摘要:

344 SALT : MICROPHOTOMETRIC DETERMINATION OF v o l . 76 Microphotometric Determination of Carboxy- haemoglobin in Blood BY H. B. SALT (Presented at the meeting of the Society on Wednesday, February 7th, 1951) A simple method is described for the determination of carboxyhaemo- globin (HbCO) in the presence of oxyhaemoglobin (HbO,) in blood by means of a photo-electric absorptiometer with a narrow band light of peak trans- mission at 6000 A. The HbO, is estimated in terms of extinction increment brought about by reduction with sodium dithionite (Na$,O,). In this method, the HbCO remains unaltered. The total pigment in another portion of the sample is converted to HbO, and then estimated in a similar manner. The fractional proportions of HbO, and HbCCl are computed and converted into absolute values from a knowledge of the concentration of total haemoglobin. The analytical method will detect as little as 4 per cent.of HbCO and enable determinations of higher proportions of HbCO to be made with an accuracy within k2.5 per cent. DIFFERENCES in light absorption by oxyhaemoglobin (HbO,) and carboxyhaemoglobin (HbCO) have been used as a basis for elaborate spectrophotometric methods for estimating these substances when present together in blood.lS2 Hill3 has used similar principles in a method involving light absorption measurements a t different wavelengths by means of vacuum photo-cells and an electrometer. A spectrophotometric method for detection of HbCO in blood, described by Heilmeyerl, made use of the reduction of HbO, to Hb with sodium dithionite (Na$,O,) and measurement of light absorption values at three wave- lengths.In this procedure, HbCO was unaffected by sodium dithionite. Another method, quoted by Snell and Snell,, included the use of sodium dithionite and the determination of light absorptions at two wavelengths with a visual step-photometer. In the simple method now described, the necessary measurements are made with a photo-electric absorptiometer and the use of only one narrow waveband region, the pigments being differentiated by reduction with sodium dithionite. The photometer may be of any design, provided it allows the use of plane-parallel optical cells of 1.0 cm optical depth and 6.0 ml working capacity and is sufficiently sensitive to permit the use of narrow wavebands such as are selected by the Ilford series of spectral filters, which yield transmittance values of an accuracy within 0.5 per cent.DEVELOPMENT OF THE METHOD Light absorption measurements made by u$ng solutions of HbO,, HbCO and dithionite- reduced Hb, with each of the Ilford spectral colour filters Nos. 601 to 609, showed that the orange colour filter No. 607, of peak transmissilon ~ O O O A , provided a waveband of light that was absorbed by HbCO and by Hb with maximal differentiation, whereas the absorptions by HbCO and by HbO, were equal. It was found to be a further advantage to combine three thicknesses of heat-absorbing glass (Chance ON13) with the No. 607 colour filter, the effect being to absorb much of the red wavelengths otherwise partially transmitted by the orange filter and thereby narrow the selected waveband.This arrangement was additional to the one piece of heat-absorbing glass used ordinarily in the photometer; the resulting filter combination provided a narrow spectral orange waveband. The method depends on the principle that HbO, and HbCO have the same light extinction (E) values for narrow spectral orange light, whereas Hb has an E value 2-41 times as great. When a solution of blood containing HbO, and IElbCO, of extinction value Eg, is reduced with sodium dithionite, the extinction value increases to ER proportionately to the content of HbO, in the mixture. The component of EB due to the HbCO is unchanged by sodium The photometer described by Salt5 was used in all the experiments.June, 19511 CARBOXYHAEMOGLOBIN IN BLOOD 345 dithionite, so that the component due to the HbO, (EHbO,) thoretically may be derived from the relation- This relationship presupposes simple proportionality between concentrations of each of the three pigments concerned and the appropriate E values.In practice, as may be seen from Fig. 1, curvilinear relationships were found for HbO, and Hb, although simple proportionality EHbO, = (ER - E~)/(2.41 - 1.0) 0 3 C .- U E U Lu .- 01 Haemoglobin. mg per 6.0 ml Fig. 1. Relationship between haemoglobin, mg per 6-0 ml, and extinction values. Curve A, HbO,; curve B, HbCO; curve D, sodium dithionite (Na,S,O,) - reduced Hb; curve C, values in D minus values in A, i.e. (EHb - EHbO,) was demonstrated for HbCO. When ( E R - EB) values were plotted against the concentrations of HbO, from which they were derived, a simple proportionality was observed, although the constant in the equation differed slightly from the theoretical value. In practice, it was found that the results obtained from analyses of diluted blood containing varying pro- portions of HbO, and HbCO, when calculated on the basis of this relationship, were not consistently accurate.Attempts to establish an empirical relationship between EHbO, and (ER - EB) values for standard solutions containing known mixtures of HbO, and HbCO proved to be difficult. Saturation of whole blood with carbon monoxide (water-washed coal gas) before dilution was unsatisfactory; saturation of diluted blood with carbon monoxide left an excess dissolved in the water, which converted more pigment to HbCO when the solution was mixed with a solution of HbO,.It was shown experimentally that, when HbO, and HbCO standard solutions were prepared separately and their E values determined separately, the E values could be appropriately added together to yield extinctions corresponding to actual mixtures of HbO, and HbCO when present together in the same solutions, but this intricate method of standard- isation was inconvenient. Ultimately, it was found that when the HbO, present was determined in terms of extinction increment (ER - EB) brought about by reduction with sodium dithionite, and the total haem pigment, after initial conversion to HbO,, was determined similarly by means346 SALT : MICROPHOTOMETRIC IIETERMINATION OF Fol.76 of another equal sample, the ratio provided a good estimate of the fractional proportion of HbO, originally present. The HbCO fraction was derived by simple difference and the fractional proportions were converted into absolute values from a knowledge of the concentra- tion of total haem pigment. Since the HbO, content of a sample is related to the increment in extinction ( E R - EB) brought about by reduction with sodium dithionite, in a manner independent of the absolute quantity of total pigment present, this. quantity theoretically need not be controlled, but it was established experimentally that best results were obtained when 6.0 to 12.0mg of haem pigment were present in 6.0ml of solution. METHOD For ordinary blood specimens, prepare the sample by diluting 0.1 ml of blood to 12.0 ml with 0.007 N ammonia. For anaemic blood specimens of total haemoglobin content between 6.0 and 12-0 g per 100 ml, take double the quantity of blood; similarly, for severely anaemic specimens of total haemoglobin content below 6-0 E; per 100 ml, take four times the quantity of blood.These adjustments in sampling are not essential but will ensure maximum accuracy in the determination. Divide the sample into two portions of 6.0 ml each and transfer one portion to an optical cell. Determine the extinction value EB for the diluted blood at 1.0 cm optical depth, using the narrow spectral orange waveband. Add a few granules of solid sodium dithionite, but avoid an excess, which may cause a turbidity. Mix to dissolve the dithionite and determine the extinction value ER for the reduced blood sample.To ensure that reduction is complete, add a second small quantity of dithionite, still avoiding turbidity, mix and check that ER has the same value. Pass oxygen through the other portion of the diluted blood sample to convert all the pigment into HbO,. Complete conversion will be ensured by bubbling oxygen through the solution in daylight at room temperature for 30 minutes at the rate of about six bubbles per second from a glass jet 1-0 mm in diameter. This procedure uses nearly 2 litres of oxygen and is adequate for the quantity of haemoglobin chosen, even if all is initially present in the form of HbCO. Transfer the solution to an optical cell, determine the extinction value EHbO,, reduce the solution with sodium dithionite and determine the new extinction value EHb, in a manner similar to that described above.Calculate the oxyhaemoglobin in the sample a:; a fraction of the total pigment from the ratio- (ER - EB)/(EHb -- EHbO2) Subtract this value from unity to obtain the fractional proportion of carboxyhaemoglobin in the sample. If the total haemoglobin content of the blood is not previously known, determine the total pigment from the value EHb by means of a sta.ndard reference graph in which extinction values for known amounts of dithionite-reduced Hb are plotted against the concentrations (see Fig. 1, D). Calculate the absolute quantities, of HbO, and HbCO in the sample from the fractional proportions and the total pigment concentration. SENSITIVITY AND ACCURACY The method is capable of detecting with certainty as little as 4 per cent.of HbCO in the presence of 96 per cent. of HbO,. This compares favourably with other techniques currently used in clinical chemistry. Wikoff and Carsons have compared the sensitivity of twelve methods (none of which involves sodium dithionite or photometry) for the detection of carbon monoxide in human blood, with the conclusion that tests using tannic acid or p-yrotannic acid will detect 4 per cent. of HbCO, whereas the Hartridge reversion spectroscope will detect amounts only as low as 8 per cent. The remaining tests were still less sensitive. It has been shown by analyses of samples, in which the total haemoglobin varied between 6-0 and 12.0 mg and the proportion of HbCO varied from 4 to 100 per cent.that the percentage proportions of the two pigments can be estimated by the microphotometric method to an accuracy within +2*5 per cent.June, 19511 CARBOXYHAEMOGLOBIN I N BLOOD 347 REFERENCES 1. 2. 3. 4. 5. 6. THE BIOCHEMICAL LABORATORY THE ROYAL INFIRMARY Heilmeyer, L., “Spectrophotometry in Medicine,” translated by Jordan, A., and Tippell, T. L., Horecker, B. L., and Brackett, F. S., J . Biol. Cham., 1944, 152, 609. Hill, D. K., J . Physiol., 1940, 98, 2 3 ~ . Snell, F. D., and Snell, C. T., “Colorimetric Methods of Analysis,” Vol. 11, 3rd Edition, D. Van Nostrand Company, Inc., New York, 1949, p. 841. Salt, H. B., J . Lab. Clin. Med., 1950, 35, 976. Wikoff, H. L., and Carson, G. B., Amer. J . Clin. Path., 1948, 18, 648. Adam Hilger Ltd., London, 1943, pp.91, 94. WORCESTER DISCUSSION MR. P. J. GOODE asked whether it was considered necessary to take any special precautions to prevent oxygenation of the blood before taking the first extinction measurement. MR. SALT replied that if the blood sample, which need not exceed 0.1 ml, was taken by finger-prick and diluted immediately, the solution need only be kept tightly stoppered and in the dark, until the analysis can be performed. Venous blood was better collected with sodium citrate as anti-coagulant rather than with potassium oxalafe, other precautions being as before. As carbon monoxide has an affinity for haemoglobin about 210 times as great as the affinity of oxygen, there was very little chance of loss during the brief period occupied by the subsequent absorptiometric analysis.MR. T. P. WHITEHEAD asked whether the determination as described could be made with a Spekker absorptiometer and the appropriate filters, or even with the single-cell type of colorimeter. MR. SALT replied that the determination could certainly be made with the modern form of Spekker absorptiometer, or with any single- or double-cell instrument, in which the Ilford series of spectral filters and accurate plane-parallel optical cells were used. DR. K. A. WILLIAMS drew attention to the practice of using round test tubes instead of plane parallel- sided cells for absorptiometric work. This practice had been advocated on the grounds of economy but led to inaccurate results, and was therefore to be deplored. He was pleased to see that Mr. Salt’s economically constructed apparatus had not followed this trend, but made use of a properly constructed photometric cell. MR. SALT agreed, and added that i t had become evident in his personal experience that the greater cost of optical cells as compared with that of round test tubes was more than justified by the gain in accuracy in all absorptiometric methods. There was also the advantage that optical data and not merely arbitrary “colour comparisons’’ could be recorded. The present micro-procedure for carboxyhaemoglobin required optically accurate cells, and i t would be helpful if manufacturers could be discouraged from designing instruments that did not possess the degree of precision demanded by modern analytical methods.

