摘要:

JANUARY, I969 THE ANALYST Vol. 94, No. I I 14 Determination of Fluorine by Prompt y- Radiation Proton Bombardment Part I. Theory and Experimental Method BY J. M. BEWERS AND F. C. FLACK (Physics Department, University of Exeter, Exeter, Devon) from An analysis is presented of those factors which are important for quanti- tative determination of fluorine by means of proton bombardment. An expression for the limit of detection has been derived, and the important part played by the stopping cross-section pointed out. The experimental arrangements are discussed in preparation for the presentation of results obtained with fluorine-containing solids, liquids and gases, which are given in Part I1 of this paper. A FEATURE of many nuclear reactions induced by bombarding particles of variable energy is the presence of multiple resonances.Reactions with strong resonances have been used as an extension of conventional neutron and y-ray activation analysis by several authors.l,2,5,4,536,7 98 In the instance of charged particles with energy of about 1 MeV, whose range in solids is small, this technique has recently been developed as a system of surface analysis.9 A group of workers at A.E.R.E., Harwell,lo among others, is devoting-considerable effort to an examina- tion of the general field of quantitative and qualitative micro analysis of surfaces by charged- particle beam scanning. The reaction 1QF (p,ay) l60 has a higher yield of y-radiation for incident protons with energy of less than 2 MeV than any other proton-induced reaction. It therefore warrants study in its own right, and is a suitable reaction for a general estimate of the merits and problems liable to be encountered in this field.Accordingly, the theory underlying the method and the practical arrangements required are reviewed here. Part I1 of this paper presents the results from experimental determina- tions of fluorine by proton bombardment in gases, liquids and solids. EXPERIMENTAL BASIS In principle, there are two features on which identification and measurement of a material in a given target can be based, assuming proton bombardment; first, the resonance energy, which is generally determined within a few keV; secondly, the y-ray spectrum, which is unique both for a given isotope and a given resonance. With fluorine, y-rays of high energy (6.13, 6.7 and 7.1 MeV) are produced, the relative amounts depending on the particular resonance, although the 6-13-MeV line is generally predominant.Basically, the procedure involves preparation of the sample in the form of a layer, which may be either thick or thin with respect to the range of the protons in the material (a few microns). For these studies, a thick layer, in this sense, is probably preferred, even although only a thin layer on the surface will, in any event, be sampled by the method. The accelerating voltage can then be set to deliver equal numbers of protons of energy just above and then just below the resonant energy. The difference in the corresponding y-ray spectra (recorded, e.g., in a scintillation or junction detector) gives the contribution from that resonance, being specific to the particular reaction under study.In such a difference spectrum, the intensities of the y-rays related to the substance sought are directly proportional to the amount present , assuming normal isotopic abundances. 0 SAC and the authors. 12 BEWERS AND FLACK: DETERMINATION OF FLUORINE BY PROMPT [ArtdySt, VOl. 94 Because of the high energy of the fluorine radiation, it has not been found necessary, either in this work or in work reported by earlier investigators, to use the difference method. Rubin, Passell and Bailey1 simply compared the y-ray yield of sodium fluoride standards with that from fluorine in glass, and claimed a precision of k0.1 per cent. at the 3 per cent. level. By similar means, Sippell and Glover2 used proton bombardment for determining fluorine, among many other elements.With 2-MeV protons they covered a range of solids containing 200 p.p.m. up to 2.5 per cent. of fluorine, and determined the limit of detection as 100 p.p.m. In 1966, Moller and Starfelts carried out surface analysis for fluorine contamination in various zircalloy samples by using the 19F (p,ay) le0 reaction. The samples contained surface fluorine in the range 0.15 to 0.6 pg, and the dstribution in depth was investigated by varying the proton energy, thus altering the depth at which protons reached a particular resonant energy (1375 keV). Pierce, Peck and Cufflo have also published results showing a linear relationship between y-ray yield and fluorine content for manufactured targets in the range 0 to 500 p.p.m. THEORETICAL ASSESSMENT OF LOWER LIMIT OF DETECTION- In ordinary activation analysis, a typical thermal-neutron flux may be 10’2 neutrons per cm2 per second. Consider sodium-23 as a typical capturing nucleus of activation cross- section 0.54 barn.For comparison, a beam of 1 pA of protons consists of about 6 x 10l2 particles per second, and the #eak cross-section for the reaction 19F (p,ay) lSO, at the 340-keV resonance, is 0.16. barn. Further, the resulting high-energy radiation is not much spread out in time, being present only during bombardment. Thus, for a sample of sodium in the neutron flux for 15 hours, then cooled for 15 hours (one half-life), the resulting activity is about 3 x lo9 disintegrations per second per gram of sodium.The y-ray yield from a “thick” target for fluorine is 2 x 106 photons per second. It will be appreciated that the reacting nuclei all lie in the first few microns of the bombarded layer. The count-rates from equal sampled masses would, in the given circumstances, be comparable. The calculation of the y-ray yield may follow the standard treatment of Fowler, LauIitsen and Lauritsen.11 Assume that the Breit-Wigner dispersion equation describes the variation of reaction cross-section (a), with energy ( E ) , viz., 1-14 where oR is the cross-section at resonant energy ER and I’ the full width of resonance at half maximum intensity. This equation applies to simple proton-capture reactions, and a similar equation will apply, with provisos, in an instance like that of fluorine bombarded by protons, when the y-radiation is emitted from a nucleus generated by a-particle emission.The provisos are that the resonance must be narrow and well separated from adjacent levels, and that none of the reaction parameters (e.g., a-particle emission energy) varies appreciably over the resonance. Suppose the target material to be of such thickness that a proton of incident energy E loses energy F in passing through it. The stopping cross-section for the incident protons per disintegrable fluorine-19 nucleus is given by 1dE N dx’ E = - i.e., E is the energy loss (a) per unit path length (dx) divided by N , the number of dis- integrable nuclei per unit volume. The yield, Y, of y-rays per incident proton is then Provided E and elementary integration gives do not depend on E over the energy range covered by the (single) resonance,January, 19691 y-RADIATION FROM PROTON BOMBARDMENT.PART I. THEORY 3 The maximum value of Y is reached when E = ER + E,/2 when, for a given value of e, o r Y(max., (5) = -$- tan-1 E/r, provided that > I?. Thus, the largest yield arises when E, is effectively infinite, viz., URr Y(max.) (00) = 4 2 -- E It is with the aid of this