Bernsteinasg followed the approach of Gorsuch in using wet oxidation with nitric, sulphuric and perchloric acids, with a trap into which lower-boiling fractions are distilled during the digestion, thus raising the oxidation potential of the digest in the flask. They used the method for determinations on dressed grain with 96.5 3.8 per cent. efficiency. The method also gave 97.7 per cent. recovery of phenylmercury 8-hydroxy quinolinate from apples.No details have been given for use of the method with other vegetables or fruits, or with animal materials. Ward and McHughlO developed a method for determining the mercury content of vegetation from 0-4 p.p.m. upwards, involving wet oxidation with nitric, sulphuric and perchloric acids, followed by quantitative determination with dithizone. In their method the perchloric acid is added initially, and this could lead to uncontrolled reaction of easily oxidised substances before they have been partially decomposed with nitric acid. Hydrogen peroxide is used to complete the oxidation of the vegetable material after digestion with the mixed acids. 0 SAC; Crown Copyright Reserved.144 SMART AND HILL: DETERMINATION OF MERCURY RESIDUES I N POTATOES, [ArtalySt, VOl.94 Epp+ has used digestion with mixed nitric and perchloric acids for determining mercury residues in rice. The method, however, is only sensitive to the nearest 0.1 p.p.m. of mercury. Addition of sulphuric acid is desirable to control the reaction rate and also to raise the effective concentration of the perchloric acid, thus giving a more complete oxidation.12 Kinoshita13 has used nitric, sulphuric and perchloric acids for determining milligram amounts of mercury compounds, but the method is not applicable to residue determinations. Following the suggestion of Gorsuch, independently of the above workers, we propose methods for determining mercury residues in potatoes, grain, eggs, hens’ muscle and lambs’ livers, based on established standard methods in which wet oxidation with nitric, sulphuric and perchloric acids (without hydrogen peroxide) is used, which are more rapid than the original methods.Fruits, vegetables and animal material that might contain mercury residues were investi- gated using the organomercury compounds likely to arise in commercial practice in the United Kingdom. Determinations were made at the order of residue levels that may arise, as judged by previous experience. Workers who use the method should familiarise themselves with the “Notes on Perchloric Acid and its Handling in Analytical Work,” published by the Analytical Methods Committee of the Society for Analytical Chemistry.14 EXPERIMENTAL Apart from wet oxidation of the sample and the reversion step for animal tissue, the method is the same as that recommended by the Joint Mercury Residues Panel.2 A 1-litre flask is preferred for the wet oxidation. REAGENTS- Mercury Residues Panel.The following reagents are required in addition to those listed in the report of the Joint Perchloric acid, 72 per cefit.-Analytical-reagent grade. For animal material the following are also required. Sodium nitrite, 5 per cent., aqueous. EDTA, disodium salt, 2.5 per cent., aqueous. Urea, 10 per cent., aqueous. Hydrogen peroxide is not needed. Organomercury compounds were added in 1 to 5 ml of acetone (ethanol was used with phenylmercury urea) to the foodstuff contained in the wet-oxidation flask to obtain recovery results . POTATOES, EDIBLE RICE AND SEED GRAIN- Potatoes should be cut into quarters, which are thoroughly mixed, an aliquot diced and a 50-g sub-sample taken for analysis.Representative samples of 25 g of edible rice or 10 g of seed grain are taken. Place the prepared sample in the reaction flask, together with a few glass beads, and mix with 0.1 g of selenium powder. Add 25 ml of water to the rice and 10 ml of water to the seed grain. Place the flask in the heating mantle and fit the condenser system (with water flowing rapidly through it) and the tap funnel. Add 25 ml of mixed nitric and sulphuric acids (1 + l), slowly and intermittently, over a period of 10 minutes, taking care that the mixture at no time froths appreciably. The contents of the flask should be swirled from time to time. Then add a further 10 to 20 nil of nitric acid for potatoes and seed grain, or 50 to 70 ml for rice (to prevent charring). Switch on the heating mantle and slowly increase the rate of heating for about 30 minutes.The reaction should not be allowed to become