ISSN:0003-2654    

DOI:10.1039/AN9517600344

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

348 SMALES: DETERMINATION OF STRONTIUM IN SEA WATER BY [Vol. 76 The Determination of Strontium in Sea Water by a Combination of Flame Photometry and Radio chemis t r y BY A. A. SMALES (Presented at the meeting of the Physical Methods Groztp on Tuesday, January 3rd, 1950) A novel combination of flame spectrophotometry and radiochemistry is described and applied to the solution of an old problem, namely, the deter- mination of strontium in sea water. A measured amount of strontium chloride containing 89Sr is added to the sea water and the fraction of the total strontium that is co-precipitated with barium sulphate on addition of barium chloride is measured by the decrease in the activity of the filtrate and also by estimating, by means of flame spectrophotometry, the amount of strontium that must be added to the filtrate in order to re-adjust the strontium emission to what i t was before the precipitation by barium chloride. The original strontium content of the sea water can then be calculated by means of a simple equation.Some evidence is also presented on the effect of the other constituents of sea water on the strontium emission in a propane - oxygen flame. IN spite of a large volume of literature on the chemistry of sea water,l including much on strontium, few references appear to have been made to the determination of this element in sea water, and even fewer give details of the method used. Desgrez and Meunier,2 by crystallisation of calcium and strontium sulphates on evaporation of the sea water, followed by spectrographic determination of the strontium in the dried material, found 13.5 mg per litre in a sample from the English Channel.This figure was confirmed by Thomas and Thompson (reported by Thompson and Robinson3) in Pacific Ocean water. Ramage4 reported 40 to 50 mg per litre, but this has been generally regarded as high and ignored in the literature. N ~ l l , ~ who used X-ray spectroscopy, gives a value of 6.2 mg per litre, Miyakes found 14.4 mg per litre by a gravimetric method and Vinogradov7 gives a value of 8 mg per litre. There is no definite evidence to show whether the proportion of strontium varies widely with position, although on general grounds such variations would not be expected; it seems more probable that differences in the analytical methods lead to the apparent variations. When a request was made for a determination.of strontium in the sea water used in some fish experiments being carried out by the Ministry of Agriculture and Fisheries Research Laboratory at Lowestoft, the method of Desgrez and Meunier2 was tried, but it proved extremely difficult to obtain consistent results with it. An alternative method was therefore sought and, as the generally accepted figure of Sverdrup,l 13 mg per litre, should be within the sensitive range of direct flame photometry, it seemed likely that such a method would be applicable. The work described shows how this method was applied, and how the difficulty of obtaining true standards for comparison was overcome by the use of radiochemistry. It is of a preliminary nature and may require further experimental substantiation before its validity can be firmly established.APPARATUS AND REAGENTS The Beckman flame photometer attachment with the model DU spectrophotometer was used; as this is fully described in the manufacturers’ catalogue, it needs no detailed explanation. It is worth noting, however, that the horizontal pre-heating chamber ensures that only a dry dust stream enters the flame, which in this work was oxygen-propane, while the dipping jet assembly permits rapid replacement of the sample. Since radio-active solutions (although extremely weak) were used, an extension chimney consisting of 3-inch piping was fitted to the top of the existing chimney and connected with the laboratory fume exit. No further description of the instrument is necessary, except that the flame is used asJune, 19511 COMBINATION OF FLAME PHOTOMETRY AND RADIOCHEMISTRY 349 the light source in place of the tungsten or hydrogen lamp that is used with the Beckman monochromator and that the sensitivity and time constant are increased by the replacement of a 2000-megohm input load resistor in the photo-tube by one of 10,000 megohms.The instrument is used as for absorption spectrophotometry, i.e., as a null instrument, the transmission-scale variable resistance being used to balance the photo-cell output. Radio-active strontium chloride solution was prepared by irradiating strontium carbonate with slow neutrons in the Hamell pile and dissolving it in the minimum of hydrochloric acid, the solution being allowed to stand for several days before use in order to allow short-lived radio-active components to decay.A P-ray absorption and decay curve determined with the aid of an end-window Geiger counter demonstrated effectively the radiochemical purity of the product (ssSr~+SgSr; half life 55 days, maximum P-energy 1.5 MeV). Although almost carrier-free tracer strontium might be desirable, this is only conveniently obtained from fission products and contains a major proportion of 90Sr, which, because of its long half life, has a much lower health tolerance value; in addition the presence of the daughter 90Y would be a disadvantage forathis work. Liquid counting was carried out with a glass- walled (35 mg per sq. cm) Geiger counter with 10-ml samples and the usual power pack, scaler and timing units.FLAME PHOTOMETRY- Oxygen and air pressure-Preliminary experiments with a solution of inactive strontium chloride (10 pg of strontium per ml) established that an oxygen pressure of 35 inches of water, an air pressure of 12 lb per sq. in. and sufficient propane to give a flame approximately mid- way between the explosive and the luminous gave suitable conditions, i.e., reasonable stability and sensitivity for strontium. The air pressure necessary varies, of course, with the particular jet used; in these experiments 12 lb per sq. in. was not quite enough to fully saturate the flame. Gas pressure-one minor practical disadvantage of the Beckman flame photometer is that the gas pressure can only be read on the manometer provided before turning on the oxygen supply, and so no visible control of gas is available during operation except the rather coarse one of flame appearance. This means that standard conditions are difficult to reproduce from day to day; as recommended by the makers of the photometer, it was found essential each time the apparatus was used to carry out tests with standards as well as with the samples.One possible way that has been used to ensure approximately standard conditions is toset the air and oxygen pressures and then, while feeding a standard solution into the flame, to adjust the gas pressure to give a chosen reading on the transmission scale. Wavelength setting-The exact position of the wavelength control setting is best found by slowly turning this control in the appropriate range while the flame is being fed with a solution containing the element under investigation; the slit should, for preference, be narrow. The position of maximum intensity is noted.In this work the strontium line used was at 4607 A, but the wavelength setting is so acute that visual setting on this wave- length, even with a calibrated wavelength scale, is useless and the procedure described above must always be followed. Slit width-For the work described, the instrument was used with all controls at maximum sensitivity except for the slit width, which was usually set a t a value between 0.15 and 0-05 mm. As the slit width is increased the flame background increases (as measured with distilled water feeding into the flame) and a compromise between sensitivity and selectivity is necessary.For example, at a slit width of 0.05 mm very little radiation other than that of the desired strontium and, of course, the continuous flame background, is apparent, while at 0-15 mm stray radiation is definitely noticeable when sea water is being used, but sensitivity is, of course, greater. The term “sensitivity” is used in this paper in the sense of diference in reading between the emission from a sample and the flame background, rather than as the ratio of these readings. EFFECT ON STRONTIUM EMISSION OF OTHER ELEMENTS IN SEA WATER A number of authors have discussed the possible sources of error in flame photometry; a summary of this work is given by Leyton.s The usual procedure adopted to avoid these errors is to calibrate the photometer with solutions approximating in composition to that under test.350 SMALES: DETERMINATION O F STRONTIUM IN SEA WATER BY r o l .76 Accordingly, an artificial sea water was prepared as described by Lyman and FlemingJg except that strontium was omitted, and with this as a base further solutions containing known amounts of strontium were prepared in order to plot a calibration curve relating photometer reading to strontium content. However, little useful infomation was gained by this method because the artificial sea water gave a much greater photometer reading than did the distilled water flame back- ground. The greater reading might have been du.e either to stray radiation, e.g., from the large amount of sodium or calcium, or to strontium impurity in the reagents used for the artificial sea water.If stray radiation was the cause, the sample would presumably behave in the same way, i.e., the high reading would be unimportant; if the reagents were a t fault, accurate determination in the sample would not be possible, owing to uncertainty in the position of the intercept on the strontium-content axis of the standard curve. There were two possible ways of investigating this problem: (a) by comparing the photo- meter reading for the artificial sea water on the strontium 4607 A line with the reading on either side of the strontium line; when this was done the comparison clearly showed the presence of some strontium in the artificial sea water; (b) by treating the artificial sea water, slightly acidified with hydrochloric acid, with barium chloride solution in order to co- precipitate at least some of the strontium; the filixate, which would remain essentially the same in terms of major constituents, might then be used as the base solution.(In fact, the concentration of chloride in the filtrate would very slightly increase and that of sulphate would very slightly decrease.) This procedure also showed that some strontium was originally Strontium, p.p.m. Strontium, p.p.m. Fig. 1. Graph showing effect of sodium and magnesium on strontium emission (slit width 0.05 mm) present as an impurity, as the photometer reading for the filtrate was lower than before the precipitation, although still higher than a distilled water flame background. Hence some uncertainty remained, although a possible way out of the difficulty would be repeated co- precipitation of strontium on barium sulphate.Before this means of purification could be accepted as sound, however, it was essential to prove that the strontium flame emission was independent of small changes in acidity and in concentration of chloride, sulphate and possibly calcium. It was felt that it might also be useful to consider the effect of the other major constituents of sea water simultaneously, since one of the possible sources of error mentioned by Leytong is the mutual interference between elements in the flame. It can be seen from Fig. 1 that there is some diminution of the strontium emission by sodium and'magnesium when present in amounts similar to those in sea water, but small changes in acidity and in concentration of chloride, sulphate and calcium were shown to be without effect. No real evidence of true increased strontium emission was found, although Fig. 2 shows an apparent increase, which, however, is shown to be a stray light effect when allowance is made for the photometer reading with no strontium added. The preparation of a strontium-free artificial sea water was then attempted by adding slowly to 250ml of the boiling, acidified, artificial sea water 50mg of barium chloride (BaC12.2H20) dissolved in a small quantity of wa.ter and then allowing the precipitate toJune, 19511 COMBINATION OF FLAME PHOTOMETRY AND RADIOCHEMISTRY 351 stand at least overnight before the barium sulphate was filtered off.Separate radiochemical experiments indicated that up to 70 per cent.of the strontium could be co-precipitated in this manner from an artificial sea water, but that yields might occasionally be only 10 per cent., particularly if the barium chloride was added quickly. With this solution, after two barium sulphate precipitations, the strontium content of the sea water sample was indicated as 10 p.p.m. (mg per litre), as shown in Table I. TABLE I STRONTIUM IN SEA WATER DETERMINED BY COMPARISON WITH THE PURIFIED BASE SOLUTION Photometer reading Artificial sea water, after two BaSO, precipitations- No added strontium .. .. .. .. 8.2 5 p.p.m. of strontium added . . .. .. 9.7 10 p.p.m. of strontium added . . .. . , 12.1 Sea water sample . . .. .. .. .. . . 12-2 Sea water sample, after two BaSO, precipitations .. 8.6 Nevertheless no real knowledge of the number of precipitations necessary to remove strontium completely could be obtained other than by using the tracer strontium in the artificial sea water itself, and even then at least two and often more precipitations were necessary. In addition there was always the possibility that the constitution of the real sea water samples was not completely identical with that of the artificially prepared sample, although this was guarded against to some extent, as shown in Table I, by precipitating barium sulphate from the sample of real sea water. Further consideration of the possible uses of tracer strontium suggested a much simpler method of overcoming the difficulty. COMBINATION OF THE TWO TECHNIQUES OF FLAME PHOTOMETRY AND RADIOCHEMISTRY As mentioned earlier, the partial purification of artificial sea water by co-precipitation of strontium on barium sulphate gives a filtrate that, from the flame photometry aspect, is essentially the same as the original, except in strontium content. If real sea water is similarly treated with barium chloride, part of the strontium is removed, but in other respects the filtrate is still essentially sea water.Now it is possible to measure the fraction of the strontium so removed, ag., by use of a radio-active tracer. If the filtrate after partial precipitation of strontium is examined in the flame photometer it will, of course, give a lower reading at the strontium line than did the original sea water. But, by adding weighed amounts of strontium to the filtrate until the reading is the same as that given by the original sea water, a relationship can be found between the fraction of the original strontium removed by precipitation and the equivalent mass of strontium; a simple calculation then gives the original concentration of strontium in the sea water.In other words the flame photometer is simply used as a null instrument to overcome the difficulties of the presence of strontium in the reagents, mutual interference and differences in surface tension and viscosity between standard and sample. Let the original sea water contain x p g of strontium per unit volume. To this add weightless tracer strontium without altering the volume. Then remove a fraction p of the total strontium by co-precipitation on barium sulphate, again without altering the volume. Suppose a pg of strontium per unit volume must now be added to the filtrate in order to restore the reading on the flame photometer to that found for the original sea water.Then obviously px = a, and since p is measured by radiochemistry and a is measured by flame photometry, x can be calculated. In practice, although weightless strontium tracer, i.e., 90Sr, half life 20 years,1° obtained from fission products, may be used, as pointed out earlier it was decided to use 89Sr in pre- ference. Since this is obtained by slow neutron irradiation of inactive strontium, the tracer solution contains a significant mass of strontium and this necessitates a slight extension of the mathematical treatment but does not invalidate the simple argument presented above ; in fact it makes possible an additional arithmetical relationship that serves as a useful check.352 SMALES: DETERMINATION OF STRONTIUM I N SEA WATER BY [Vol.76 Consider a volume of sea water containing x pg of strontium per 25 ml, to which y pug of strontium per 25ml in the form of a radio-active Strontium chloride solution is added without sensibly affecting the volume. Carry out a barium sulphate precipitation on part of this mixed solution after slightly acidifying it with hydrochloric acid, allow the precipitate, barium sulphate and part of the strontium as strontium sulphate, to settle overnight, and then filter. Measure the activity of the original solution and the filtrate, under comparable conditions, in a liquid counter and express the fraction of the activity removed as 9; on the assumption that complete exchange has taken place, this will also be the fraction of total strontium removed.Now to 25-ml portions of the filtrate, add various increasing measured amounts of strontium, again without sensibly affecting the volume, so that the solutions cover the range DETAILS OF METHOD- - - -. MgCI, strontium -Pure s t ron ti um Strontium, p.p.m. Strontium, p.p.m. (4 (b) Fig. 2. Effect of sodium, magnesium and calcium on stronium emission (slit width 0.15 mm). Graph ( a ) , absolute readings; graph ( b ) , corrected for reading of solution containing no strontium of strontium content from that present in the filtrate to above that of the solution before precipitation, and take the flame photometer readings, under standard conditions, of these solutions together with readings for (i) the enriched sea water before precipitation, and (ii) the original sea water.Plot a graph showing the relation between the number of micrograms of strontium added per 25ml and the flame phckometer readings. From this graph find the amount of strontium that must be added to 26 ml of the filtrate to give the same photo- meter reading as the “enriched” sea water, say apg. Then we have- P k 3 - Y ) = a a z.e., x = - - y P and since a, p and y are known, x may be calculated. The additional arithmetical relationship mentioned earlier may be obtained from the extra measurement so far unused, i.e., the flame photometer reading for the original sea water. Dependent on the extent of co-precipitation of the strontium with the barium sulphate, i.e., the value of fraction p , and on the amount of added strontium, y , the final strontium content of a 25-m1 portion of the filtrate, after pi-ecipitation, may either be greater or less than that of a similar volume of original sea water.If it is less, as is usual, then the difference can be found from -the graph plotted as explained above, since flame photometer readings of both were recorded. The strontium content of the filtrate can also be expressed as a fraction of the original “enriched” content; since a fraction p was removed then (1 - p ) must remain, i.e., (1 .- p ) ( x + y ) is the strontium content of the filtrate, in units of pg per 25ml. The strontium content of the original sea water is x pg per 25 ml and the difference between these is equal to 6.Thus we have the relationship- Call this difference b p g . x - (1 - P)(x +. y ) = b b + YP - $1 P or x =June, 19511 COMBINATION OF FLAME PHOTOMETRY AND RADIOCHEMISTRY 353 I t is possible for b to be either zero or negative, e.g., when there is more strontium in the filtrate than in the original sea water and the value of b must be obtained by extrapolation, but provided the correct value and sign are included in the formula this does not affect the argument. The procedure depends on achieving complete exchange between added tracer strontium and the inactive strontium present in the sea water. On general grounds this would appear to be ensured as strontium in both solutions is likely to be present as Sr” ions.To confirm Fig. 3. Reading for synthetic solutions containing 50 p.p.m. of strontium. Graph (a), Sr* added, Graph (b) Sr* y = 320 pg per 25 ml; fraction of activity removed, p = 13/18; slit width 0.07 mm. added, y = 640 pg per 25 ml; fraction of activity removed, p = 517; slit width 0.08 mm that this exchange does in fact occur, a synthetic solution was prepared containing the major elements of sea water, 2.3 per cent. of sodium chloride, 0-5 per cent. of magnesium chloride and 0.4 per cent. of sodium sulphate, together with 50 p.p.m. of strontium and allowed to “age” to simulate as closely as possible the conditions of a sea water. This solution was examined together with two samples of North Sea water, collected by the Fisheries Research 300 W W 200 ._ Y e x 40 A Fig.3. Photometer reading Lowestoft sea water (1). Sr* added, y = 320 pg per 25 ml; fraction of activity removed, p = 49/62; slit width 0.15 mm Laboratory at Lowestoft, and three samples of Atlantic Ocean water from depths of 600, 1200 and 2000 fathoms. After addition of the tracer strontium the solutions were allowed to stand for some hours before barium sulphate was precipitated by the dropwise additionSMALES: DETERMINATION OF STRONTIUM IN SEA WATER BY vol. 76 354 to the boiling solution of a small amount of hydrochloric acid (1 to 2 ml of 10 N acid per 250 ml of solution) followed by 20 mg of barium chloride, BaC1,.2H20, dissolved in 5 ml of water. Boiling was continued until the original. volume was reached and the precipitate was allowed to stand overnight before filtering through a Whatman No.40 filter-paper. RESULTS- The radiochemical evaluation of $ is recorded in Table 11; an example of the flame photometric data is shown in Figs. 3 and 4, and all the values of a and b obtained from similar graphs, together with the calculated value of x , alre recorded in Table 111. TABLE I1 RADIOCHEMICAL EVALUATION OF THE FRACTION OF STRONTIUM PRECIPITATED, @ Water Synthetic solution A . . Synthetic solution B . . Lowestoft 1 . . I . .. Lowestoft 2 . . .. Atlantic, 600 fathoms . . Atlantic, 1200 fathoms . . Atlantic, 2000 fathoms . . Sr* added, y , pg per 25 ml 320 640 320 320 320 320 320 Activity before precipitation, corrected for background (-101 c.p.m.), counts per 10 min.per 10 ml 3600 & 2% 7000 6200 6000 5600 5650 5660 Activity after precipitation, corrected for background, counts per 10 min. per 10 ml 2000 1300 2400 1600 2000 1500 1000 & 3% Fraction of strontium removed , P 13/18 49/62 5/7 3/5 5/7 365/565 415/565 TABLE I:[I FLAME PHOTOMETRIC EVALUATION OF FACTORS a AND b Values of x calculated Values taken from from equations graphs I - x=E- b + Y ( l -P) P Synthetic solution A . . 1160 840 1280 1290 Synthetic solution B . . 1350 750 1250 1300 Lowestoft 1 . . .. 445 115 245 220 Lowestoft 2 . . .. 340 36 245 270 Atlantic, 600 fathoms. . 390 70 226 224 Atlantic, 1200 fathoms 360 40 238 240 Atlantic, 2000 fathoms 435 80 270 225 p Y x= Water a, b, pg per 26 ml pg per 25 ml pg per 25 ml pg per 25 ml Strontium content found 7 from a, p.p.m.51 50 10 10 9 10 11 h----7 from b, p.p.m. 52 52 9 11 9 10 9 It is not claimed that the method is highly accurate, but the difference between the previously accepted figure of 13 to 13-5 p.p.m. and the figure of 10 p.p.m. obtained by the method described seems to be significant. From the flame photometry or spectrochemical viewpoint this technique of combination with radiochemistry to ensure similarity between sample and standards has obvious applica- tions when a part of the element to be determined may be removed without essentially altering the composition of the solution so far as major elements are concerned. This differential or null method of overcoming difficulties caused by impurities might also be extended to other elements. The assistance of Mr. W. T. Fullwood in much of the experimental work and of Mr. G. M. Spooner of the Marine Biological Laboratory, Plymouth, in bringing to the notice of the author some of the less easily accessible literature references is acknowledged. Thanks are due to the Director, A.E.R.E., for permission to publish.June, 1951 J COMBINATION OF FLAME PHOTOMETRY AND RADIOCHEMISTRY 355 REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10, Sverdrup, H. U., Johnson, M. W., and Fleming, R. H., “The Oceans,” Prentice-Hall Inc., New Desgrez, A., and Meunier, J., Compt. Rend., 1926, 183, 689. Thompson, T. G., and Robinson, R., Bull. Nut. Res. Council, 1932, No. 85, p. 174. Ramage, H. S. E., Naturalist, 1933, 38, 54. Noll, W., Chemie der Erde, 1934, 8, 507. Miyake, Y., Bull. Chem. SOC. Japan, 1939, 14, 55. Vinogradov, A. P., Us$. Chim., 1938, 7, 645-686. Leyton, L., “Annual Reports,’’ The Chemical Society, London, 1949. Lyman, J., and Fleming, R. H., J . Marine Res., 1940, 3, 13P46. Powers, R. I., and Voigt. A. F., Phys. Rev., 1950, 79, 176. York, 1946. ATOMIC ENERGY RESEARCH ESTABLISHMENT, HARWELL, BERKS.