equation that thick target yields can be evaluated. Absolute yields from calcium fluoride (CaF,) targets have been reported by Fowler, Lauritsen and Lauritsenfl and by Chao, Tollestrup, Fowler and Lauritsen,la most results being available for resonances at 340, 872 and 935 keV. van Allen and Smith,ls as well as Bonner and Evans,14 have also reported absolute y-ray yields at 340 keV, their results being slightly lower than those of Fowler, Lauritsen and Lauritsen.The relevant data on the level parameters, drawn from the compilation of Ajzenburg- Selove and Lauritsen,16 are shown in Table I. Resonances up to 1090 keV are included, and the more recent results of Seagrave, Brolley and Beery,la and Marion,17 on the widths of the 340 and 872-keV resonances have been used. TABLE I RESONANCES FOR 6 TO 7-M~v y-RAYS I N 19F (p,ay) AFTER Relative component intensities r E P* 0, r. YCaF, 6.13 6.7 7-1 keV mb keV MNe* x 108 MeV MeV MeV 224 0-2 1.0 13.86 0.002 - - - t 340.5 102 2.4 13.196 1.74 0.96 - 0.04 57 6.0 13.511 4.8 0.76 0.06 0.19 A JZENBURG-SELOVE AND LAURITSEN" L \ 32 0.9 13.332 0.54 0-8 - 8.2 7-1 30.0 1 3 4 0 2.7 671.6 834.8 19 6-5 13.667 2.0 902.3 23 5- 1 13-730 1.5 0-95 0.06 - 935.1 180 8.6 13.761 20 0-76 0-026 0-2 1 1090 13 0-7 13-909 0.24 - - - Total yield to- 450 keV 1 -742 1050 keV 70.282 1600 keV 21 5.1 82 {it;:: - - - t 872.1 66 1 4-7 13.703 37 0.65 0.24 0.1 1 * The ratio of yield at 1050 keV to that at 450 keV is 40.35, and agrees well with that of t Energies and widths according to Mari0n.l' 41.57 by van Allen and Smith.l* If, near the point where the protons strike the target, a y-ray detector of efficiency q(Ey) for radiation of energy EY (q = fraction of y-rays emitted by the target that interacts with the detector) then, for most radiation detectors, a fraction, f, of all the pulses produced will fall in the "peaks" selected by the spectrum analysis system.A correction factor, p, will be needed to account for absorption in the target container, etc., and perhaps also to allow for reactions whose radiation is not isotropic about the proton-beam direction. Thus, the number of y-ray counts, Nu, that contributes to the desired energy range in the spectrum is given by N Y = fr.l?lYfl where rt is the number of incident protons. The particular value of Y depends on the stopping power per fluorine atom (E) of the target; at 340 keV the value of Y is YM0 = - 3*85 x 10-22 y-rays per proton.4 BEWERS AND FLACK : DETERMINATION OF FLUORINE BY PROMPT [ArtdySt, VOl. 94 For a pure fluorine compound A,BnFq, the stopping cross-section per fluorine atom is and so, at 340-keV bombarding energy, one expects where t is the bombardment time in minutes and I the current of protons in microamps (assumed steady).If the radiation detector exhibits a background count in the appropriate energy band of B counts per minute, and the number of reaction events is said to be detectable when it reaches a number equal to the background, then the limit of detection L is given by N , = 3-75 x 1014trfc1?1~ x Cp.p.m. B 3-75 x 1014 Itjpr)y L = where C is the number of parts per million (w/w) of fluorine in the substance AmBnFq,i.e., atomic weight of fluorine c = q x molecular weight of AmBnFq' Taking calcium fluoride as typical then, for a beam current of 0.3 PA, a background of 200 counts per minute and detection efficiency ( f p ~ ) of 5 per cent., one obtains, at 340 keV, incident energy La40 = 1000 p.p.m. and, at lo00 keV, incident energy Llooo~= 27 p.p.m.In practice, background is the limiting factor, and the sensitivity can only be improved by greater detection efficiency, better shielding and the use of increased beam currents. Limita- tions on beam current are imposed by the amount of heating and damage tolerable in the target or beam exit foil. EXPERIMENTAL ARRANGEMENT COUNTING ASSEMBLY AND ACCELERATORS- A 500-keV Cockcroft-Walton accelerator was kindly made available by A.E.R.E., Harwell, for studies at the 340-keV resonance in fluorine. The 2 and 5-MeV van de Graaff accelerators at A.E.R.E. were also used. A standard 3 x 3-inch sodium iodide crystal and photomultiplier was used in all of the experiments, and the multi-channel analyser, together with a single-channel analyser, were set to accept the full energy peak, together with the first and second escape peaks from the 6 and 7-MeV radiations.TARGET ARRANGEMENTS AND CHARGE COLLECTION- Targets were made up in a variety of ways, which are described in the appropriate sections in Part 11. The portion of the target assembly receiving beam current was well insulated from the proton accelerator flight-tube. The total charge received during a measurement was recorded on a Type 3008, Harwell beam-current monitor, which displayed errors of less than 1 per cent. between ranges. To avoid possible errors caused by leakage current, the minimum beam current used was amp. Counting was automatically stopped when a pre-determined charge had passed to the target.Precise collimation of the beam is essential, as only the target must be bombarded. Protons incident elsewhere than at the target, but which are measured by the target current integrator, will produce errors. Beam wander across the target must be minimised to prevent changes in detector-source geometry. In practice, all collimating stops expose unwanted material to the beam. Despite the fact that stops may be placed at points fairly remote from the detector, which can be shielded from them, the y-ray background may be increased. Throughout this work, high purity tantalum showed least background when used as stop material, and a 2-inch lead cylinder was placed round the crystal detector (generally used at the 90" position with respect to the beam) to shield it from radiation produced at the stops. Liquid nitrogen traps were used on beam lines to reduce target contamination and also to prevent material volatilised from target surfaces from contaminating the flight-tube.amp, the maximum being 5 x BEAM TRANSPORT-January, 19691 7-RADIATION FROM PROTON BOMBARDMENT. PART I. THEORY 5 CALIBRATIONS- In most instances reported here, targets were "thick," and it was important to ensure that the incident protons were well above the resonant energy. Accordingly, excitation curves were plotted with high fluorine content materials, typical curves being shown in Fig. 1. Such measurement was especially important when proton beams were brought out through exit foils as, occasionally, foils broke or needed to be replaced, and any slight variation in foil thickness could seriously affect the emergent proton energy.Foils appeared to have fairly uniform thickness, as judged by the straggle in proton energy, revealed by the rising edge of the thick-target excitation functions, with targets placed first in front of and later behind the foils. Such straggle amounted to about 2-5 per cent. in energy at 1600-keV incident energy when the foils were about 700 keV thick. 41 1 350 400 450 Proton energy, keV Fig. 1. 