violent. When solids disappear, add 15ml of 72 per cent. perchloric acid and reflux for 1 hour at full heat. At the end of this time the reaction mixture should be almost colourless with potato, and pale yellow with rice and seed grains. Cool the digest and wash down the condenser with 50 ml of water. Proceed as described in the recommended method of the Joint Mercury Residues Panel (p. 613, line 9).February, 19691 ANIMAL TISSUE- Place 50g of a representative sample of avian or mammalian tissue in the reaction flask, together with a few glass beads, and mix with 0.1 g of selenium powder.Place the flask in the heating mantle and oxidise as above for potatoes and grain. An additional 20 ml of nitric acid are required for eggs and livers and 10 ml for muscle. The final digests are markedly yellow. Allow the cold, partly neutralised digest (see recommended method of the Joint Mercury Residues Panel) to stand with 10 ml of 20 per cent. hydroxylammonium chloride solution. Extract an aliquot, depending on the expected level of mercury, for 1 minute with 10 ml of stock dithizone solution and then twice each with 10 ml of dilute dithizone solution. Combine the dithizone extracts in a 100-ml separating funnel. Wash the combined extracts with 25 ml of 0.1 N hydrochloric acid and 5 ml of hydroxylammonium chloride for 1 minute. Run the dithizone layer into a third 100-ml separating funnel, together with 2 to 3 ml of chloroform used for washing the second separating funnel.Add 10 ml of 0.1 N hydrochloric acid and 1 ml of 5 per cent. aqueous sodium nitrite solution. Shake the mixture for 1 minute, separate and discard the chloroform layer. Wash the aqueous layer with 2 to 3ml of chloroform and discard the washing. Add 1 ml of 20 per cent. hydroxylammonium chloride and allow to stand for 15 minutes, with occasional shaking. Add 1 ml of 10 per cent. urea solution, 1 ml of 2.5 per cent. EDTA (disodium salt) solution and 10 ml of dilute dithizone solution. Shake the mixture for 1 minute. Allow the layers to separate and run the dithizone solution down a 0.8 x 3 to 4-cm column of tightly packed absorbent cotton-wool into the 4-cm spectrophotometer cell.Also run some of the aqueous layer on to the column to ensure that enough dithizone solution runs into the cell for the spectrophotometer beam to pass through the solution. Read the optical density at 490 nm and refer to the calibration graph. GRAIN AND ANIMAL TISSUES USING PERCHLORIC ACID DIGESTION 145 RESULTS POTATOES AND GRAIN- with the perchloric acid oxidation method are given in Table I. The recoveries of added organomercury compounds from potatoes, edible rice and barley TABLE I RECOVERY OF ADDED ORGANOMERCURY COMPOUNDS FROM POTATOES, RICE AND BARLEY Mercury, pg (net), recovered from r A \ phen y lmercury chloride phenylmercur y phen ylmercury ethylmercury p W G ) barley (10 g) barley (10 g) barley (10 g) 4-7 4.4 99 97 9.5 4.7 4.6 95 93 8.8 4.7 5-1 98 97 9.7 4.9 4.7 90 98 8.7 4.8 4-6 94 100 8.9 4.7 4.7 97 98 9.2 4.9 4.5 97 101 8-6 4.7 (5) 4.6 (5) 95 97 ( - 5 ) 8.9 (5) (5 pg of mercury) added to urea (100 pg of acetate (100 pg of chloride (10 pg of mercury) added to mercury) added to mercury (added to Mean- (96%) (93%) (9570) (974%) (89.5%) Standard deviation- f0.1 f 0.2 &3 zt3 f 0 - 3 Total reagent and crop blanks for determinations were 0.7 and 1.Opg of mercury for potatoes; 0.7, 0.6, 0.4 and 0.3 pg of mercury for rice; 0.1, 0.1, 0.0 and 0-5 pg of mercury for barley.ANIMAL TISSUE- liver by the perchloric acid oxidation method are given in Table 11. The recovery of added organomercury compounds from eggs, hens’ muscle and lambs’146 SMART AND HILL: DETERMINATION OF MERCURY RESIDUES IN POTATOES, [A’?Za&St?, VOl.94 TABLE I1 RECOVERY OF ADDED ORGANOMERCURY COMPOUNDS FROM EGGS, HENS’ MUSCLE AND LAMBS’ LIVERS Mercury, pg (net), recovered from phenylmercury acetate (5 pg of mercury) added to r eggs (50 g) muscle (50 g) liver (50 g) 4.1 4-65 4.4 4.5 4.3 4.2 4.1 4.3 4-6 4.4 4.1 4.3 4.2 4.1 4.1 4.0 4.2 4.5 4.2 4.2 4.3 4.0 (5) 4.3 4.4 (81 %) (85 % ) (88%) f 0.2 f 0.2 f0.2 A \ Mean- Standard deviation- ethylmercury chloride (5 pg of mercury) added to r - v eggs (50 g) muscle (50 g) 4-0 4.1 3.9 4.4 4.4 4.2 4.1 4.3 4.2 4.1 3.9 4.2 3.9 4.5 phen ylmercury chloride (6 pg of mercury) added to liver (50 g) 4.3 4.5 4.6 4.6 4.2 4.4 4.6 4.2 4.2 (84%) (85%) f0.2 f 0-2 4.3 (86%) f 0.2 Total reagent and crop blanks for determinations were 0.4 and 007pg of mercury for eggs; 0.6 and 0.7 pg