ISSN:0003-2654    

DOI:10.1039/AN9517600348

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

June, 1951 J COMBINATION OF FLAME PHOTOMETRY AND RADIOCHEMISTRY 355 The Determination of Potash by Measurement of its Radio-Ac t ivi t y BY H. N. WILSON, D. S. LEES AND W. BROOMFIELD (Presented at the meeting of the Physical Methods Group on Tuesday, May 23rd, 1950) A method is described for the determination of potassium in potash salts and fertilisers by measurement of the radio-activity of powdered samples. A description of the apparatus and its calibration, together with suggestions for checking its reproducibility, is included. As no preliminary preparation is required other than grinding the samples, the method is rapid. The results, although they are not of the highest order of accuracy, can be obtained in from one t o two hours. The coefficient of variation (S.D. expressed as a percentage of the mean) for pure potassium chloride is 1.45 and the standard deviation of the Geiger count from the results of chemical analysis on 60 routine samples of fertiliser, ranging from 14.5 to 16.0 per cent. of K,O, was 0-28 with a coefficient of variation of 1-84.The method has proved of value in works process control but, as some fertilisers contain radio-active elements other than potassium, it must be applied with caution. IN 1943 a description appeared1 of a method for determining potash in fertiliser by measuring the radio-activity. The fertiliser was brought into solution, and introduced into an annulus, between an outside jacket and a very large Geiger - Muller tube, made of glass, connected to a counting circuit of the usual type.We, therefore, made enquiries as to availability of equipment, but found that suitable Geiger tubes were not to be obtained in this country. Although potassium is y-active as well as /%active, the radio-activity is very weak and a tube of large area was required. We later discussed the subject with Professor F. Paneth and members of his staff at Durham, who suggested that the radiation from the solid should give a measure of the K,O content, provided that the distance from the solid to the window of the Geiger tube remained constant, that the area of solid was sufficiently large and that the solid was in a layer thicker than the “critical depth.” An ordinary Geiger tube, mounted just above the sample, in a lead castle of conventional type, could then be used for measuring the radiation.Experi- ments, with carefully prepared mixtures of pure potassium chloride, ammonium sulphate and ammonium phosphate, proved that this was so. While this work was in progress a description appeared2 of a scheme similar to ours, but rather less convenient. The power pack and counting circuit have been modified as experience has been gained; Geiger tubes of 2 inches in diameter have been specially made for us by the General Electric Company to replace the original 1 inch tubes; these increase the rate of counting fourfold. The sample holder also has been modified. The apparatus is fully described below. COUNTING RATES AND “BACKGROUND”- Even in the absence of the sample, and inside the lead “castle,” which keeps out much stray radiation, the Geiger tube will register a certain number of “counts.” These are due356 WILSON , LEES AND BROOMFIELD : THE DETERMINATION OF [Vol.76 partly to radio-active dust (always present) in the laboratory atmosphere and partly to cosmic rays. These background counts cannot be distinguished from the true radio-active counts and a blank count must be made from time to time, which is to be deducted from the gross count, to obtain the net count due to radio-activity in the sample. Unfortunately the rate of the background counts is not constant (it was not found practicable to devise a coincidence circuit to eliminate the background), but varies somewhat from minute to minute, and also exhibits a (smaller) long-term variation so that background in successive weeks might well vary from 34 to 40 counts per minute.This uncontrollable random variation limits the precision of the method and it has been found3 that the standard deviation is equal to the square root of the number of counts. Thus, if the total number of counts was 25, the deviation would be 5, that is 20 per cent. of the number of counts; if it was 100, the deviation would be 10 = 10 per cent. and for 10,000 it would be 100 or 1 per cent. of the number of counts. I t is, therefore, clear that although accuracy increases with the number of counts, one soon reaches the region where, for small gains in accuracy, the time consumed is increased intoler- ably, e.g., to obtain a standard deviation of 0.5 per cent., one would need 40,000 counts, and for 0.2 per cent. deviation, 250,000.It has been our aim to approach a standard deviation of 0.5 per cent., which is about as good as that of chemical analysis of fertilisers. There are other uncontrollable sources of error, in particula:r, traces of other radio-active substances present in some fertilisers, which make it improbable that it is worth while trying to improve much upon this accuracy, which is adequate for many purposes, i.e., a net count of more than 10,000, preferably 40,000. With a 1-inch Geiger tube in our apparatus, pure potassium chloride gave a gross count of about 107 counts per minute, the background being 8 counts per minute. A 2-inch tube gave figures of about 590 per minute and 34 per minute respectively. Pure potassium chloride is equivalent to 63.1 per cent.of K,O, so that to obtain 10,000 counts from 40 per cent, potash salts, one would require with the 2-inch tube to count for half an hour; a mixed fertiliser containing 10 per cent. of K,O would require 2 hours, this compares favourably with the quickest chemical method of comparable accuracy, which takes two and a half hours. But as no weighing or chemical manipulation is required the time spent in actual manipulation is only about 5 minutes per sample. APPLICABILITY OF THE METHOD- The method is more restricted than was expected, because of the widespread presence of other radio-active elements in many fertilisers. It is applicable to fertilisers of the type made at Billingham, but not to fertilisers containing superphosphate made from rock con- taining radio-active elements, and it should only be used after most careful investigation, In the proved absence of excessive amounts of extraneous radio-active elements it gives speedy results of moderate accuracy, and in our hands has proved its worth.The electronic part of the apparatus has, however, not been quite trouble-free and it would be necessary to improve its reliability before absolutely reliable operation in a plant routine testing laboratory can be expected. SIZE OF SAMPLE- It has been found most convenient to have the sample in a tray with a recess 5 cm in diameter and 4 mm deep; the tray is accurately positioned in the castle, perpendicularly below the window (4.3 cm diameter) of the Geiger tube. This area is large enough to fill the window of the counter and the depth is in effect “infinite.” The maximum range of /?-particles in potassium chloride is 0.35 g per sq.cm. This means that a depth of greater than 2 mm (critical depth) is sufficient. Any ordinary fertiliser material will be within the range of 4mm, the depth of the sample cell. The sample should be more or less dry and should be ground to about 60-mesh size, as is usual for analytical samples. APPARATUS- The apparatus is in 6 units. (1) The Geiger - MuZZer tube-Fitted with a window not more than 0.0025 inch thick, to minimise absorption of @-particles. (2) Pre-amplifying unit.June, 19511 POTASH BY MEASUREMENT OF ITS RADIO-ACTIVITY 357 (3) Lead castle-For housing the counter, with 4 cm of lead shielding in every direction. (4) Scaling unit-Capable of registering up to 2000 counts per minute.(5) High voltage power pack-1000 to 2000 volts. (6) SampZe trays. (1) THE GEIGER - MULLER TUBE- The body of the Geiger - Miiller tube is a 2-inch diameter copper tube 24 inches long, with one end closed except for a small hole and the other open and provided with a flange. The flanged end is covered by a sheet of duralumin 2-5 thoiisandths of an inch thick, sealed on with a thermo-setting plastic cement. The small hole carries a glass bead through which a tungsten electrode is sealed. The body contains a mixture of 90 per cent. of argon and 10 per cent. of alcohol at a pressure of 0-1 atmosphere. When the tube is in operation the central tungsten wire is at about 1500 volts positive to the body, which is earthed. P-Particles enter through the thin duralumin “window,” and as each one reaches the gas space the electrons produced by ionisation of the gas are accelerated in the strong electric field and produce a momentary discharge.This discharge is transmitted t o the pre-amplifier in the form of a negative pulse. The construction of a Geiger - Muller tube is not to be lightly undertaken; ours are purchased from the General Electric Company’s Research Laboratory a t Wembley. (2) THE PRE-AMPLIFIER- The pre-amplifier is a single valve, resistance-capacity coupled, placed within a few inches of the Geiger-Muller tube in order to reduce stray capacities. Its voltage gain is in the region of 10 and its output power is adequate for transmission along a hundred feet of screened cable. It draws its power for filament heating and anode supply from the scaling unit. The pre-amplifier is made by Messrs.Dynatron of Maidenhead. (3) THE LEAD CASTLE- The lead castle is simply a lead box to protect the Geiger - Muller tube from stray radiation and cosmic rays. It is made in five main parts and a door, so that when dismantled it is reasonably light to handle. It provides a 4-cm thickness of lead all round, contains a shelf for the Geiger - Miiller tube and has slots for locating the sample tray. The whole unit is made by the E.R.D. Engineering Co., Ltd., Slough. Although it is designed specifically for the 1-inch G.E.C. tube, the modifications necessary for fitting the 2-inch tube are simple. Two brass plates 3 inches square and bored 2+ inches are fitted in place of a shelf and screwed ring.(4) THE SCALING UNIT- The scaling unit is a somewhat complicated electronic device for counting the pulses generated by the Geiger - Muller tube, and passed forward by the pre-amplifier. The pulses arrive at irregular intervals, and are sometimes in the form of a multiple discharge in which a number of very closely spaced pulses are produced by only one P-particle. Direct registering by means of an electromagnet operating a mechanical counter is impracticable because many of the pulses are too close together to be distinguished by such a device. On the other hand it is necessary to prevent the occasional multiple discharges being counted as more than a single particle. The scaling unit used for potassium counting is designed to ignore all but the first pulse of a multiple pulse, but to accept all pulses separated by more than one thousandth of a second.The pulses are then “scaled down” in number by a factor of 10, and only every tenth pulse is passed forward to the electromagnet and ratchet counter registering tens, hundreds, thousands, etc. (the units are shown by the neon lights). The maximum rate handled by the ratchet mechanism is 3 per second, or about 200 per minute, corresponding to a count of 2000 per minute. The design of the scaling unit follows closely the system described by Rotblat and S a ~ l e , ~ except that the scaling down factor is 10 instead of 100, and some of the extra refine- ments have been omitted. The unit is mains operated, and has a conventional unstabilised power pack, which also serves the pre-amplifier.The first two valves “shape” the pulse received into a 30-volt square-topped pulse lasting 1/1000 of a second, and during this pulse358 WILSON, LEES AND BROOMFIELD THE DETERMINATION OF [Vol. 76 the unit ignores any other signals received frctm the pre-amplifier. The square-topped pulse is fed to a “scale of two” unit consisting of a double diode and a double triode, connected so that if one half of the triode is conducting the other is not. When this system receives a pulse the two halves change roles, and at the second pulse they change back again, and so on. When they change back at the second, fourth, sixth, etc., pulse they pass another pulse on to another “scale of two” unit, which in turn passes on pulses at fourth, eighth, sixteenth pulses.This system of dividing by two at each unit could be carried on indefinitely, but in the apparatus used for this work a special electronic switch is arranged to bypass the second and third scale of two units at alternate operations of the fourth (and last) unit. In this way the fourth unit becomes a scale of 10 instead of a scale of 16, and the arithmetic is made simpler for the operator. The mechanical counter thus registers the tens, hundreds, thousands, etc., and the units are indicated by neon lights connected to the scale of two circuits. The first two valves (the “pulse shapers”) can be converted by operating a switch into an oscillator generating pulses every second, or alternatively every tenth of a second. These pulses are used for routine tests of the scaling circuit and mechanical counter.