340-keV Resonance in reaction '@F@, 0 c y ) ~ ~ 0 , fluorspar target: A, by direct bombardment; and B, by bombardment through a 0.000026-inch exit foil STOPPING CROSS-SECTION- The y-ray yield has been shown to depend on the stopping cross-section per fluorine atom (q). The particular value assumed by E in a given target is illustrated in Table 11, where the stopping cross-sections and corresponding yields are given for various pure fluorine compounds at the 340-keV resonance.The stopping cross-sections were derived from the publication of Whaling,l8 and it was assumed that the additivity rule applies to such cross- sections, as indicated in results published by the present a~th0rs.l~ TABLE I1 THEORETICAL Y-RAY YIELDS FROM SOME PURE FLUORIDES AT THE 340-KEV RESONANCE IN THE REACTION 'QF ( p , a ~ ) "0 Material Fluoride ion .. .. Beryllium fluoride .. Calcium fluoride . . .. Potassium fluorotitanate . . Mercury(I1) fluoride . . Lead fluoride .. .. Caesium fluoride . . .. Mercury(1) fluoride . . Chemical form F -Fa CaF* K,TiF,.H,O Hg Fa PbF, CsF HgPa Stopping cross-section x 10l6, eVcma per F atom 11.6 14.1 22-1 24-3 29.0 31.1 41.0 46.4 y-ray yield x y per proton 3.32 2.73 1 *74 1.58 1-33 1 -24 0.94 0.83 108, In a substance that consists of fluorine-containing grains mixed with grains of other materials, the stopping cross-section at any instant will depend on the particular substance in which the proton is found.Therefore, great importance attaches to grain sizes compared6 BEWERS AND FLACK with the range of the protons during their resonant phase. In most analyses in solid materials, the sample is a mixture of grains that are extremely large compared with the resonant range of the bombarding protons. In such a situation, a bombarding proton is retarded below resonance energy in any single grain and the analysis is effectively surface analysis.As such grains consist, usually, of large numbers of molecules of a single type, the reaction yield is inversely proportional to the E value for this particular molecular type.19 When the bombarded sample consists of an inactive matrix, in which there are small numbers of individual fluorine atoms or well dispersed single fluorine-containing molecules, the proton is presented with a number of reacting centres proportional to its resonant range in the matrix. Thus in a weak solution, the only significant contributor to the stopping power is the matrix, as the atomic or molecular dimensions are small compared with the resonant range of the proton. As the concentration of the fluorine-containing solute is increased, corrections must be applied for the increasing r61e of these atoms or molecules in the stopping processes.It is clear, therefore, that a good deal of information is needed about the form of the fluorine and the materials with which it is mixed or combined in any given sample if absolute values of fluorine content are to be derived. Fortunatelv. with dilute aaueous solutions such knowledge is often readily available, and experiments 4with such solufions are reported in Part 11. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. REFERENCES Rubin, S., Passell, T. O., and Bailey, L. E., Analyt. Chem., 1957, 29, 736. Sippell. R. F., and Glover, E. D., Nucl. Instrum. Meth , 1960, 9, 37. Siie, P., C. R . Hebd. Sbanc. Acad. Sci., Paris, 1956, 242, 770. Point, J. J., Int. Conf. Radioisotopes Scient. Res., 1957, 2, 180. Hunt, S. E., Pope, R. A., Evans, W. W., and Hancock, D. A., Brit. J . AppL Phys., 1958, 9, 443. “Practical Aspects of Activation Analysis with Charged Particles,” Euratom Report EUR 2957, Pierce, T. B., Analytica CJtim. Acta, 1965, 33, 686. -, Proc. SOG. Analyt. Chem., 1966, 3, 160. Moller, E., and Starfelt, N., Aktiebolaget Atomenergi, Report AE 237, 1966. Pierce, T. B., Peck, P. F., and Cuff, D. R. A., Amlyst, 1967, 92, 143. Fowler, W. A., Lauritsen, C. C., and Lauritsen, T., Rev. Mod. Phys., 1948, 20, 236. Chao, C. Y., Tollestrup, A. V., Fowler, W. A., and Lauritsen, C. C., Phys. Rev., 1950, 79, 108. van Allen, A., and Smith, N. M., Ibid., 1941, 59, 501. Bonner, T. W., and Evans, J. E., Ibid., 1948, 73, 666. Ajzenburg-Selove, F., and Lauritsen, T., Nucl. Phys., 1959, 11, 1. Seagrave, J. D., Brolley, J. E., and Beery, J. G., Rev. Scient. Instrum., 1964, 35, 1290. Marion, J. B., Rev. Mod. Ph-ys., 1966, 38, 660. Whaling, W., Handb. Phys., 1948, 34, 193. Bewers, J. M., and Flack. F. C., Nucl. Instrum. Meth., 1968, 59, 337. Grenoble, 1965. Received June 5th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400001

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, January, 1969, Vol. 94, $9. 7-14 7 Determination of Fluorine by Prompt y- Radiation from Proton Bombardment BY J. M. BEWERS AND F. C. FLACK (Physics Department, University of Exetev, Excter, Devon) The practical application of the nuclear reaction 1@F(p,ay)160 to the determination of fluorine is outlined. Analysis in liquid and gaseous phases suffers from the need to use a proton transmission foil. In principle, solids can be analysed inside the accelerator vacuum, but it is shown that powder targets can be handled more easily by using a foil. Lack of independent standards has prevented absolute calibrations, but it has been demonstrated that fluorine in gases can be determined at relatively high levels (5 to 100 per cent.) by a simple system that could easily be adapted, by the use of thicker foils and higher bombarding energies, to reach much lower levels.Determinations in liquids have been carried out in the range 20 p.p.m. to 10 per cent. of fluorine in aqueous solutions. There is no obvious reason to rule out acidic and alkaline solutions that do not attack the proton window or its holder. Pellet targets were tried for analysis in solids, and it is shown how such targets, made by mixing fluorine salts with iron powder, are usually in- homogeneous. Thus, without independent solid standards the method appears to present difficulties, especially in view of the “thin target” behaviour of the pelleted samples. Bombardment of powders, through a foil window, is shown to be a practical alternative, and it has been demonstrated in a selection of ground rock samples that fluorine is distributed homogeneously throughout a given sample.The importance of the stopping power, and hence a knowledge of the molecular form in which the fluorine occurs, has been emphasised for solids; the situation is far less critical for solutions, and for very weak solutions it has been demonstrated that chemical form is immaterial. Interferences are not important, except for aluminium and possibly lithium. The theoretical limits of detection are in good agreement with those observed in practice, being about 20 p.p.m. for the particular background and experimental arrangement used. PART I of this paper describes the general principles governing the quantitative measurement of fluorine by proton bombardment.In particular, the limit of detection was calculated, and it was shown that a knowledge of the stopping cross-section for the material is of great importance. Presented here are the results of experimental investigations of these points, and also the nature of the relationship between the thick-target yield of y-rays from the reaction 19F(p,ay)160 and the fluorine concentration (w/w) in the target is established. Experiments were carried out with gases, liquids and solids. The experimental tech- nique was altered slightly for determinations in each phase. The technique had to be adapted to suit the particular accelerator that happened to be available and, with gas targets, this restricted, to some extent, the range of measurements that it was possible to make.0 SAC and the authors.8 BEWERS AND FLACK: DETERMINATION OF FLUORINE BY PROMPT [A?i%abSt, VOl. 94 GAS MIXTURE STUDIES MATERIALS- Sulphur hexafluoride was chosen as the gas because of its high chemical stability, even in the presence of heavy ionisation, such as that produced by the passage of a concentrated beam of protons. In choosing a gas with which to dilute the sulphur hexafluoride, considera- tion was given to hertness and parity of density to avoid separation of the components under the fairly severe conditions in the bombardment chamber. Xenon was chosen, as its density is 5.896 g per litre compared with 6.602 g per litre of sulphur hexafluoride, both at S.T.P. TECHNIQUES- Gases were bombarded by allowing protons to traverse a thin foil window and enter a differentially pumped chamber.The arrangement is shown in Fig. 1. The accelerator used was a 500-keV Cockcroft-Walton set that required the use of very thin transmission foils; a thin, non-porous nickel foil manufactured by the Chromium Corporation of America was chosen. It was found that lo4 inch thick foils could be attached to the brass foil holder by using a tin - lead (1 + 1) solder, thus producing an unstrained foil surface in perhaps nine instances out of ten. There is evidence that the use of such a solder and its flux introduces fluorine contamination to the foil surfaces, so that other techniques should preferably be devised if low concentrations of fluorine-containing gases are to be determined by this method. ce Locking 4 ring - Nal crystal Fig.1. Cell for proton bombardment of gases RESULTS- The effects of allowing protons to enter the cell when filled with sulphur hexafluoride at 24-cm pressure of water, and when the cell is evacuated, are shown in Fig. 2. I t is clear that fluorine contamination is present on the front and rear faces of the foil, and on the end w d of the chamber. The foil was about 50 keV thick to protons of 300 to 400 keV. The yield of y-rays from the foil contamination was about 3 per cent. of that from the sulphur hexafluoride. The foil interior seems to have been reasonably free from fluorine. x 300 350 400 450 500 300 350 400 450 500 Proton energy, keV Proton energy, keV Fig. 2. y-Ray yield as a function of proton energy: (a), gas cell containing sulphur hexafluoride at 2Pcm pressure of water; and (b), gas cell evacuated: A, foil front face; B, foil interior; C, foil rear face; and D, cell end wallFig.3. Target arrangement for bombardment of liquids [To face page 9January, 19691 Y-RADIATION FROM PROTON BOMBARDMENT. PART 11. RESULTS 9 It was found that the yield of y-rays from gas in the chamber fell significantly with time. Subsequent examination of the foil windows showed them to have become optically transparent in annular regions near the edges of the windows, through which leakage may have occurred. inch) foils from the same manufacturer were used, although these required operation of the high-voltage generator at its uppermost limit, when its somewhat variable performance caused problems. With the accelerator operating at 490 keV and a foil of the thicker variety, the yield of y-rays was shown to vary linearly with the concentration (w/w) of sulphur hexafluoride in xenon.Currents of protons were limited to 0.3 pA to prevent foil damage, and a small pump was used to ensure good mixing of the gases. The results show the relationship between the number of counts and the percentage concentration (w/w) to be linear from a few per cent. up to 100 per cent. concentration. Fitting by least squares produced a standard deviation of less than 1.2 per cent. For the particular arrangement used, 185.7 counts per per cent. concentration were received for 1OpC of protons. Problems associated with the fluorine contamination of the foil and holder assembly limited the sensitivity of the experiment to about a 2 per cent.fluorine concentration in the gas target. LIQUID TARGETS TECHNIQUES- It would appear from the considerations of stopping power given in Part I, that the proton bombardment method may best be applied to analysis in the liquid phase. The present experiments were designed to test the speed and the sensitivity that could be achieved by using simple methods. As with the gas irradiation, transmission foils are needed to pass the protons from the high vacuum of the accelerator flight-tube into the liquid. A much stronger foil was needed, however, as liquids were handled at atmospheric pressure. Fortunately, van de Graaff accelerators capable of producing protons up to several MeV were available, and resonances in the l9F(pJay)16O reaction up to 1 MeV could be reached, even when energy losses in the foils were as high as 1 MeV.Preliminary experiments suggested that the lowest levels of fluorine contamination occurred in rolled tantalum and in tungsten foils. Accordingly, all stops, collimators and traps in the flight-tube were manufactured from these materials, while 3 x 10" inch thick tantalum foil (Metallwerk Plansee A.G. , Austria) was found suitable for the transmission windows. Reliable mounting of the foils on to the brass mounts was achieved with Loctite 404 adhesive. In many hundreds of hours of use, only four foil failures were experienced; in all instances failure was caused by the foil splitting along the rolling striations during the initial evacuation of the flight-tube when the cell had just been coupled to it.Foil thickness was established by comparing the proton energy required to produce a particular resonance in a fluorine-bearing material bombarded first in the flight-tube and then behind the foil. The foils were found to be 725 keV thick at 1625-keV incident proton energy, and they produced an energy straggle in the transmitted protons of about 2 i per cent. The foils passed beam currents of 0.3 pA continuously when immersed in the target solution. The simplicity of the target arrangement, when advantage was taken of a beam line at 45" to the vertical on a 2-MeV accelerator, is shown in Fig. 3, where the window on the end of the flight-tube dipping into a beaker containing the solution under study is visible. The crystal of the y-ray detector in its lead shield can be seen on the left-hand side.Only the foil holder and foil surface can become contaminated with this arrangement. Fig. 4 illustrates the way in which the contamination is reduced by replacing the solution under test with distilled water; immersion in distilled water overnight removed practically all of the contamination. Other accelerators that had horizontal beam lines were provided with target cells basically similar to that used with gases. Target solutions were changed by pouring out the old solution, flushing with distilled water and pouring in the new target. RESULTS- Liquid samples were prepared by dissolving various weighed amounts of soluble fluorides or fluorosilicates in measured volumes of distilled water at room temperature. Storage of Accordingly, thicker (2.5 x10 .