of mercury for hens’ muscle; and 0-7 and 0.5 pg of mercury for lambs’ livers.POTATOES AND GRAIN- The modified perchloric acid digestion gave about 95 t 2 per cent. recoveries for this group of materials. In this country, phenylmercury chloride is the only organomercurial used on potatoes, and was consequently used to obtain recovery results. This compound is widely used to control rice blast (Piricularia oryzae) in many parts of the world, and was recovered from rice at the 0.1 p.p.m. level, as mercury residues in rice have been shown to be of this order of magnitude.l Organomercueals used as seed dressings on wheat and barley do not give rise to residues in harvested grain. However, it is sometimes necessary to check whether grain has been dressed with organomercurials and, if so, to what extent.The proposed method has, therefore, been tested for barley, using phenylmercury acetate, phenylmercury urea and ethylmercury chloride, the organomercurials most commonly used (in this country), as fungicidal seed dressings at the levels at which these compounds are usually present on dressed seed. Methoxyethylmercury silicate was not tested because of difficulties encountered in preparing a satisfactory standard solution for addition to the grain for recovery tests. The modified digestion shortens the method by about 3 hours, so that six analyses can by completed by one worker in 1 day. The method, therefore, takes about half the time required for the Panel’s original method.It is also shorter and considerably simpler, for potatoes and grain, than the method given by the Metallic Impurities in Organic Matter Sub-committee of the Analytical Methods Committee of the Society for Analytical Chemistry.s ANIMAL TISSUE- The modified perchloric acid wet oxidation, and determination of mercury with dithizone by the thiosulphate reversion technique, used for potatoes and grain was not satisfactory because of excessive oxidation of the dithizone. However, the nitrite reversion stage of the Metallic Impurities in Organic Matter Sub-committee of the Analytical Methods Committee method proved satisfactory in place of the thiosulphate reversion. The final dithizone extract is cleaned up on a short cotton-wool column; 85 & 2 per cent.recoveries were obtained for these types of avian and mammalian tissue. The organomercury compounds chosen for study with eggs and hens’ inuscle are those which would arise from poultry that eat mercury- dressed grain. Although not recommended as good agricultural practice, animals occasionally graze in sprayed orchards, and livers from sheep poisoned in this way are sometimes presented for analysis. DISCUSSIONFebruary, 19691 147 The nitrite reversion takes 16 to 30 minutes longer than the thiosulphate reversion step, but the modified method, as applied to avian and animal materials, is still appreciably shorter than the original recommended method. This modified method should only be applied to other materials with caution, and is not suggested for determining mercury residues in tomatoes and apples, for which the Panel’s original method is more suitable. GRAIN AND ANIMAL TISSUES USING PERCHLORIC ACID DIGESTION 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. REFERENCES Smart, N. A., in Gunther, F. A., Editor, “Residue Reviews,” Springer-Verlag, Berlin, Volume 23, 1968, p. 1. Report by the Joint Mercury Residues Panel of the Advisory Committee on Poisonous Substances Used in Agriculture and Food Storage, the Analytical Methods Committee and the Association of British Manufacturers of Agricultural Chemicals, Analyst, 1961, 86, 608. Analytical Methods Committee, Ibid., 1965, 90, 616. International Union of Pure and Applied Chemistry, Pure Appl. Chem., 1965, 10, 77. Association of Official Agricultural Chemists, J. Ass. 08. Agric. Chem., 1962, 35, 80. Gorsuch, T. T., Analyst, 1969, 84, 136. Analytical Methods Committee, Ibid., 1960, 85, 643. Hordynska, S., Legatowa, B., and Bernstein, I., Chemia Analit., 1962, 7 , 667. Legatowa, B., Hordynska, S., and Bernstein, I., Roczn. Pdnst. ZakI. Hig., 1963, 14, 221. Ward, F. N., and McHugh, J. B., Prof. Pap. U.S. Geol. Surv., No. BOlD, 128, 1964. Epps, E. A., J. Ass. Off. Analyt. Chem., 1866, 49, 793. Diehl, H., and Smith, G. F., Talanta, 1959, 2, 209. Kinoshita, S., Microchem. J., 1964, 8, 79. Analytical Methods Committee, Analyst, 1959, 84, 214. Received July 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400143

出版商:RSC    

年代:1969

数据来源: RSC