(5) THE HIGH VOLTAGE POWER PACK- The high voltage power pack provides the 1500-volt supply for the Geiger - Miiller tube. Different tubes require different voltages, and the power pack is adjustable between 1000 and 1800 volts; it is stabilised to maintain the correct output during wide variations of the mains voltage. A transformer and half-wave rectifier valve provide about 2500 volts, which is smoothed by condensers and resistances. A micro-ammeter reading to 200pA in series with 10 megohms serves as an output voltmeter, and a pair of valves serve as a shunt load to draw off surplus current if the voltage rises above the pre-set figure, A standard 120-volt H.T. battery provides a reference voltage to control the valves.The scaling unit and the H.T. power unit described above were made up from surplus Government material, but versatile scalers and power packs can be obtained commercially which provide all the facilities needed for p0tassiu.m counting and can also be used for higher or lower counting rates and can provide H.T. voltages over a wider range. (6) SAMPLE TRAYS- The sample trays are brass plates 6 mm thick, with a central circular recess of 5.8 cm in diameter and 4 mm deep. They are flanged so as to slide into any one of four slots machined in the base of the castle, which permits the selection of four distances between the Geiger window and the sample. Usually the top slot is used. PROCEDURE Connect the H.T. power pack and the counting unit to the mains and to the Geiger Connect the tube to the pre-amplifier and the latter to the counting circuit. Adjust (This varies from tube The specified voltage should not be exceeded by If this is.tube. the H.T. voltage to that specified by the makers of the tube in use. to tube, and is usually about 1500 volts. more than 50 volts.) correct, close the castle and proceed to ascertain the “background,” over several hours. Test the scaling unit by the internal impulse for fast and slow counting rates. CALIBRATION- Dry AnalaR potassium chloride, after grincling to pass a 60-mesh sieve, and fill the sample tray. Tamp down the sample lightly with a large clean glas surface, level off carefully with a spatula and brush off any fragments outside the depression.Insert into the top slot of the castle, read the scale, start the stop-clock and switch on the scaling unit. Allow to count for 20 to 30 minutes. Calculate the gross counts per minute. Repeat three times and take the average. This represents counts for 63.1 per cent. of K,O. Prepare mixtures of AnalaR potassium chloride, pure ammonium sulphate and pure ammonium phosphate, by grinding together in a mortar weighed quantities of the materials. The exact amounts are immaterial, but it is convenient to have the K,O content in regular steps, e.g., 50, 40, 25 and 15 per cent. of K,O. Each mixture should be counted four timesJune, 19511 POTASH BY MEASUREMENT OF ITS RADIO-ACTIVITY 359 for long enough to give more than 10,000 counts each time and the average taken.The values, plotted on a graph of percentage of K20 against counts per minute, should lie on a straight line, which if produced will coincide with the background value at “0 per cent. K20.” From our experience this checking is probably unnecessary, and a straight line graph can be drawn from “background” to 63.1 per cent. of K20. To do this the pure potassium chloride point must be ascertained with the greatest of care. The value obtained is a property of the Geiger tube in use, because of the variation in sensitivity to radiation from tube to tube. Tubes do not last for ever, and must be checked from time to time.‘ We have found that permanent standards are very useful. These are glass discs made to fit into the sample trays. One is a uranium glass, which gives a count of about 500 per minute with an average Geiger tube, the other two are potash glasses, with counts of 350 and 200 per minute respec- tively.The use of these standards is a great help in ensuring reproducibility. As uranium radiation is more penetrating than potassium radiation, two Geiger tubes may not have the same ratio of counts for uranium and potassium, and a third potash glass standard of higher K,O content might be preferable. An average figure has already been given. NOTE-The Geiger tubes do not function in a satisfactory manner at a temperature of less than 10°CI and films of moisture on high tension equipment are never beneficial. It is therefore essential to have the equipment in a dry warm room, preferably about 20” to 22’ C. CostrelP gives a general account of the theory of Geiger - Muller counters.DETERMINATION OF K20 IN SAMPLES, e.g., POTASH SALTS- If damp, dry the sample at 105” C, grind to approximately 60 mesh, charge into the sample tray and proceed as described for pure potassium chloride. Allow to count for at least 10,000 counts, preferably for 20,000. Calculate the gross counts per minute and read the K,O content from the graph, or calculate from the factor. It is advisable to check the background for half an hour first thing in the morning and last thing at night, as a check on the cleanliness of the inside of the castle. A test piece should be used as a check about every 48 hours for half an hour or so. NOTE-It has been found that when a Geiger tube is reaching the end of its life, it may diminish in sensitivity when used continuously and recover, or partially recover, if left unused.Thus, a sample could show, say, 250 counts per minute in the morning and if it were left inposition for some hours, the rate could drop to 200 counts per minute or even less. This source of error should be guarded against, especially with old Geiger tubes. DETERMINATION OF K20 IN BILLINGHAM FERTILISERS MADE FROM MOROCCO ROCK- Because Morocco phosphate rock contains traces of elements that are immensely more radio-active than potash, errors are likely to be higher in the determination of K20 in the fertilisers than in potash salts, although but little of the extraneous radio-active elements find their way into the finished product. The magnitude of the effect can be seen from the following figures obtained under comparable conditions.Pure potassium chloride . . . . -500 to 550 counts per minute (net) C.C.F. containing -15% K,O .. -150 99 $9 99 Morocco rock ... . . .. .. 170 to 240 9’ 99 It is difficult to differentiate between the radiation from the two sources, but provided that the product is reasonably uniform, a suitable allowance can be made. So long as the method of manufacture and the raw materials remain unchanged, the contribution made by radio-active substances other than potash will be a function of the P20, content of the fertiliser, as the extraneous radio-activity is associated with the P20,, and hence the P20, content may be used to calculate a correction. The calculation of the K20 content from the Geiger counts may be generalised as- Percentage of K20 = (Net counts/min.- yo P,O, x radio-activity of Morocco rock x a) x b where the radio-activity of the rock is expressed as net counts per minute and a and b are constants; b indicates the net response of the Geiger tube to potash and must be determined for each tube. As long as similar Geiger tubes are in use, changes in a will not be great from tube to tube; and as the term including a is a second order correction, no great error will result if a is taken as constant for all similar Geiger tubes. With recent deliveries of360 WILSON, LEES AND BROOMFIELD: THE DETERMINATION OF [Vol. 76 Morocco rock and the present supply of tubes, a has a value of 0.004. For a typical 2-inch Geiger tube, the formula then becomes- Percentage of K20 = (Net counts/niin.- yo P,O, x 0.7) x 0.105 RESULTS STANDARDISATION WITH PURE POTASSIUM CHLORIDE- In the earlier part of our work, the same tube was standardised with pure potassium chloride fifteen times, the tests being spaced over a month, each count being continued until between 20,000 and 30,000 impulses had been recorded. The range was from 625 to 666 net counts per minute, the average 653, and the standard deviation from the mean 9-51 counts per minute. This gives a coe8cient of variation (i.e., S.D. expressed as a percentage of the mean) of 1-45 per cent. KZO IN FERTILISER SAMPLES- A range of fertiliser samples with a K20 content of 14.5 to 16.0 per cent. was. analysed with the results shown in Table I. DETERMINATION OF K,O Net counts per minute 154 151 153 154 151 153 153 155 155 154 162 150 146 155 151 152 158 149 154 156 155 154 163 153 151 150 147 150 154 155 Average K,O Y O by chemical analysis, 15.0 15.3 3 5-0 15.3 14.9 14.8 15.0 15.3 15.2 15.1 14.7 14.8 14.5 15.2 15-1 15.2 15.6 15-0 15.3 15-1 15.3 15.3 15.7 15.3 14.9 14.9 14.6 15.1 15.3 15-1 .... K2O % by Geiger counter, 15.3 15.0 15-2 15.3 15-0 15.2 15.2 15.4 15.4 15.3 15.1 14.9 14.5 15.4 15.0 15.1 15.7 14.8 15-3 15.5 15.4 15.3 16.2 15-2 15.1 14-9 14.6 14.9 15.3 15.4 .. .. TABLE I BY CHEMICAL ANALYSIS AND GEIGER COUNTER Net counts Difference per minute + 0.3 - 0.3 + 0.2 0.0 +0.1 + 0.4 + 0.2 +O*I + 0-2 + 0.2 + 0.4 +om1 0.0 + 0.2 -0.1 - 0.1 +om1 - 0.2 0.0 + 0.4 + 0.1 0.0 + 0.5 -0.1 + 0.2 0.0 0.0 - 0.2 0.0 + 0.3 155 155 156 152 154 152 153 161 161 161 157 155 156 156 159 158 156 155 158 152 154 157 150 153 158 156 151 156 155 157 Standard Deviation of Geiger count from chemical analysis “Coefficient of variation” 1-84 K2O by chemical analysis, % 15.2 15.3 15.3 14.8 15.1 14.9 15.3 15.3 15.6 15.0 15.0 15.4 15.2 15.7 15.9 15.6 15.7 15.4 15.3 15.3 15.7 15.6 15.1 14.6 15.4 15.0 14.6 15.2 15.7 15.4 15.19 KzO Y O by Geiger counter, 1-54 15.4 15.5 15.1 15.3 15-1 15.2 15.8 15.8 15.8 15.4 15.2 15.3 15.3 15-6 15.5 15-3 15.2 15.5 14-9 15.1 15.4 14.7 15.0 15.5 15.3 14.8 15.3 15.2 15-4 15-25 0.28 Difference + 0-2 +0m1 + 0.2 + 0-3 + 0-2 + 0.2 -0.1 + 0.5 + 0*2* + 0.8 + 0.4 - 0.2 +0.1 - 0.4 - 0.3 -0.1 - 0.4 - 0-2 + 0.2 - 0.4 - 0.6 - 0.2 - 0.4 + 0.4 +0.1 + 0-3 + 0.2 +0.1 - 0.5 0.0 0.22 * A new Geiger tube was put in at this point in the series.June, 19511 POTASH BY MEASUREMENT OF ITS RADIO-ACTIVITY 361 NOTES ON TABLE I- (1) Each sample was analysed less than 24 hours after mixing.(2) The K,O was chemically determined by Perrin’s modification of the chloroplatinate method.6 The standard deviation of this method is not known (but is believed to be small), and hence the standard deviation of the Geiger results cannot be separated between the chemical and physical tests. (3) All the results in Table I have been obtained by counts of about 20,000. (4) Evidence is accumulating that the radio-activity of fertiliser samples increases with time. A batch of 19 samples were tested by Geiger count twelve months after manufacture. All the results were high, the average bias being +l-8.By counting the same sample at intervals results were as follows- Date . . .. .. . . 4.8.49 8.8.49 17.8.49 7.9.49 17.11.49 Net counts/min. . . . . 164 167 169 176 190 Hence the method can only be applied to fresh samples. APPLICATION TO OTHER SAMPLES Once the absence of other radio-active elements is established, rapid approximate determinations of potash can be made. Even potash contents as low as 1 or 2 per cent. have been estimated and as short a time as 10 minutes counting indicates the order of the amount present; the errors are, of course, rather large; they can be roughly calculated from the deviations given on page 356, but in practice these deviations are likely to be exceeded. Some results for muriate of potash are shown in Table 11. TABLE I1 MURIATE OF POTASH K,O by chemical Date analysis, counter, 1.12.48 40.4 40.0 17.12.48 54.0 53.7 8.3.49 41.6 41.8 7.4.49 58.5 60.0 20.5.49 50.3 50.5 7.6.49 51.4 52.8 8.6.49 50.8 51.2 30.6.49 40- 1 39.3 6.7.49 49.6 48.9 12.7.49 41.0 40.1 22.8.49 59.1 58.7 9.9.49 60.0 60.7 All the analyses were made on dried samples. (Minimum number of counts = 20,000.) K,O by Geiger % % In the above series of determinations the standard deviation 0, the counter method from the chemical method is 0.78, and the coefficient of variation is 1-57. This is rather higher than the standard deviation given for fertiliser samples in Table I, but the “coefficient of variation,” i.e., the standard deviation expressed as a percentage of the mean, is lower. As the muriate of potash is free from other radio-active elements, this might be expected. REFERENCES 1. 2. 3. 4. 5. 6. PHYSICAL CHEMISTRY GROUP RESEARCH DEPARTMENT Barnes, R. B., and Salley, D. J., Ind. Eng. Chem., Anal. Ed., 1943, 15, 4. Gaudin, A. M., and Parnell, J. H., Anal. Chem., 1948, 20, 1154. Kamen, M. D., “Radio-active Tracers in Biology,” Academic Press Inc., New York, 1947, p. 86. Rotblat, J., and Sayle, E. A., J . Sci. Instr., 1948, 25, 33. Costrell, L., J . Res. N a t . Bur. Stand., 1949, 42, 241. Perrin, C. H., And. Chem., 1949, 21, 985. I.C.T., BILLINGHAM DIVISION