- u - a 0 - f 0 - 5 . Q U Y c 0 ’ - U BEWERS AND FLACK: DETERMINATION OF FLUORINE BY PROMPT [Artd’ySt, VOl. 94 \ I I 1 1 I 1 I 5 10 I5 5 10 15 Time after target removal, minutes Time after target removal, minutes Fig. 4. Decontamination of transmission foil following removal of: (a), 660 p.p.m. ammonium fluorosilicate solution and (b), 7860 p.p.m. sodium fluoride solution the solutions in polythene bottles for periods up to 2 weeks appeared not to affect the inter- relationships between samples in a set. With such samples, results were obtained for the y-ray yields from 100 targets prepared from a variety of solutes with a wide range of concentrations. One crucial experiment was intended to establish that the y-ray yield for a given fluorine concentration was independent of the fluorine salt used to make the target.It was essential to obtain a set of soluble salts that were pure enough to allow accurate calculation of target concentrations; water of crystallisation or adsorbed water were clearly undesirable attributes of the solutes. The suitability of the solutes used for this experiment was established by bombardment of the salts themselves, the results for which have been presented elsewhere.’ Twenty-five targets of various salts were prepared as solutions with concentrations of fluorine in the range 6 to loo0 p.p.m. w/w, the choice of such low concentrations being dictated by the fact that many of the suitable materials have low solubilities. The results for bombard- ment of these samples, which are shown by the graphs of Fig.5 (a and b), fit the same straight line to within 3.2 per cent. standard deviation, although there is a suggestion that the six sodium fluoride samples were consistently lower in concentration, by 2.5 per cent., than the values calculated from the weights. Target concentration, p.p.m. Target concentration, p. p.m. Fig. 5. Variation of y-ray yield as a function of fluorine concentration for aqueous solutions of various pure fluorides: (a), range 0 to 140 p.p.m. and (b), range 0 t o 700 p.p.m. -*-, (NH,), SiF,; 0, K,SiF,; +, NaF; A, CdF,; x, BaF,; 0, PbF,; t, KF; A, LiF; and 0 , KBF, After establishing that the method gave consistent results that were independent of the form of combination of fluorine, single-solute targets were used to investigate the variation of yield with concentration.Statistical errors in counts obtained were less than 2 per cent.January, 19693 7-RADIATION FROM PROTON BOMBARDMENT. PART 11. RESULTS 11 To a first order, the results show the relationship between yield and concentration to be linear over the range 500 p.p.m. to 8 per cent. However, detailed examination of the results reveals that the deviation from the least-squares fitted. curve is markedly improved at the low concentration end by addition of a second-order term in concentration. The term, negative in sign and of magnitude about 10" of the first order, reduces the R.M.S. residuals to about 3 per cent. Almost all of the contribution to these residuals comes from a few Iow-con- centration points. Fig.6. Variation in A, y-ray yield per p.p.m. of fluorine and in B, mean stopping cross-section as a function of fluorine concentration in a target solution of potassium fluoride in water. Bom- barding proton energy = 900 keV Fluorine concentration, p.p.m. The curve in Fig. 6 shows the way in which the y-ray yield per p.p.m. varied with con- centration of potassium fluoride solution according to the best second-order fit. Theory predicts that the reaction yield is inversely proportional to the stopping cross-section, the change in which is also shown in Fig. 6. As the concentration of solute increases, then its atoms play an increasing r81e in stopping. It is clear that serious deviations (5 per cent.) from a linear relationship should not occur until the fluorine concentration exceeds about 10 per cent.Certainly, it can be said that with concentrations below 10oO p.p.m. , no stopping power effects should arise. Thus, a satisfactory explanation of these second-order effects is not yet forthcoming, assuming, of course, that, e.g., there are no systematic non-linear errors in the balance used for weighing the solutes (about 10mg for lo00 p.p.m.). It is regretted that it has not been possible to obtain independent analyses of these samples for comparison. Once calibration (e.g., with potassium fluoride solutions) has been effected, it is possible to carry out analyses in a time that depends on the need to acquire sufficiently good counting statistics and, in turn, on concentration. Such times varied in the present experiment from about 30 minutes in the 10 to 100 p.p.m.range to 5 minutes at concentrations in excess of 104 p.p.m. SOLID TARGETS TECHNIQUES- In addition to the work on pure, powdered fluorine compounds reported elsewhere,f measurements were made with two other kinds of target. Attempts were made to prepare pelleted mixtures of fluorine compounds in inert binders, as other workers2J had used such techniques. FinaUy, a selection of rock samples was irradiated to examine the homogeneity of fluorine distribution in such samples. Pelleted targets could be exposed directly to the proton beam of the 6;00-keV Cockcroft- Walton accelerator inside the vacuum system. We are grateful to Dr. T. B. Pierce and his colleagues for permitting us to use a sample changer that enabled successive exposure of many pellets to be made without breaking the vacuum.Fine iron powder (shown to be12 BEWERS AND FLACK: DETERMINATION OF FLUORINE BY PROMPT [AdYSt, VOl. 94 virtually free from fluorine contamination) of about 200 mesh was admixed with various amounts of calcium fluoride by shaking them in a flask, the resulting mixtures being pressed at 30 tons pressure into thick 20-mm diameter pellets of mass of about 10 g. Tumble mixing was later used, together with the substitution of lead, cadmium and mercury(1) fluorides for calcium fluoride in an attempt to equalise the densities of the mixture constituents, thus improving the efficiency of the mixing process. The powdered rock samples were irradiated by passing protons into them through a 3 x 104-inch tantalum exit foil, with a chamber similar to that used for liquid samples.RESULTS- The first effect to be noticed in connection with the pelleted targets, as might be expected, was a lack of homogeneity in any given sample, although the targets made by the tumbling process were considerably better in this respect. The change to a smaller density range of the pelleted powder, and addition of a grinding stage before tumbling, finally reduced the variations (as measured by the standard deviations of the residuals in a least-squares fit) from an original +19 to +5 per cent. 6 I n 5 .- C $4 Q 23 2 I- - .