ISSN:0003-2654    

DOI:10.1039/AN9517600355

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

362 WINTERINGHAM : RADIOMETRIC ASSAY I N TRACER EXPERIMENTS [Vol. 76 Radiometric Assay in Tracer Experiments BY F. P. W. WINTERINGHAM (Presented at the meeting of the Physical Methods Group on Tuesday, M a y 23rd, 1950) The assay of radio-active isotopes at tracer levels of activity is briefly discussed in the light of sensitivity, precision and convenience. Experi- mental procedures adopted for tritium (3H), carbon-14 and for bromine-82 are described. IN simple radio-active tracer technique a suitably labelled compound is introduced into the system under investigation. After a time, during which chemical or purely physical changes take place, samples are taken and assayed radiometrically. The radio-activities determined then provide data on the final concentrations and locations of the molecules originally labelled.Many factors have, of course, to be taken into consideration; for example, whether or not there has occurred isotope exchange with an unlabelled compound, when the activity as measured may no longer be significantly associated with the original molecule at all. Ideally one should be able to assay a radio-active isotope simply by bringing a suitable detector to a fixed point near the sample, whether it be solid, liquid or gaseous. Few isotopes lend themselves to reliable assay in this manner, particularly at the levels of activity usually associated with the samples of a tracer experiment, e.g., 0.001 pc (microcuries). The method adopted will depend upon the nature and energy of the radiations associated with the decay of the isotope, and upon the sensitivity, precision or degree of convenience required.1 Up to the present time the Geiger-Muller counter has generally been found to be the most suitable detector in tracer work (cf.Libby3 and others). This paper briefly describes some typical assay procedures developed in the course of radio-active tracer experiments. In these experiments carbon-14 has been used to label methyl bromide in a study of its chemical action on the protein of wheat under the conditions of fumigation against insect pests. By treating the resulting methylated protein constituents with such reagents as sodium hydroxide and hydriodic acid, followed by the determination of the 14C-labelled methyl groups in the volatile products, it has been possible to identify the principal protein groups methylated.Provision has been made for the assay of tritium so that it could be used as an auxiliary tracer for methyl groups. Bromine-82 has been used to label a bromine analogue of DDT that was then injected into normal and “DDT-resistant” houseflies. The enormous sensitivity of the radio-active tracer technique combined with paper partition chromatography enabled unchanged DDT and its metabolites to be identified and determined in single flies. TRITIUM, 3H This isotope decays with a half-life of about 11 years and emits a negative 13-particle of maximum energy 0.017 MeV.2 The maximum range, r, of the P-particle calculated from the empirical equation3 Y = E6’3/150, is 0.75 mg per sq. cm, E being measured in KeV. This means that if the tritium were assayed in a 1-mg sample spread over 1 sq.cm below a thin window counter, more than 90 per cent. of the emitted P-particles would be absorbed in the sample alone. Tritium has, therefore, usually been assayed as a gas or vapour introduced into a quartz ionisation chamber attached to a Lauritzen electroscope4 or into the sensitive volume of a Geiger - Muller counter itself.6 In Melander’s method the tritium is assayed as tritium-labelled hydrogen inside the counter at a pressure of 9 cm of mercury together with methane at a partial pressure of 1 cm of mercury. Melander’s apparatus was relatively simple, all gas manipulations taking place from higher to lower pressures and an ordinary rotary oil-immersed vacuum pump being used for evacuation.This method has been found to give satisfactory results, but provision has been made for transfer of gas from lower to higher pressures by means of a Topler pump and for the recovery of active samples. The laboratory-made gas counter shown in Fig. 1 consists of a brass tube, D, of internal diameter 2.5 cm and length 15 cm, the ends C being turned to take standard B 29 Pyrex cones. The counter is attached at A to the high-vacuum assembly by means of a standard B 10 joint. The anode, E, is 42 S.W.G. tungsten wire attached at one end to the glass bridge, B, and at the other end to a platinum lead, H, sealed in at G after the tungsten wireJune, 19511 WINTERINGHAM : RADIOMETRIC ASSAY IN TRACER EXPERIMENTS 363 has been drawn taut. The relevant part of the high vacuum assembly is shown in Fig. 2.The principal manifold is exhausted at B, through the usual cold traps, by an oil diffusion pump backed by a rotary oil pump. The active hydrogen sample, prepared by reducing the corresponding water sample: enters the system at A and is stored in one of the reservoirs R2, R3, etc. Residual gas is transferred from the manifold to the reservoir by means of the Topler pump, TP, residual pressures being indicated by a Pirani gauge at P. The Topler pump is operated by opening the reservoir to the air via a drying tower at L or to an auxiliary vacuum pump at W. Pressures of gas samples are indicated by the mercury manometers M1, M2, etc. The reservoir R1 is charged with inactive hydrogen, which is used for dilution Fig. 1. Internal gas counter for tritium assay and washing out.The suitably diluted sample is admitted to the counter, GMC, through tap T5, to a pressure of 9 cm of mercury. Methane is added from a cylinder via the “doser” bulb, D, to a pressure of 1 cm of mercury. Not less than 10 minutes are allowed for the sample to mix properly before counting is begun. The counter is connected at X and 2 to conventional extra-high tension, pre-amplifying, quenching and scaling circuits. After counting, the sample is pumped by the Topler into the oxidation unit, OU, which can be evacuated at G by the auxiliary vacuum pump. Oxygen is admitted from a cylinder through tap T18 to about twice the pressure of the hydrogen - methane sample. The total pressure within OU before ignition is never allowed to exceed 30cm of mercury.The mixture is flashed quietly by bringing the 39 S.W.G. platinum filament to red heat for a few seconds. The water produced may be trapped in TR and sealed off after pumping away unwanted carbon dioxide; alternatively, it may simply be absorbed in the soda lime tube, S. Active gas samples are never passed into the laboratory atmosphere. Between assays the counter and manifold are flushed out with inactive hydrogen and methane; after a series of assays they are pumped out at high vacuum. As used, the counter showed a background of 60 counts per minute, but this figure could no doubt be reduced by surrounding the counter with similar counters connected in anti-coincidence.6 The threshold was + 1500 V, the plateau extending for a little over 30 V.It was therefore found advisable to check the plateau limits with an external radium source just before each count and to operate the counter at the mid-point of these limits. The absolute activity of the samples assayed was not known and it was not possible to estimate the total counting efficiency. Theoretically, every dis- integration within the sensitive volume of the counter that does not occur during the quench period should give rise to a recorded pulse, so that a high counting efficiency would be expected. The geometry of the sample is dictated by the shape of the counter and is therefore constant, an important factor in precision or reproducibility. Melander quotes a figure of +2 per364 WINTERINGHAM : RADIOMETRIC ASSAY I N TRACER EXPERIMENTS p o l .76 cent. in his original paper.6 The manipulation tends to be tedious; it was not possible to deal with more than one prepared active sample per hour. CARBON-14 Carbon-14 decays with a half-life of 6400 years' and emits a negative /.?-particle of maximum energy 0-154 MeV.* The maximum range is about 28.5 mg per sq. cm, so that losses by self-absorption and absorption in the window-foil of the counter may seriously reduce the total counting efficiency for externally-mounted solid samples. Like tritium, carbon-14 has been assayed as carbon dioxide inside an ionisation chamber9 and within the sensitive volume of a Geiger - Muller counter in admixture with carbon disulphide to improve the gas counting characteristics.1° Assay by the latter method has been estimated to be more than 95 per cent.efficient.ll Solid samples can be assayed inside demountable counters through which flows a suitable gas mixture.12~13 In this way window-foil absorption losses are avoided and certain geometrical effects are reduced. A helium - ether flow mixture was found to give better counting characteristics than a helium - alcohol mixture,ls but the higher threshold voltage (over 3000 V) is a disadvantage, because such high stabilised voltages are not always available. The use of a helium - alcohol mixture flowing through a demount- able counter electronically quenched by an external circuit has been found to be very satisfactory. In a routine work this has proved a useful compromise between the efficient gas-counting and the convenient external-sample techniques.Carbon-14 has been assayed directly in mounted tissue specimens,l4 but in order to achieve a high degree of reproducibility it is better to prepare samples in some standard chemical form, e.g., as barium or calcium carbonate, although this necessitates preliminary combustion or wet oxidation. Calcium carbonate has the advantage of a higher carbon content than the barium salt, self-absorption being correspondingly less for a given activity. The following procedure has been developed for the assay of carbon-I4 at tracer levels of activity.June, 19511 WINTERINGHAM : RADIOMETRIC ASSAY IN TRACER EXPERIMENTS 365 Carbon dioxide resulting from a wet or dry oxidation is purified and adsorbed, in con- ditions that preclude the presence of atmospheric carbon dioxide, in a train of three specially- designed units of the type shown in Fig.3. Each unit is first flushed out with entraining 15 U T 2 0 c m i I G H Fig. 3. Absorption of C*O, and precipitation as CaC*O, gas, e.g., nitrogen or a helium - oxygen mixture, that is free from carbon dioxide, The entraining gas enters at P, passes through bulbs E and D and leaves at C, which is a small indicator bubbler charged with aqueous barium hydroxide. When flushed out, bulb D is disconnected at 0 and the opening is immediately stoppered. The joint N is attached to a burette, J, containing the required reagent, free from carbon dioxide. The tip of the burette is normally protected from carbon dioxide by the soda-lime tube L. The required volume of reagent is added, tap T2 being opened to release the displaced atmosphere in D.The unit is quickly re-assembled and reflushed. The reagents are transferred from bulb D to bulb E, which contains a sintered glass bubbling-plate, while the entraining-gas passes through the system, entering each unit at P and leaving at Q. The transfer of the reagent is assisted by applying a slight positive pressure in D through the soda-lime tube B by means of the bellows, A. Reagents and washings are finally sucked into flask F, which may be exhausted at S with taps T3 and T6 closed. To wash the apparatus, D is filled to the neck, N, with distilled water free from carbon dioxide, or with any specially required washing agent. By exhausting at S, opening taps T4 and T2 and closing tap T3, the top half of the washing agent swills round the head above 0 and drains into F.By closing tap T2 and opening tap T3 the lower half of the washing agent sprays the barrel of E from the pin-holes of the small bulb at the bottom of the capillary extension vertically below T3. During the operation of washing, tap T1 is opened to B and the bellows, A, is detached. Silicone grease is used for lubricating joints and taps.366 WINTERINGHAM : RADIOMETRIC ASSAY I N TRACER EXPERIMENTS [Vol. 76 For the purification and absorption of the carbon dioxide resulting from a wet or dry oxidation, the first unit is charged with N alkaline sodium permanganate and the second unit with N alkali containing 0.3 N hydrazine. The third unit is charged with N sodium hydroxide alone.When the absorption of carbon dioxide is complete, an excess of sulphuric acid is added to the alkaline reagent in the first unit and the second unit is acidified with hydrochloric acid, the flow of entraining gas being maintained. When oxygen is used in the extraining gas, e.g., during the dry combustion of methyl iodide labelled with 14C and 1311 recovered from the labelled methylation of protein by a modified Baernstein method, the hydrazine is not added in the second unit until a few minutes before acidification. Sulphur dioxide and halogens are absorbed in the first bubbler, but some free halogens may reach the second bubbler where they will be reduced. On acidification all the labelled carbon dioxide passes over to the third bubbler where it is absorbed, the alkaline washings being IA 10 cm J Fig.4. Demountable flow- type counter for soft P-assay finally sucked into F. Excess hydroxide in the final washings of the third unit is largely removed by boiling with 5 N ammonium chloride, any excess of this reagent being carefully avoided. The carbonate is precipitated from the hot solution by the dropwise addition of slightly alkaline calcium chloride. The precipitate is filtered off at the pump in the filter-stick G. The soda-lime tube, F, excludes atmospheric carbon dioxide during the initial stage of the filtration when the suspension is still alkaline. The precipitate is collected on a tared 1.75-cm Whatman No. 50 filter-paper backed by two Whatman No. 4 filter-papers. These are pressed flat by the rim of G, which is clamped firmly and vertically in the small Biichner funnel, H.“Creeping” of the calcium carbonate is prevented by the addition of absolute alcohol during the filtration process. The precipitate is finally washed with ether and then G is released from the clamp and gently raised from H. The precipitate remains as a clean circular disc; by selection of the.right internal diameter for the tube G the disc has an area of 1.50 sq. cm. A drop of a neutral solution of cellulose nitrate in redistilled amyl acetate (prepared from fresh “Durofix”) is added to the disc, which is re-dried while the edge of the filter circle is held flat in a small brass cup. The disc, which can now be handled without fear of damage, is re-weighed and mounted in the brass cup, F, of the demountable counter shown in Fig.4. The counter shown in Fig. 4 was designed for simplicity of construction and manipula- tion; it has functioned reliably and reproducibly for several months. The brass base, J, carries a standard B 29 solid brass cone. The counter envelope is a B 29 Pyrex socket closed at one end, as shown. The short glass capillary, A, carries a short platinum wire lead connected within the capillary to the 42 S.W.G. tungsten anode, D. The anode ter- minates in a 2-mm Pyrex bead, E, which clears the surface of the brass cup by 1 mm. The cathode surface, C, is prepared by painting the inside of the glass envelope with “Aquadag 660 B” followed by baking at 200” C. The cathode surface extends to the lower rim of the socket, so making electrical contact with the brass base, which is earthed at I.During counting, a helium - alcohol vapour mixture enters at B and leaves the counter through G and H. The operational assembly is shown in Fig. 5. Commercial helium from a cylinder enters at tap T2 and passes through the bubbler, A, containing absolute alcohol cooled in The disc is dried under an infra-red lamp and weighed.June, 19511 WINTERINGHAM: RADIOMETRIC ASSAY IN TRACER EXPERIMENTS 367 iced-water. The gas, saturated with alcohol vapour in equilibrium with the liquid at 0” C , passes through a packed wool-dust filter and through the spiral B in which it is brought to equilibrium with room temperature. The rate of flow of the emergent gas is indicated on the flowmeter E. The counter is flushed out for 5 minutes a t 5ml per second before use, but the gas flow is reduced to one or two bubbles per second in F during counting.The counter operates in conjunction with conventional extra-high tension, quenching and scaling circuits. The threshold of a typical counter of this design was +1500V and the “flat” plateau extended over a range of at least 200V. When screened by 1i inches of lead, the mean background rate was 30 counts per minute. By starting with a 14C-labelled sodium carbonate solution it was possible for one person to make about 12 assays in one day. It was of interest to compare the efficiency of assay by this method with the carbon dioxide gas-counter technique. A sample of calcium carbonate whose absolute activity was known from a comparison with a calibrated 14C standard was assayed by the demountable counter technique described, and as an external sample mounted below a commercial counter having the thinnest end-window, 1 4 m g per sq.cm, that was available. The sample was then converted to carbon dioxide and assayed in a typical gas-counter assembly. With a precipitate 1 mg per sq. cm thick, the demountable counter was 29 per cent. efficient, the R C Fig. 5. Assembly for soft P-assay using demountable flow-type counter thin end-window counter was 10 per cent. efficient and the efficiency of the gas-counter was 65 per cent. The efficiency of the gas-counting could, however, have been increased by using a counter with a higher sensitive-volume to total-volume ratio than was available a t the time. In a series of assays made on a purely routine basis with the demountable counter, the coefficient of variation for all results was less than 2 5 per cent. The random nature of the decay accounted for 1 per cent.of this variation. BROMINE-82 This isotope decays with a half-life of 34 hours2 and emits a negative /%particle of maximum energy 0.465 MeV and y-rays of energies 0.545, 0.787 and 1.350 MeV in cascade.2 The maximum range of the P-particles is about 150 mg per sq. cm, while the y-rays can be detected through several centimetres of lead. Samples with an activity of the order of millicuries can be compared accurately at a distance (eg., 50 cm) from a shielded ?-ray counter or ionisation chamber, self-absorption corrections being unnecessary. The results of some measurements made with a y-ray counter agreed to within 1 to 2 per cent.with those obtained by counting the much diluted samples in a liquid counter of the type described368 WINTERINGHAM : RADIOMETRIC ASSAY I N TRACER EXPERIMENTS [Vol. 76 by Veall.ls This type of counter proved extremely useful in tracer experiments with this isotope because solutions or suspensionsle can be assayed directly. For high-efficiency counting, the bromine of the labelled substance is separated chemically and precipitated as the silver bromide, carrier bromide being added when necessary, and the precipitate is mounted on an alkathene-faced lead disc below a thin end-window c0unter.l' The coefficient of variation for all results by this method is less than 1 2 per cent. on total counts of 10,OOO.With a counter of window thickness 1.8 mg per sq. cm, it was estimated that 15 per cent. of the s2Br disintegrations of a silver bromide precipitate weighing 15 mg per sq. cm were being detected. In these conditions the counting rate was 50 times greater than that observed with a sample having the same activity but assayed in solution in the liquid counter. SELF-ABSORPTION Of the many factors that affect the total efficiency of B-assav of solid samples. self- absorption is probably the most important. As it is"not conve&ent to stand&-dise the 0.9 0 *8 0.7 0.6 0-5 gO, 0.3 0.2 04 0 04 0.2 0 3 0.4 0.5 0.6 0.7 0.8 0.9 Sample thickness/Maxirnum Beta range Fig. 6. Theoretical and experimental /I-self-absorption curves thickness of the precipitate for all samples, it is essential to construct a self-absorption curve to provide a ready basis for comparing assays made on solid mounts of different thicknesses.Experience at this laboratory has suggested, fo'r general work at least, that corrections based on Libby's equation3 are quite satisfactory. On the basis of the apparent exponential absorption of /3-particles Libby has deduced the expression- I d (1 - e-") kd ' I0 _- -June, 19511 WINTERINGHAM : RADIOMETRIC ASSAY IN TRACER EXPERIMENTS 369 where I d is the observed count, corrected for “dead” time losses, etc., and I , is the theoretical count at “infinite thinness,’’ i.e., were there no self-absorption. The value d is the thickness of the sample in terms of the maximum range of the p-particles; this datum is usually more readily available than the absorption coefficient that appeared in earlier equations.The constant k was originally evaluated as 5, but more recent data show that 8 is a better value. The continuous line in Fig. 6 is the curve of Libby’s equation. The circles represent the experimental points of a series of 1% assays made by the demountable counter technique as described. The triangles represent the experimental points of a series of 82Br assays made with the alkathene-mounted silver bromide discs to which reference has been made. All points lie reasonably near the theoretical curve, and it is of interest that in the 82Br assays the y-ray component was of no significance in this method of counting. The author is indebted to Mr. J. V. P. Long of the Chemical Research Laboratory for making comparative l4C assays by the gas-counting technique. This paper is published by permission of the Department of Scientific and Industrial Research. REFERENCES 1. Maddock, A. G., Analyst, 1948, 73, 644. 2. Seaborg, G. T., and Perlman, I., Rev. Mod. Phys., 1948, 20, 585. 3. Libby, W. F., Anal. Chem., 1947, 19, 2 . 4. Henriques, F. C., and Margnetti, C., Ind. Eng. Chem., Anal. Ed., 1946, 18, 420. 5. Melander, L., Acta Chem. Scandin., 1948, 2, 440. 6. Putman, J. L., J . Sci. Instr., 1949, 26, 198. 7. Hawkings, R. C., Hunter, R. F., Mann, W. B., and Stevens, W. H., Canad. J . Res., 1949,27, 645. 8. Solomon, A., Gould, R., and Anfinsen, C . , Phys. Rev., 1947, 72, 1097. 9. Henriques, F. C., and Margnetti, C., Ind. Eng. Chem., Anal. Ed., 1946, 18, 417. 10. Miller, W. W., Science, 1947, 105, 123. 11. Hawkings, R. C., Hunter, R. F., and Mann, W. B., Canad. J . Res., 1949,27, 555. 12. Kelsey, F. E., Science, 1949, 109, 666. 13. Tait, J. F., and Haggis, G. H., J . Sci. Instr., 1949, 26, 269. 14. Hogness, J. R., Roth, L. J., Leifer, E., and Langharn, W. H., J . Amer. Chem. Soc., 1948,70, 3840. 15. Veall, N., Brit. J . IZadiol., 1948, 21, 347. 16. Howarth, F., Nature, 1949, 163, 249. 17. Winteringham, F. P. W., Ibid., 1949, 164, 183. DEPARTMENT OF SCIENTIFIC AND INDUSTRIAL RESEARCH SLOUGH, BUCKS. PEST INFESTATION LABORATORY LONDON ROAD

ISSN:0003-2654    

DOI:10.1039/AN9517600362

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

370 CROPPER: A STUDY OF 'THE POTENTIOMETRIC [Vol. 76 A Study of the Potentiometric Titration of Fluoride with Lead BY F. R. CROPPER (Presented at the meeting of the Microchemistry Group on Wednesday, September 27th, 1950) A preliminary study has been made of the conditions suitable for the potentiometric titration of fluoride by means of lead nitrate in presence of ferric and ferrous chloride indicator and of excess of sodium chloride. The fluoride is precipitated as lead chlorofluoride, and as the end-point is approached the ferrifluoride complex ion dissociates and gives a change in the Fe"' - Fe" redox potential; the equivalence-point is indicated by maximum change in the potential. Reproducible results can be obtained under specified conditions, and provided that the lead nitrate be standardised against sodium fluoride solution, the method would be accurate to within 1 per cent.With N lead nitrate the end-points are reasonably sharp for 40 to 120mg of fluorine, but on the micro scale they are such that careful working is essential. The preliminary treatment of organic fluorine compounds, prior to potentiometric titration with lead, is being investigated and will be the subject of a subsequent publication. A NUMBER of papers have been published during the last two decades on the quantitative precipitation of fluoride as lead chlorofluoride. Since the early work of Hawleyl on the effects of time of precipitation and of varying amounts of fluoride on the composition of the precipitate, studies have been made with the object of finding the optimum conditions for this method; Hoffmann and Lundel12 investigated the effects of pH and of chloride concentra- tion for the range 10 to 100mg of fluorine.Elving and Ligett* described the potassium fusion method for decomposing organic fluorine compounds, and the conditions for subsequent precipitation of lead chlorofluoride; they claimed good results for a variety of fluorine com- pounds. Details have been published for the lead chlorofluoride method as adopted by the Association of Official Agricultural Chemists4 for the analysis of insecticides. Although the gravimetric method has received some attention, the usual procedure has involved washing the precipitate with a little saturated lead chlorofluoride solution , re-dissolving the precipitate and determining the chloride content by the Volhard method.Recent evidence of the difficulties still met with in this determination is given by Chapman, Heap and Sa~nders,~ who studied the effects of variations in pH, chloride and fluoride ion concentrations and times of stages in the precipitation ; these authors recommended applica- tion of certain corrections and claimed that for 50 to 60 mg of fluorine the corrected results were accurate to about 1 per cent. of the true value. Since that date, Kaufmans has also studied the effect of pH on the precipitation. Recent work in these laboratories on a procedure similar to that of Chapman, Heap and Saunders, has shown that it is difficult to get reproducible correction curves. A method that would not necessitate filtration and washing, nor be in error even if chloride ions are adsorbed on the precipitate, would probably be more reliable than the existing method.This paper describes a preliminary study of this possibility. Farkas and Uri7 describe the determination of lead by potentiometric titration with alkali fluoride in presence of alkali chloride. Excess of sodium chloride is added to a solution containing lead ions so that the lead is partially precipitated as lead chloride; when titrated with standard alkali fluoride, the lead in solution is precipitated as lead chlorofluoride, the precipitated lead chloride then dissolves and is re-precipitated as chlorofluoride , until eventually all the lead has been converted to chlorofluoride; an accuracy of +0*5 per cent.is claimed. The test solution containing fluoride is treated with excess of sodium chloride and titrated with standard lead nitrate solution. During the titration small amounts of ferric chloride and ferrous chloride indicator solutions are added, so that most of the ferric ions become (FeFs)"'; as the end- point is approached, this complex ion dissociates and gives an increase in the Fe"'/Fe" In the present investigation, this titration has been reversed.June, 19511 TITRATION OF FLUORIDE WITH LEAD 371 ratio, and hence in the potential of a platinum electrode. The end-point is reached when maximum potential change occurs. The advantage of titration in this direction, compared with that of Farkas and Uri, is that lead chloride does not have to redissolve after each addition of titrant, and equilibrium should be attained more quickly.EXPERIMENTAL APPARATUS- TinsZey potentiometer-Graduated in millivolts, 0 to 1.9 volts. Platinum electrode-About 1 cm square. Saturated calomel electrode-Fitted with a saturated potassium chloride bridge. Tall- form beaker-Fitted with an efficient glass stirrer i o give rapid rotation. uCapacity 100 ml. Burette-Graduated in fiftieths of a ml. The platinum electrode was cleaned frequently by immersion in hot concentrated nitric acid and thoroughly washed with distilled water. Capacity 0 to 5 ml. STANDARD SOLUTIONS- Sodium fluoride-Hopkins and Williams, pure, dried at 130" C before use. Tests did not reveal any chloride and since the sulphated ash figure was correct, it was taken as being pure.N and 0.10 N solutions were used. Lead nitrate-A.R., standardised gravimetrically as lead sulphate ; solutions of approxi- mately N and 0.10 N concentration were used. Ferrous chloride-B.D.H. laboratory reagent. Two grams were dissolved in 10 ml of water and 3 ml of concentrated hydrochloric acid; about 0.5 g of granulated zinc was added and the mixture was allowed to stand for several hours, until a test portion showed no reaction for ferric iron on adding ammonium thiocyanate. It was then made up to 100 ml and about 0.5 ml of mercury added to form an amalgam with the residual zinc. This solution is satis- factory for use for about a week. Ferric chloride-A solution of 0.04 g in 100 ml of distilled water. Sodium chloride-A solution of 25 g in 100 ml of distilled water.Bu$er solution, pH 3.5-Prepared with 5 ml of N sodium acetate and 95 ml of N acetic acid; the pH measured was 3.6. METHOD The tests were carried out on suitable volumes of N or 0.10 N sodium fluoride, or on individual weighed portions of dried sodium fluoride, in a 100-ml tall-form beaker, fitted with a rapidly rotating glass stirrer so that a vortex was formed corresponding to about one-third of the height of the liquid at the walls, a saturated potassium chloride bridge, and a platinum electrode; the bridge led to a saturated calomel electrode cell, and the two electrodes were connected in the usual manner to the Tinsley potentiometer. To the fluoride solution were added 4 ml of 25 per cent. sodium chloride solution and, when required, the 20 ml of buffer solution, 20 ml of ethanol and water to make the total volume about 40ml.About 90 per cent. of the expected volume of lead nitrate solution was added with vigorous stirring and then the ferrous and ferric chloride solutions. The titration was continued, usually in increments of 0.1 ml at a time, until the end-point had obviously been passed. The exact end-point titre was calculated by means of the second diff erential.8 TITRATIONS WITH N LEAD NITRATE (40 TO 1 2 0 ~ ~ OF FLUORINE)- Late addition of indicator-In the work of Farkas and Uri, lead solutions were treated with excess of chloride, ferric - ferrous indicator was added, the pH was adjusted to between 4-0 and 4.2 and the mixture titrated with standard sodium fluoride solution, preferably in the presence of ethanol.In the present investigation, the first experiment was simply a reversal of this procedure, i.e., 4 ml of a solution containing 76 mg of fluorine, i.e., N NaF, was treated with 1 g of sodium chloride, 1-0 ml of ferric chloride solution, i.e., 0.4 mg of FeCl,, and 1 ml of ferrous chloride solution, i.e., 20 mg of FeCl,, 20 ml of ethanol and 20 ml of N buffer of pH 4; the mixture was then titrated with lead nitrate. The result was only 95 per cent. of the theoretical titre. Deliberate addition of the indicator solutions at later372 CROPPER: A STUDY OF THE POTENTIOMETRIC [Vol. 76 stages in the course of the titration showed that higher results were obtained if the indicator was added at about 90 per cent. of the theoretical titre. TABLE I EFFECT OF LATE ADDITION OF INDICATOR SOLUTIONS Theoretical titre,* 4.06 ml Indicator solutions added at, Observed titre Percentage of ml ml theoretical titre 0 3.85 94.9 3.0 3-86 94.9 3.2 3.93 96.8 3.6 4-05 99.7 3.6 3.95 97.3 3.6 4.0 98.5 3.6 3.85 97.5 * Throughout the paper this figure is calculated for 0.983 N or 0.10 N lead nitrate solution.Subsequent work in the absence of alcohol or buffer and in the absence of both alcohol and buffer, confirmed that under these conditions also results rather closer to the theoretical values were obtained when the indicator was added at about 90 per cent. of the theoretical titre; the amount of ferric chloride used (0.4mg) corresponds to but 0.26mg of fluorine (Lea, to less than 0.5 per cent.) in the formation of FeF6"', and it is difficult to see why the late addition should have the observed effect. Variation in the amount of ferric chzloride-The influence of late addition of the indicator solutions suggested that the amount of ferric ion, and hence the ferric - ferrous ratio, might also have an appreciable effect on the level of the results.In the following tests 75 to 80 mg of fluorine (as NaF) were titrated with N lead nitrate in presence of 1 g of sodium chloride and 20 ml of ethanol; the indicator solutions comprising 1 ml of ferrous chloride (20 mg FeCl,) and x ml of ferric chloride (0.0'7 to 5.0 mg of FeC1,) were added at a titre of 3.2 ml. The pH was brought to about 4.0 by adjusting to neutrality against methyl orange after the addition of indicator. The results in Table I1 show that the amount of ferric ion is not an important factor provided not more than 1 mg of FeCI, is used; a fixed amount (0.4 mg FeCI,) was therefore used in subsequent work.TABLE I1 EFFECT OF VARIATION IN THE AMOUNT OF FERRIC CHLORIDE FeCI, added, Observed titre, Theoretical titre, Percentage of mg ml ml theoretical titre 0.07 3-88 4.09 94.8 0.1 3-82 4-00 95.5 0.1 3.85 3-97 97.0 0.2 3.90 4.06 96.1 0.4 3.93 4-14 94.9 0.4 3.88 4-12 94.2 0.6 3-89 4.09 95.1 1.0 3.90 4.11 95.1 5.0 3.74 4.11 91.2 Variation of fiH vaZzle-The effect of pH on the precipitation of lead chlorofluoride has been studied extensively2J'96; for amounts exceeding 20 mg of fluorine, a pH between 4-3 and 4.8 is recommended by Kaufman.6 The stability of the ferrifluoride complex is known to be poor below pH 3, and even above this level the redox potential is influenced considerably by pH variations.Titration of fluoride by lead nitrate in an unbuffered system is accompanied by a significant decrease in pH value, but this effect can be largely counterbalanced by adjusting to neutrality against methyl orange after addition of indicator at about 90 per cent. of the theoretical titre; although this procedure was used in the preceding section a more effective control of pH was desirable. The results given in Table I11 were carried out with 20 ml of N buffer solutions of different pH values added at the commencement of the titration ; 1 g of sodium chloride and 20 ml of ethanol were present, and the indicator solutions were added at 3.6 ml.June, 19511 TITRATIOS OF FLUORIDE WITH LEAD TABLE I11 EFFECT OF VARIATION IN pH pH of buffer solution Observed titre, Theoretical ti tre, Percentage of ml ml theoretical titre 5.0 No end-point 4.06 - 4.00 3.86 3.95 r 4.0 1 3.93 3.87 3.6 { 3.80 3.86 4.06 3.96 4.06 4.06 4-03 4.03 4.06 373 98.5 97.5 97.3 96.8 96.0 94.3 94.8 Curves A, B and C of Fig.1 illustrate the type of titration obtained; the pH of the added buffer must be well below 5; with pH 4 and 3.5 buffers the end-points are satisfactory as regards slope, the former pH giving rather a greater slope but with a smaller over-all e.m.f. change. The pH of the titrated system changes by about 1.5 units from the beginning to the end of the titrations, even in presence of 20ml of N buffer solution; this no doubt in itself causes some decomposition of the ferrifluoride complex.The eflect of ethanoZ-In all the above tests the e.m.f. tended to drift after each addition of titrant so that sometimes 2 to 3 minutes had to elapse before a steady reading could be obtained. Farkas and Uri found that the equilibrium potential was attained more slowly and greater potential changes were obtained when ethanol was present in the titrated system. Although the delay in reaching equilibrium appeared to be caused solely by the slow rate at which lead chloride dissolved, a study of the titration in absence of ethanol appeared to be desirable. The tests shown in Table I V were carried out on 40 to 120mg of fluorine in a total volume of 40 ml of water (including 1 g of sodium chloride and 20 ml of N buffer pH 36), with the ferric - ferrous indicator solutions added at 0.3 to 0.6 ml before the expected end- point. TABLE IV TITRATIONS CARRIED OUT IN THE ABSENCE OF ETHANOL Fe"' - Fe" indicator added at, ml 3-5 3.5 3.5 3.5 3-6 3.6 1.7 5.4 Observed titre, ml 4-14 4.07 4.06 3.99 4.06 4.07 2.04 6.08 Theoretical titre, ml 4-15 4.13 4.11 4.01 4.06 4-06 2.03 6.09 Percentage of theoretical titre 99.8 98-6 98.8 99.4 100.0 100.2 100-5 99.8 The immediate improvement in results, both in accuracy and in freedom from drift of potential after each addition of titrant, was marked; but the end-points were much less sharp and the over-all e.m.f.change was less than in presence of ethanol (see Fig. 1, curve D). Increased concentrations u f sodium chloride-As the analysis of organic fluorine compounds in the range of 40 to 120mg of fluorine would probably necessitate the Parr bomb method for decomposing the sample, the effect of increasing the amount of sodium chloride in the titration system was investigated.The tests shown in Table V were made in presence of 20 ml of N buffer, pH 3.5, and 6 g of sodium chloride at a dilution of 40 ml in the absence of ethanol; the indicator solutions were added a t 0.4 ml before the expected end-point. The end-points in these titrations were as sharp as before (see Fig. 1, curve E), but a rather longer time was required to reach equilibrium potential after each addition of lead nitrate, and the end-point occurred rather too early. Nevertheless, the results are repro- ducible and it is evident that if the lead nitrate solution were standardised against sodium fluoride under these conditions the analysis of unknown fluoride solutions would be at the correct level and would be reproducible to about k0.7 per cent.CROPPER: A STUDY OF THE POTENTIOMETRIC [Vol.76 374 At a later date the effect of variation in the amount of ferrous chloride was investigated; this should have no effect other than a change in the general level of potentials owing to a change in the Fe"' - Fe" ratio. It was found, however, that in the presence of ethanol, 1 g of sodium chloride and buffer of pH 4, correct results (100.0 to 100.5 per cent. of theory) TABLE V RESULTS IN PRESENCE OF 6 c OF SODIUM CHLORIDE Fe"' - Fe" indicator added at, Observed titre, Theoretical titre, Percentage of ml ml ml theoretical titre 1.6 1.97 2.03 97.1 2.6 2.98 3.04 98.0 3.6 4.00 4.06 98.5 4.6 4.95 5-07 97.6 - Mean 97.8 4 0.7 with a sharp end-point could be obtained by reducing the amount of ferrous chloride from 20 to 5 mg, the ferric chloride content of the titration system being kept at 0.4 mg.When the sodium chloride content was increased to 6 g the end-points were very unsatisfactory, the potentials taking a very long time to reach equilibrium, and the results were distinctly i i i i i i I I i i I i 8 i i i i i I j,/ I ID i i I I E i i i I I I I I I 3 4 5 3 4 5 3 4 5 Volume of N Pb (NO,),, ml Fig. 1. Titration curves. Curve A, pH 3.5, 1 g of sodium chloride, alcohol present; curve B, pH 4.0, 1 g of sodium chloride, alcohol present: curve C, pH 6.0, 1. g of sodium chloride, alcohol present; curve D, pH 3.5, 1 g of sodium chloride, alcohol absent; curve E, pH 3.5, 6 g of sodium chloride, alcohol absent low.A reduction in the ferrous chloride content in a titration solution in absence of ethanol, but in presence of 6 g of sodium chloride, gave a low result comparable to those given in Table V, and no advantage was to be gained by reducing the ferrous chloride content under these conditions.June, 19511 TITRATION OF FLUORIDE WITH LEAD 375 TITRATIONS WITH O.1ON LEAD NITRATE (4 TO 16MG OF FLUORINE)- The initial titrations of 4 to 16 mg of fluorine were made in presence of 20 ml of N buffer, pH 4.0, 0.4 mg of ferric chloride and 20 mg of ferrous chloride, at a total volume of about 40 ml, in the absence of ethanol. The sodium chloride content was reduced to 0.1 g and 0.10 N lead nitrate was used as titrant.The titration curves were unsatisfactory; the slope at the end-point being far too small. Reduction of buffer concentration to one-tenth of the previous value, and even in absence of buffer (with methyl orange to indicate adjustment to correct pH after addition of indicator) considerable improvement in the sharpness of the end-point was effected, although it was still not nearly as sharp as in the N titrations and the result was less than 95 per cent. of theory. Reduction in the amount of ferric chloride to one-tenth of the original value, Le., to O.O4mg, slightly decreased the slope of the titration curve, but raised the level of the results to 98 to 98.5 per cent. of theory; as for the N titrations, addition of the ferric - ferrous indicator solutions at the beginning of the titration caused the results to be lower.The tests shown in Table VI were made on 4 to 16 mg of fluorine in presence of 0.1 g of sodium chloride, 20 mg of ferrous chloride and 0.04 mg of ferric chloride, in the absence of ethanol. TABLE VI RESULTS OF TITRATIONS WITH 0.10 N LEAD NITRATE Fe"' - Fe" indicator added at, Observed titre, Theoretical titre, Percentage of ml ml ml theoretical titre 1.6 3.6 5.6 7.6 1.98 3-98 6.00 8.00 2.03 4-06 6-09 8-12 Mean .. 97.5 98.0 98.5 98.5 98.1 2 0-5 - Addition of excess of sodium chloride (1 to 1-5 g) did not alter the level of the result, but slightly decreased the slope of the curve at the end-point. Reduction of ferrous chloride content from 20 to 0-5 mg had little effect in the presence of alcohol, with or without buffer, and in the absence of both alcohol and buffer the slope was poorer, so that the titration as a whole was then considered impracticable. The conclusion reached was that accurate results could be obtained provided the 0.10 N lead nitrate solution was standardised against sodium fluoride under controlled conditions ( i e . , 0.1 g of sodium chloride, 20 mg of ferrous chloride and 0.4 mg of ferric chloride both added at about 90 per cent. of the expected titre and the system neutralised to methyl orange after the addition of indicator), but the end-point is not by any means as sharp as is desirable, and careful attention to detail is necessary to be sure of obtaining reliable results. 1. 2. 3. 4. 5. 6. 7. 8. REFERENCES Hawley, F. G., Ind. Eng. Chem., Anal. Ed., 1926, 18, 573. Hoffmann, J. I., and Lundell, G. E. F., J . Res. Nut. Bur. Stand., 1929, 3, 589; 1938, 20, 610. Elving, P. J., and Ligett, W. B., Ind. Eng. C h e w , Anal. Ed., 1942, 14, 449. Donovan, C. G., J. Assoc. Ofl. Agric. Chem., 1944, 27, 549. Chapman, N. B., Heap, R., and Saunders, B. C., Analyst, 1948, 73, 434. Kaufman, S., Anal. Chem., 1949, 21, 582. Farkas, L., and Uri, N., Ibid., 1948, 20, 236. Furman, N. H., Ind. Eng. Chem., Anal. Ed., 1942, 14, 367. IMPERIAL CHEMICAL INDUSTRIES LIMITED DYESTUFFS DIVISION ANALYTICAL LABORATORIES HEXAGON HOUSE, BLACKLEY MANCHESTER, 9