- U y ’ I 0.7 0.8 0.9 1.0 1.1 Proton energy,MeV Fig. 7. Excitation curves for reaction loF(p, ay)’*O for: A, a fluorspar crystal; and B, a pellet target con- taining 3000 p.p.m.of calcium fluoride in iron powder A brief investigation of target behaviour with a van de Graaff accelerator was made by measuring the yield of y-rays as a function of bombarding proton energy. Fig. 7 shows that a calcium fluoride - iron pellet with 3000 p.p.m. of fluorine exhibits “thin target’’ forms of resonances compared with the yield from thick target resonances in a crystal of fluorspar. No change in behaviour was found when machining a few thousandths of an inch from the face of the pellet to reveal an internal section. These results suggest either that there is some migration of fluorine to the first few microns of the pellet or the range of the protons in some of the fluorine-containing grains is comparable with the thickness of the grains.Dr. T. B. Pierce kindly allowed us to investigate some of the targets, the results for which he has reported,2 and it was found that they showed the same behaviour. It is thus clear that analysis by this method in the solid phase is closely bound up with problems of homogeneity. Comparison with standard samples must ensure that the effective target thicknesses of sample and unknown are equal unless resonances are measured in each sample. The stopping cross-section per fluorine atom for the material in which the fluorine is bound must also be known, as explained in Part I.January, 19691 7-RADIATION FROM PROTON BOMBARDMENT. PART 11. RESULTS 13 The final results to be reported here were obtained with a series of ground samples, for which fluorine assays had been made independently by Khan.' A sample of W.l standard rock powder was also kindly supplied by Mr.A. R. Coote of the Bedford Institute of Oceano- graphy, Dartmouth, Nova Scotia. Following the remarks made earlier in this section concerning the need for some qualitative knowledge of the form in which the fluorine is contained in the samples, the ptincipal object of the investigation was to determine the homogeneity of the geological samples and to compare their yields of y-rays with a set of lead fluoride - iron mixtures made by tumble r+xing. The powders were poured into the target chamber, bombarded via the tantalum wmdow and then the chamber was emptied, cleaned and re-filled with a further portion of the same sample.In this way the homogeneity of the samples was tested. Table I shows the results that indicate the extremely homogeneous nature of the fluorine content of these rocks, despite the relatively coarse particle size (about 120 mesh), while a typical lead fluoride - iron mixture exhibited the aforementioned inhomogeneities expected in this type of preparation. The third column of Table I gives the mean counts for each sample. As the molecular form in which the fluorine is bound has not been identified, it is not possible to make a reliable quantitative estimate of fluorine-containing molecules. The stopping cross-sections per fluorine atom, at 870-keV proton energy, were calculated for several common minerals known to be sources of fluorine. The values range from 12-54 x lO-lS eV cm2 per fluorine atom for fluorspar through 32.6 x 10-l6 eV cm2 for topaz (AlF2Si0,) to a value in the low 70's for the common micas.With the arbitrary assumption that topaz has a stopping cross-section typical of the fluorine-bearing constituents of the rock samples, estimated values for fluorine concentration were arrived at by comparison with the lead fluoride - iron standard mixtures. This assump tion was chosen as it also corresponds approximately to the expected stopping power for a 1 + 1 mixture of the most commonly occurring forms of fluorine, Gz., fluorspar and the micas. Column 4 of Table I gives the results of these comparisons, while column 6 presents determinations of fluorine content arrived at by independent analyses (see foot of table).TABLE I RESULTS FROM BOMBARDMENT OF A SET OF GRANULAR ROCK SAMPLES Rock \ Estimated assessment sample Means from ten Over-all mean fluorine content, of fluorine content, number successive runs ininus background p.p.m. p.p.m. w. 1 1170, 1184 966 285 290* 81 1 874, 813 650 194 9t y-Counts per 10 pC Independent r A 1160, 1134 847, 911 851, 772 823, 931 671a 38,107, 38,772 38,254 13,953 11,300 147a 38,492, 38,246 38,181 13,740 - 039a 66,361, 55,349 55,667 16,683 16,600t 181 39,663, 39,787 39,552 12,071 12,200 E. 1 33,143 33,013 9834 13,750t WA3/DK 2310, 2340 2146 640 600t WA3/dK 2292 2074 618 600t 36b 51,228, 60,245 50,510 14,903 17,400 38,488 443n 42,134 42,011 12,614 { :;:::: * After Fleisher.K t After Khan.' Other values for fluorine quoted above are approximations derived from analyses of similar rock types.The agreement between the estimated values of concentration and the independent assessment is fairly good for most samples, with the exception of Aplite 811. The mica and topaz-bearing rocks 039a, 671a, 443a, 181 and 36b might be expected to give comparable results, but it is surprising that the values for the clays and shales (WA3/DK and WA3/dK) and W.l agree so well, as the mode of occurrence of fluorine is an unknown factor.14 BEWERS AND FLACK INTERFERENCES AND LIMIT OF DETECTION- Any serious competing reactions must give y-rays of energy about 6 MeV, or higher. The most likely interfering nuclei to undergo simple proton capture, followed by y-emission are, therefore, lithium-7, sodium-23, aluminium-27 and chlorine-36.Strong solutions of sodium, potassium and caesium chlorides were bombarded, and in no instance was the observed number of y-rays, in the appropriate part of the spectrum, found to exceed three times that from the bombardment of pure water. Lithium produces 17-6 and 14MeV y-rays when bombarded with protons, and some of these will interact with the crystal detector to generate pulses in the 6-MeV region of the spectrum. Tests on solid targets of lithium chloride showed a y-ray count in the single- channel analyser equivalent to several hundred p.p.m. of fluorine, although how much of this was caused by fluorine contaminant in the chloride it is not possible to say. Finally, bom- bardment of aluminium fiuoride - aluminium pellets generated appreciable numbers of y-rays, which interfered sdficiently to distort the fluorine spectrum.The conclusion reached from these observations is that large amounts of aluminium should be avoided. Lithium could also present difficulties unless a complete spectrum was observed on each occasion, so as to make a determination of the y-ray energies involved. All normal concentrations of other light elements appear to be tolerable. In principle, spectrum-stripping should remove all interfering lines. A calculation was presented in Part I, which suggested a limit under specified conditions to detection of fluorine of about 27 p.p.m. at 1-MeV incident proton energy. Examination of the experimental results for the liquid targets suggests that theory and experiment are in good agreement. It did not seem possible to reduce the background radiation below about 200 counts per minute when the beam was on target because of radiation from reactions in the flight-tube components, many of which bore minute traces of fluorine. Conceivably, a brand new purpose-built system could be kept cleaner than one fitted to an accelerator used by many groups for diverse purposes. Further improvements in transmission foils to allow larger currents would also reduce background problems and shorten analysis times. The authors are indebted to the Analytical Sciences and the Nuclear Physics Divisions of the Atomic Energy Research Establishment, Harwell, for making available accelerator and other facilities. Thanks are given to the Science Research Council for a maintenance grant held by one of us (J. M. B.) during this work. REFERENCES 1. 2. 3. 4. 5. NOTE-Part I of this series appears on p. 1. Bewers, J. M., and Flack, F. C., NucZ. Instrum. Meth., 1968, 59, 337. Pierce, T. B., Peck, P. F., and Cuff, D. R. A., AnaZyst, 1967, 92, 143. Sippell, R. F., and Glover, E. D., Nucl. Instrum. Meth., 1960, 9, 37. Khan, S. A., “Analysis of B, F, Li in Silicate Rocks and Minerals,” Ph.D. Thesis, University of Fleisher, M., Geochim. Cosmochim. Acta, 1965, 29, 1263. Exeter, 1966. Received June 5th. 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400007