ISSN:0003-2654    

DOI:10.1039/AN9517600370

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

376 SYMPOSIUM ON BI0:LOGICAL ASSAY [Vol. 76 Symposium on Biological Assay AT a joint meeting of the Biological Methods Group of the Society and the Biometric Society (British Region) held on March 14th, 1950, the following papers were read and discussed, under the chairmanship of Dr. J. W. Trevan, F.R.S.- “The Graphical Calculation of the Errors of Biological Assays with Graded Responses,’’ “The Estimation of Error in Certain Types of Biological Assay,” by E. C. Wood. “The Problem of Combining the Results of Independent Assays,” by E. C. Fieller. Afterwards, under the chairmanship of Mr. N. T. Gridgeman, the meeting devoted its time to an unscripted discussion of a number of questions, raised by previous invitation, on matters of technique and interpretation not normally covered by the standard procedures.A rksum6 of the discussion follows. by N. T. Gridgeman. Question-How can non-parallelism of response curves be dealt with ? A nswev-The Chairman said he believed that the inescapable inference from non-parallelism of the standard - response and test-material - response curves was that some or all of the active principle in the test material differed from that in the standard not only chemically, but in quanti- tative action ; therefore an assay exhibiting non-parallelism was, normally speaking, biometrically invalid. To this, Mr. M. J. R. Healy added the rider that non-parallelism had in fact been used to detect the presence of foreign substances in pharrnacological preparations of uncertain composi- tion. Dr. H. 0. Schild feared that this line of argument led to the conclusion that only substances chemically identical with the standard preparation could validly be assayed, yet in practice this was not so; for example, vitamin D, could be assayed against vitamin D,.Mr. A. L. Bacharach drew attention to, and Drs. J. I. M. Jones and E. C. Wood enlarged upon, the modern differentiation of bio-assay into two types whose recognition avoided the difficulties that Dr. Schild had pertinently mentioned. The first type was that in which the assayist knew the chemical composition of the active principle in his test preparation to be identical with that in his standard; the “unknown” was its concentration in the inert diluent, and this the assay was designed to estimate. The second type was that in which the assayist could not assume chemical identity but could assume identity of biological behaviour; here the assay measured relative biological potency.Mr. D. J. Finney, referring to the second type, said that the understanding of the results might sometimes be helped by a study of the relationship between relative potency and level of response. Question-How can the atypical response curves occasionally obtained in microbiological assays be dealt with? Answer-This phenomenon, found in some assays of, for instance, pantothenic acid and folk acid, was difficult to handle because the atypical curves often traced no recognisable mathe- matical function. To evaluate such assays it was usual to draw the best freehand curve for the standard and to read off “equivalent potencies” for each test material response.This was unsatis- factory. Dr. A. F. Parker-Rhodes remarked that abnormal curves could be mathematically linearised-provided that the step was biochemically justifiable. Several speakers agreed with Mr. S. A. Price that progress would be difficult until it could be decided how far the question was statistical and how far biochemical. Dr. E. C. Wood and the Chairman could see no theoretical objection, on biochemical grounds, to linearisation procedures. Dr. Wood added that if abnormality of shape was confined to a falling off of the lower part of the curve, the addition of a trace of the assay factor to the basal medium might produce linearity. Question-What is the recommended design for an experiment in which the number of dosage groups exceeds the number of animals per litter? A nswer-Complete randomisation of the animals available for test was one suggestion; in this way between-litter and within-litter variances are confounded.The Chairman linked the idea with those presented by Dr. Wood in the paper he had delivered earlier in the day, and pointed out that if only one animal be taken for test from each litter, the assay would not only present no problems of design but would be a more representative assay than one employing litter mates. In practice, however, difficulties of animal economy might arise even if biological work was available to absorb the litter supplies; there was, therefore, much to be said for recourse to the more intricate designs available for multi-assays with litters of smaller membership than the number of dosageJune, 19511 SYMPOSIUM ON BIOLOGICAL ASSAY 377 groups required.Mr. Finney said that the problem was similar to that faced by agricultural experimentalists wishing to compare more treatments than there were plots per block. The “incomplete block” techniques could therefore be used. Mr. J. V. Smart and Mr. Healy warned of the danger of confounding interactions of importance, and the former held that pilot experiments should be conducted to decide which interactions might safely be confounded in the assays proper. Question-Is the partial rejection of results justifiable ? What is the correct method of dealing with individual animals that have lost weight in assays based on weight gains? Answev-This query might be said, as Dr.Jones commented, to go beyond the field of statistics into that of ethics, because a decision on a rejection that would affect the assay result was often a matter for the assayist’s conscience. Abnormality of unknown cause was the major problem, although animal abnormality of known cause (eg., injury or misdosage) posed the ancillary question as to whether, in that event, the whole group or litter should be discarded. The Chairman and others thought not. Attempts had been made in the past to set up objective statistical criteria of abnormality and rejection, but none was unexceptionable. It had been suggested that a result was validly rejectable if its withdrawal reduced the fiducial limits, but that was not a reliable test.Mr. Finney said that it was almost impossible to get away from ethical judgments unless one firmly adopted the objective criterion that rejection was permissible only for assignably-caused abnormality. Others concurred. On the second part of the question, the Chairman said that in growth assays, weight losses were particularly difficult to deal with when there were also deaths. An animal that died must have a lower response than one that lost more weight and yet survived, but simple inclusion of negative growths would not allow for this. The treatment he favoured was to record all weight losses and deaths as “zero growth.” Mr. Healy and others thought this solution imperfect, but agreed that a biometrically valid treatment was not obvious. Question-In the radiographic technique for vitamin-D assay, animals of abnormally low sensitivity to rachitogenesis are not used for test.In the rat-growth assay of vitamin A, animals of abnormally high sensitivity to vitamin-A depletion are also disliked , but are generally included in the test because it is usually only during the test period that this abnormality shows up. Should the results of these animals be rejected ? Should all the results of the containing litter be rejected ? Answer-On this question, which is in some respects an extension of the previous one, the concensus of opinion yielded a negative answer. That is to say, results should not be rejected on circumstantial evidence of this kind. It was nevertheless admitted that a litter, most of whose members responded feebly or died during the test period, might justifiably be rejected from the final results.Question-In some biological phenomena (e.g., the effect of insulin on blood sugar or the effect of vitamin P on capillary resistance) reponse to dose or stimulus is conditioned by the initial status. Answer-Mr. Finney and Dr. Wood, among others, contributed to the discussion on this question. They agreed that, in general, recordings of the initial status should be made and the responses subsequently adjusted by means of co-variance analysis and the regression equation (response on stimulus). By this means the equivalent responses from animals of uniform initial status could be estimated. The Chairman added that distribution of the animals into the dosage groups in such a way that the average initial status was the same for all dosage groups usually provided a satisfactory solution; if necessary, of course, the co-variance technique could also be applied-it might reduce the limits of error, although it would be unlikely to alter the assay.Mr. C. P. Cox pointed out that the relation between initial status and response could be non- linear, in which event co-variance analysis involving a quadratic, or even high order, equation would be needed. He agreed with Mr. Fieller that the considerable extra computation involved was rarely justified. What are the governing principles? How can the conditioning be allowed for? Question-Toxicologists in particular, and pharmacologists in general, frequently administer assay doses in proportion to body weight.Is not the correct procedure, unless experiment has shown a direct proportionality between response and body weight, to give equal doses within groups, and to eliminate the effect of body weight (as of any other measurable independent variable) by co-variance analysis ?378 NOTES [voi. 76 Answer-The correct procedure, it was agreed, is to give equal doses and to‘ use co-variance A not unimportant point made incidentally was that co- analysis to allow for size differences. variance analysis of the results involved less time and trouble than dosage by weight. Question-Can a statistical basis be given for the choice of levels in techniques involving stepwise dilution? This problem arose in tests designed to compare the efficacy of various sulphonamides as faecal bactericides.Pooled faeces from groups of mice fed on the drugs were suspended in peptone water and serially diluted in steps that each halved the concentration. The tubes were autoclaved and infected, and the degree of bacterial growth after different periods of incubation were recorded, in one experiment, as (-), ( - +), (+) and (+ +). In a second experiment, in which the faeces from the individual mice of a group were treated separately, the metameter was the proportion of tubes per group that showed growth. A continuous variate was not possible; turbidimetry, for example, was ruled out by the presence of suspended faecal particles. Answer-Mr. Healy, referring to the second of the two experiments, suggested the use of probits as in toxicity tests; this would sometimes facilitate the choice of optimum levels, Mr. Finney pointed out that the results of the second trial could be regarded as scores, for which there were standard methods of approach. The same speaker later proposed that the maximal bacteriostatic dilutions (or similar values) of the faeces in that trial be converted to logarithms and that the “t” test should be applied to the figures so obtained. It was agreed that the second trial lent itself to statistical treatment more readily than the first.

ISSN:0003-2654    

DOI:10.1039/AN9517600376

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

378 NOTES Notes NEW SPOT TESTS FOR ZINC THE addition of zinc ions to a solution containing ferricyanide and ferrocyanide causes an increase in the oxidation potential of the system because of the removal of ferrocyanide as insoluble zinc ferrocyanide. Under these conditions it is possible to oxidise certain organic compounds, which in the presence of ferricyanide alone are oxidised but slowly or not a t all, to give highly coloured products. Several tests for zinc based on this reaction have been described, in which organic compounds such as diethylaniline,l p-phenetidine2 and diphenylamine3 are used, Recently, we have described the use of naphthidine* (4 : 4-diamino-1 : 1’-dinaphthyl) as an indicator in the titration of zinc with ferrocyanide in the presence of ferricyanide, the colour change a t the end-point depending essentially on the reaction just described.In our opinion, naphthidine is superior to all other indicators hitherto used in this titration. We have since examined naphthidine and its homologue 3 : 3’-dimethyl naphthidines (3 : 3’- dimethyl-4 : 4’-diamino-1 : 1’-dinaphthyl) as reagents for the detection of zinc. The sensitivity of the test with naphthidine is of the same order as that given by other well-known tests based on the same principle, but 3 : 3’-dimethyl naphthidine provides a much more sensitive test. The limits of identification and the concentration limits for both reagents are shown in Table I. TABLE I Reagent Identification limit Concentration limit Naphthidine . . .. .. 1 Pg 1 in 50,000 3 : 3’-Dimethylnaphthidine .. 0.1 Pg 1 in 500,000 Ions that form coloured precipitates with the ferrocyanide ion will interfere, i.e., ferrous Ions that will oxidise the reagents must With naphthidine, the test is most sensitive under the slightly acid conditions furnished by Under more strongly acid conditions the sensitivity is decreased to Hence, using the reagent 0.1 N with respect to sulphuric or hydrochloric acids, With 3 : 3’-dimethylnaphthidine, the sensitivity is not affected by a similar increase in acid and ferric iron, copper, manganese, cobalt and nickel. be absent, e.g., permanganate, dichromate, chromate, vanadate, persulphate and iodate. the solution of the reagent. a small extent. the identification limit falls to 3 pg. concentration.June, 19511 APPARATUS 379 METHOD- Reagemi-Mix 1 part of a 5 per cent. solution of potassium ferricyanide in distilled water with 2 parts of a saturated solution of naphthidine hydrochloride or 3 : 3’-dimethylnaphthidine hydrochloride in distilled water. Procedure-Place 1 drop of freshly prepared reagent on a spot plate and add 1 drop of zinc test solution. An immediate red-violet colouration denotes the presence of zinc. REFERENCES 1. 2. 3. 4. 5. THE UNIVERSITY Eegriwe, E., 2. anal. Chem., 1928, 74, 228. Szebell6dy, L., and Tanay, I., Ibid., 1936, 106, 342. Cone, W. H., and Cady, L. C., J . Amer. Chem. SOC., 1927, 49, 2214. Belcher, R., and Nutten, A. J., J. Chew. SOC., 1951, 547. Fries, K., and Lohmann, W., Ber., 1921, 54, 2922. DEPARTMENT OF CHEMISTRY BIRMINGHAM, 15 R. BELCHER A. J. NUTTEN W. I. STEPHEN November, 1950

ISSN:0003-2654    

DOI:10.1039/AN9517600378

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

June, 19511 // u - 4 c r n 3 , APPARATUS 379 Apparatus A NEW DESIGN FOR AN ORSAT ABSORPTION PIPETTE MANY different types of gas absorption pipettes for use on Orsat-type apparatus have been described in the literature, some of which incorporate improvements such as the use of perforated plates or glass sinters that considerably increase the efficiency of absorption. Most of the I E I U i f 5 L e A 9 B h J ,D Fig. 1. New design for an Orsat absorption pipette Copper Spirals Copper Rods Fig. 2. Modified pipette for use with cuprous chloride in presence of copper pipettes require the use of two-way stopcocks, which necessitate a 180' turn for each passage of the gas to and from the gas burette, and an undesirable feature common to most is the U-bend380 APPARATUS [Vol. 76 connecting the absorption chamber to the reagent reservoir.This U-bend design is fragile and special clamps or supports are needed to hold the pipette in position on the apparatus. The pipette described was designed in an attempt to overcome these limitations and to produce a gas absorber that would combine maximum efficiency of absorption with simplicity of attachment, operation and maintenance. The apparatus, illustrated in Fig. 1, consists of a cylindrical Pyrex glass tube divided into three compartments by means of two internal seals. The compartments A and B are a modification of the Tsikliss pipette,l and the lower compartment, C, is the reagent reservoir vented to air by the side-arm D. For concentrated sulphuric acid reagents, the central capillaries are in heavy- walled 2-mm internal diameter tubing, and for aqueous reagents, l-mm internal diameter tubing.In operation the gas is passed into the pipette through a single-bore stopcock and bubbles into the compartment, B, through the long capillary tube E, the reagent being displaced into the reservoir. The gas sample collects in compartment B, the liquid level in A being depressed only to the upper end of the long capillary. On withdrawing the sample into the burette, the gas bubbles from the short capillary, F, through the reagent in compartment A. When the pipette is incorporated in a complete Qrsat assembly, absorption of carbon monoxide is normally effected by means of acid cuprous chloride reagent, reduced with stannous chloride. If it is desired to use ammoniacal cuprous chloride, which requires the presence of copper for reduction, the modified design shown in Fig.2 can be used. A tubulure is fused to the bottom of the pipette through which pulled out spirals of copper wire are dropped into the middle compart- ment, the spirals being kept in place by copper rods, as shown, and the opening is closed with a rubber bung. The dimensions given are for a pipette dealing with an initial gas sample volume of 100 ml. The structure of the pipette is robust and compact, and consequently is suitable for inclusion in a multiple absorption apparatus, the connection to, and removal from, the manifold being greatly facilitated without risk of breakage. Attachment to the apparatus frame is made by means of two clips. Cleaning and filling can normally be undertaken through the side-arm without removal of the pipette.The single bore stopcock appreciably reduces the trouble experienced with two-way and double-oblique stopcocks. The double scrubbing action reduces the number of operations and gives increased efficiency of absorption. Our thanks are due to Mr. E. Kennedy who constructed the apparatus, and we express our appreciation to the Directors of the British Oxygen Company, Ltd., for permission to publish this note. REFERENCE 1. Tsikliss, D. C., Chim. et Ind., 1947, 58, 259. ANALYTICAL LABORATORY RESEARCH AND DEVELOPMENT DEPT. BRITISH OXYGEN COMPANY, LTD. MORDEN H. W. HODGSON W. ANGEL November, 1950 A DROPPING ELECTRODE SYSTEM IN most types of dropping mercury electrode the purified mercury is in contact with rubber or plastic tubing.Attempts to overcome this drawback have resulted in several electrode designs. Kahanl has proposed an all-glass apparatus in which mercury is siphoned from the storage vessel to the capillary. Variations in mercury head are made by alteration of the height of the storage vessel with respect to the capillary. The flow may be stopped by the back pressure provided by a test tube, filled with pure mercury, fitting tightly on a ground joint above the capillary. Other workers adopt the principle of applying gas pressure to the mercury in order to attain the necessary head. Lingane and Kolthoff2 describe an apparatus in which the pressure is applied to the mercury by two levelling bulbs, the head being measured by a manometer.Muller, Garman, Droz and PetrasS use a pressure bottle and rubber bulb to attain the mercury head; a similar device is described by McReynolds,4 who incorporates a three-way stopcock just above the capillary, thus providing a means of reversing the flow of mercury through the capillary. The lubrication of this stopcock is not described, but could be a source of contamination. The electrode of Kahan is of a fragile nature and the method of stopping the flow could con- ceivably be a source of hazard. Lingane and Kolthoff’s design is cumbersome, and those of Muller and McReynolds are liable to leak.June, 19511 APPARATUS 381 The apparatus proposed in this note is presented in an attempt to overcome the defects of the assemblies dependent on applied pressure.It is robust, compact, and may be supported by a single clamp. The mercury flow is readily stopped, and the assembly allows for quick dismantling of the capillary attachment and replenishing of mercury. The apparatus, Fig. 1, is constructed in two parts; the tube A, fabricated as shown, is sealed to the polarographic capillary. The whole is thoroughly cleaned and pure mercury introduced. The ground joint is well greased and the tube inserted into jacket B. The centre tube of the jacket dips into the pure mercury in A. Socket Fig. 1. Mercury head and reservoir for dropping electrode The side-arm on B is connected by a length of pressure tubing to a levelling bottle. Ordinary mercury is introduced and, by raising the levelling bottle, pressure is applied to the pure mercury in tube A and causes it to rise up the central tube to give the required head of mercury.Electrical contact is made by means of a platinum wire sealed through the wall of the central tube. This type of dropping-electrode assembly has been found satisfactory in use and no contamina- tion of the mercury occurs over long periods, as the pure mercury is only in contact with glass surfaces. The use of thick-walled capillary tubing is recommended for the central tube; this adds to the robustness of the apparatus and also enables accurate measurement of the mercury head. Tube A is replenished with pure mercury by lowering the levelling bottle to drain the mercury from the jacket, B, the tube,,A, is then withdrawn and more mercury is added.382 MINISTRY OF FOOD [Vol. 76 Our thanks are due to Mr. E. Kennedy who constructed the apparatus; and we express our appreciation to the Directors of the British Oxygen Company, Ltd. for permission to publish this note. REFERENCES 1. 2. 3. 4. Kahan, G. J., Ind. Eng. Ckem., Anal. Ed., 1942, 14, 549. Lingane, J. J., and Kolthoff, J. M., J . Amer. Chem. Soc., 1939, 61, 825. Muller, R. N., Garman, R. L., Droz, M. E., and Petras, J., Ind. Eng. Chem., Anal. Ed., 1938, McReynolds, R. C., Ibid., 1942, 14, 586. 10, 339. ANALYTICAL LABORATORY RESEARCH AND DEVELOPMENT DEPT. BRITISH OXYGEN COMPANY, LTD. MORDEN H. W. HODGSON J. H. GLOVER November, 1950

ISSN:0003-2654    

DOI:10.1039/AN9517600379

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

382 MINISTRY OF FOOD [Vol. 76 Official Appointments PUBLIC ANALYST APPOINTMENTS NOTIFICATION of the following appointments has been received from the Ministry of Food since the last record in The Analyst (1951, 76, 320). Public Analyst Appointments DALLEY, Richard Arthur (Deputy) . . .. JAFF~, Frederick William Moore . . I . LEWIS, William Maurice. . .. .. .. LEWIS, William Maurice (Deputy) . . .. LEWIS, William Maurice (Deputy) . . .. WOODHEAD, John Ezra . . .. .. .. WOODHEAD, John Ezra . . .. .. .. TURNER, Mervyn Edward Dennant (Additional) County Borough of Wakefield. County of West Riding of Yorkshire. County of Warwickshire. County Borough of Walsall. Borough of Hereford. County Borough of GIoucester. Metropolitan Borough of Lambeth. County of the Pads of Holland, Lincolnshire. OFFICIAL AGRICULTURAL ANALYST APPOINTMENTS NOTIFICATION of the following appointments has been received from the Ministry of Agriculture and Fisheries since the last record in The Analyst (1951, 76, 320). Agricultural Analyst Appointments LEWIS, William Maurice (Deputy) . . . . County of Wanvickshire. LEWIS, William Maurice (Deputy) . . . . County Borough of Walsall. WOODHEAD, John Ezra . . .. .. . . County of the Parts of Holland, Lincolnshire.

ISSN:0003-2654    

DOI:10.1039/AN951760382a

出版商:RSC    

年代:1951

数据来源: RSC

摘要:

June, 19511 BOOK REVIEWS 383 Book Review INORGANIC MICRO-ANALYSIS. BSc., Ph.D., D.I.C., F.R.I.C. 1950. Price 12s. 6d. By H. V. A. BRISCOE, DSc., A.R.C.S., F.R.I.C., and P. F. HOLT London: Messrs. Edward Arnold & Co. The techniques of small-scale analysis are not difficult to acquire, nor do they often need, with the exception of the microchemical balance, equipment that the student cannot devise for himself. The best way to acquire such techniques is through the personal guidance of an expert; this book is intended not only to supplement such teaching but also to guide the student who must teach himself. The qualitative section of the book deals, not strictly with the “micro” scale (no separations on drop quantities are described), but with the more generally useful “semi-micro” scale, involving centrifugal separations in volumes of apparently 0.25 to 2ml.The techniques described are basic ones and the student may well, with familiarity, improve and simplify some of them. The scheme of cation analysis applies to the academic “common” ones and consists of the usual group separations, followed by identification within a group by spot and drop tests without, in general, further separation. It could be argued that the emphasis on spot tests (a third of the qualitative section deals with them) is not in the best interests of teaching, since a student’s chemical knowledge does not benefit by learning many of the reactions involved. More cogently, although such tests have an invaluable and increasing use in analysis, it can be maintained that, without adequate warning, they are often misleading to the student in their excessive sensitivity and their inability to indicate relative quantities. The quantitative section includes volumetric methods making use of the vertical and the horizontal burette, and there is a useful description of the application of micro-diffusion methods to the determination of ammonia, carbonate, chloride and bromide.Gravimetric methods make use of the microchemical balance, although students might benefit by preliminary experiments on an ordinary analytical balance, using the method of swings. It is not easy to justify the inclusion in such a book of a lengthy description of the photo-electric absorptiometer and experi- ments with it. Its unquestioned value in analysis is not sufficient reason for singling it out from other instruments of equal importance. Again, the determination of carbon in steels by a micro- combustion method seems rather specialised for the level aimed at. The text is agreeable, the illustrations clear, the index detailed and the price not unreasonable under present conditions. The score or so references, however, are generally to minor points, and the student would have benefited by a list of more specialised textbooks which would amplify the various sections. It is not possible in a book of this scope to satisfy all readers, but the student will find it constantly valuable whether he can receive personal tuition or not. Pp. vii + 171. DAVID W. WILSON

ISSN:0003-2654    

DOI:10.1039/AN951760383b

出版商:RSC    

年代:1951

数据来源: RSC