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, January, 1969, Vol. 94, $9. 15-19 15 A Simplified Procedure for the Assay of Picocurie Concentrations of Radium0226 and its Application to a Study of the Natural Radioactivity in Surface Waters *in Scotland BY B. CONLAN, P. HENDERSON AND A. WALTON (Chemistry Department, The Univevsity, Glasgow W.2) A complete and simplified emanation system for the measurement of picocurie concentrations of radium-226 is described. A scintillation detector, fabricated from Perspex, and standard commercial electronics are used. Efficiencies of 55 per cent. were obtained with backgrounds of about 1 count per minute for several detectors. Studies have been conducted of the natural radioactivity in Loch Lomond and surrounding streams. The Loch surface water has an average radium-226 concentration of 0.04 pCi 1-1 and the uranium concentration, as determined by delayed neutron technique, is 0.13 pg 1-1.These results suggest that the amounts of radium-226 in the oceans are independent of the influx from streams, and that drinking water in this region of Scotland does not constitute a significant intake of radium-226 by humans. THE measurement of radium-226 concentrations in the picocurie range is important in considerations of environmental radioactivity. Early techniques usually involved the co- precipitation of radium with barium sulphate and subsequent solid-source a-counting. Self-absorption effects create obvious difficulties with this approach. Subsequent develop- ments led to the more satisfactory technique of counting the radon-222 daughter product via gas counting in an ionisation chamberlls or some scintillation device.3~~ Secular equilibrium between radium-226 and radon-222 is virtually achieved in 30 days, a time period that is not unreasonable in most experiments, and the method has the added virtue that duplicate measurements can be performed without further chemical handling of the radium solution.The emanation technique used in the system described is simpler than those previously developed] and the apparatus can be constructed from readily available materials and equip- ment. A scintillation detector, similar to that described by Di Ferrante, Gourski and Boulenger,6 has been used with nitrogen as the carrier gas for the introduction of radon into the detector. Pyrex, 250-ml volume, equilibration flasks, complete with an inlet bubbler and outlet, were used for secular equilibrium build-up.The need to use a charcoal traps*' for isolation of radon-222 prior to radioassay was found to be unnecessary, and although the recovery of radon was somewhat reduced, this had no serious effect on the reproducibility of results. The system has been used for the routine assay of radium-226 concentrations in inland waters at concentrations as low as 0.01 pCi 1-1 following enrichment by evaporation. EXPERIMENTAL EMANATION SYSTEM- The experimental system for the determination of radium-226 via the emanation of radon-222 is shown in Fig. 1. Solutions for analysis were flushed out with nitrogen gas and left to equilibrate for a measured period, usually between 4 and 16 days.The equilibration flask was then attached to the vacuum system and the entire line evacuated. Radon rkcovery was effected by bubbling nitrogen through the solution and collecting the radon in two traps cooled in liquid nitrogen. Water was removed in the first trap by surrounding it with an abetone - solid carbon dioxide mixture. 0 SAC and the authors.16 CONLAN, HENDERSON AND WALTON : A SIMPLIFIED PROCEDURE [AutabSt, VOl. 94 The flow-rate of nitrogen gas was 40ml per minute and was controlled by a needle valve attached directly to the cylinder. As the nitrogen passed through the solution and traps it was allowed to collect in the l-litre evacuated bulb shown in Fig. 1. A mercury blow-off served as an additional visual indicator of the flow-rate of nitrogen.The flushing of the solution was complete when the mercury level reached zero. Q Lltre To -To Hg manometer Fig. 1. Emanation system for radium-226 assay Excluding the l-litre bulb the system was then evacuated and the radon present in the liquid nitrogen traps transferred to the final U-tube preceding the detector. Liquid nitrogen in the second and third traps was replaced by acetone - solid carbon dioxide cooling mixtures to retain any water. The final step in the process was to transfer the radon to the detector. This was accom- plished by the removal of the coolant surrounding the U-tube and allowing the radon to be flushed rapidly into the detector with the nitrogen gas contained in the bulb. A final pressure of about 65 cm of nitrogen (Phs radon) in the detector was attained with this simple technique , which eliminates the need for pumping devices.DETECTOR AND ELECTRONICS- The detector is essentially similar to the design of Di Ferrante, Gourski and Boulenger5 and consists of a cylinder of Perspex with a spherical cavity of 48ml. A thin coating of silicone grease followed by a layer of zinc sulphide phosphor, obtained from Nuclear Enter- prises (G.B.) Ltd., are applied to the inside surfaces of each hemisphere. The two sections, carefully machined to overlap and slip together, are held together by Scotch tape. (In the original design the sections were bonded by Araldite, which did not permit disassembly and re-coating.) Black insulating tape is then wrapped around the assembled detector, and a microstopcock attached to the outlet tube for filling purposes.The detector is placed directly on the head of the 5-cm photomultiplier, coupling being accomplished with a thin layer of silicone grease. A light-tight enclosure is threaded to the photomultiplier housing and completely surrounds the detector assembly. In this work three basic units were used in the electronics. These consisted of scintillation detector, Ekco type N691A; amplifier, Ekco type N640A; and scaler unit, Ekco type N529D. The amplifier was operated at minimum gain, i.e., x25. A11 three units were standard commercial instruments used in teaching laboratories and were not especially selected for this study. OPERATION AND CHARACTERISTICS OF THE COUNTING SYSTEM- Under operating conditions a working voltage of 750 volts was chosen at the mid-point of the plateau, about 300 volts in length.An over-all efficiency for the entire system of 54.2 & 2.8 per cent. (based on the emission of three a-particles for each radon disintegration) was determined from the assay of a 10-log radium standard in dilute acid solution. This sample was prepared from a g standard radium solution, obtained from the Radiochem- ical Centre, Amersham. Calibration was performed on ten occasions with two detectors andJanuary, 19691 FOR THE ASSAY OF PICOCURIE CONCENTRATIONS OF RADIUM-226 17 the agreement between the detectors was within the error of the standard results. Back- ground counting rates (about 1 count per minute) were rather high when compared with more sophisticated instruments but were adequate for the purpose of our experiments, and memory effects were not apparent when the maximum count-rates of about 300 counts per minute were used.J I F F 0 20 40 60 80 100 I20 140 I Time, minutes 0 Fig. 2. Decay curve of radon-222 daughter products It is interesting to note (see Fig. 2) the activity-decay curve for the post-counting period, i e . , following the removal of a radon sample from the counter and re-filling of the detector with inactive nitrogen gas. An approximate half-life for this decay curve is 35 minutes, which is consistent with the expected decay of particulate radon-222 daughter products (see below) that remain on the walls of the detector. Further, calculations showed that these decay products and radon-222 are detected with about the same efficiency in the scintillation detector. As these decay products will build up after the radon-222 is introduced into the detector it is, therefore, necessary to allow 2 to 3 hours to elapse before counting.268 mlnutes 1.32 minutes MEASUREMENT OF RADIUM-226 IN NATURAL WATER- Little effort, apart from survey studies, has been expended in this country on the assay of natural radioactivity in the environment. A systematic investigation of the Scottish lochs has commenced, the results of which will be useful in evaluations of the natural radiation environment and in geochemical balance studies. Although some total a-activity results were reported for drinking-water supplies in the United Kingdom in 1964 and 1966,s37 only a single sample from the Belfast region was investi- gated for its radium-224 and radium-226 contents.This paper gives the results of radium-226 and uranium measurements for Loch Lomond water samples, together with the concentrations of sodium, potassium, magnesium and calcium.18 CONLAN, HENDERSON AND WALTON: A SIMPLIFIED PROCEDURE [AfldySt, VOl. 94 EXPERIMENTAL PROCEDURES AND RESULTS- The experimental procedures used were standard techniques and only a brief summary is given. Ten-litre water samples were collected in polythene bottles from Loch Lomond and inflowing streams in the spring and summer of 1966. Ordnance Survey reference points for the samples are given, together with radium-226 results, in Table I. Raaum-226 determinations were carried out by using an emanation technique and a modified Perspex scintillation detector coupled to a photomultiplier tube, as above, and the uranium measurements were made by the delayed-neutron method.8 The results for sodium, potassium and magnesium were obtained by flame-emission and atomic-absorption spectro- scopy on the Unicam SP90 instrument, and those for calcium were performed in co-operation with the Clyde River Purification Board by using the EDTA titrimetric method.Samples for radium assay were first filtered through a glass-fibre filter (GF/A) before being reduced in volume to about 200 ml by evaporation. Separate samples of 2-litre volume were used for uranium assay, and these were also evaporated to about a few millilitres, before being transferred to a polythene vial in which evaporation was completed at a slower rate.Solutions of analytical-reagent grade salts were used as standards in the spectroscopic assays. Sample G-1 G-2 G-3 G-4 G-5 6 6 G-7 G-8 G-9 G-10 TABLE I RADIUM-226 AND ELEMENTAL CONCENTRATIONS FOUND I N LOCH LOMOND WATER I N 1966 Ordnance Radium-226, Pc survey reference pCi 1-1 NS-358897 0.04 NS-416907 0.06 NS-396942 0.03 NS-410897 0.02 NS-400890 0.0 1 NS404896 0.03 NS-474875 0.12 NS-384917 0.04 NS-370902 0.01 NS-426896 0.01 3tassium. p.p.m. 0.70 1.00 0.75 0.70 0.85 1.65 0.70 - I Sodium, p.p.m. 3.7 7.5 3.5 3.5 6-4 7.5 3.7 - Magnesium, Calcium, p.p.m. p.p.m. 1.2 5.2 1.2 1-6 1.2 5-2 1.2 5-2 1-2 5.2 5 4 22.4 1.2 5-6 - - RESULTS AND DISCUSSION The average concentration of radium-226 in Loch Lomond surface water is 0.04 pCi 1-1, with a range extending from 0.01 to 0-12 pCi 1-1 (Table I).Results obtained by other workers and noted by Miyake, Sugirnura and Tsubota9 indicate an average radium concentration in river waters throughout the continents to be about 0.07 to 0008pCi 1-l, which is in agree- ment with our results. River waters in Japan appear to range in concentrations from 0*04 to 0.13 pCi 1-l. A more recent publication by Moore1* suggests radium-226 concentrations in the Amazon and Mississippi rivers to be 0.01 and 0-03 pCi I-l, respectively; again in good agreement. Elemental analyses for sodium, potassium, calcium and magnesium indicate a remarkable uniformity over the southern regions of Loch Lomond. Higher concentrations observed in G-6 can be readily explained by the fact that the sample was collected from the Endrick Water at Drymen Bridge.In the catchment area of this river are Lower Carboniferous Limestone outcrops, together with considerable agriculture. Uranium results for some water and local rock samples are shown in Table 11. TABLE I1 URANIUM CONCENTRATIONS IN SELECTED SAMPLES Sample Type Concentration G-20 Blank (distilled water) < 0.005 pg per litre G-2 1 Tap water (Loch Katrine) 0.06 p g per litre (3-22 Loch Lomond water (G-1) 0.13 pg per litre G-23 Old Red Sandstone (Ardmoor) 1.02 p.p.m. G-25 Phyllite (Dunoon) 1.30 p.p.m.January, 19691 FOR THE ASSAY OF PICOCURIE CONCENTRATIONS OF RADIUM-226 19 Analytical errors in the above results are estimated to be about +lo per cent. Once again these results are within the ranges of activity found by other workers for similar materials.It is interesting to note the uranium concentrations for Loch Lomond water and tap water which, in the Glasgow region, is derived from Loch Katrine, and is sufficiently pure to warrant little further treatment. Loch Lomond is in the immediate vicinity of Loch Katrine and within a similar geological environment. Comparable results would, therefore, have been expected from the two sources. It is tempting to carry further the similarity in uranium concentrations to radium-226 results, in which case drinking water would be expected to have a radium-226 concentration of about 0.02 pCi 1-l. Circumstantial supporting data for this extrapolation are forthcoming as follows. According to recent report^,^,' the total a-activity of Glasgow (Loch Katrine) drinking water varies between 0.23 and 0.59 pCil-l (mean 0.41 pCi 1-I).The ratio between total a-activity and radium-226 concentrations found in Belfast drinking water was 8.7. If the same ratio is applicable to the Glasgow region, then the expected radium-226 concentration would be about 0.05 pCi l-l, and is in reasonable agreement with the suggested figure of 0-02 pCi1-l. Such a value suggests that drinking water in this area does not constitute a major source of radium-226 to humans. It is significant that, if the above relationships are used to evaluate concentrations of radium-226 in several rivers and springs throughout the United Kingdom, an average run-off concentration of 0-05 pCi 1-1 of radium-226 can be calculated.6 (Wells and boreholes, on the other hand, appear to contain somewhat higher concentrations, with a greater degree of variability, 0.02 to 1.2 pCi 1-l.) Miyake, Sugimura and Tsubota9 indicate a run-off concen- tration of 0.07 to 0-08 pCi 1-1 for Japanese rivers. As discussed by Moorelo these results present further supporting evidence for the premise that the major source of radium in the oceans is something other than influx from rivers.On the question of uranium geochemistry our results of 0.05 and 0.13 pg 1-1 are somewhat higher than the results determined for the Amazon but lower than those observed in Missis- sippi in 1965. It is significant in this discussion, however, to note that in Japan the weighted mean concentration of uranium in rivers is 0.57 pg 1-1, a factor of at least five greater than in Scotland and about thirteeen times the concentration in the Amazon river. It is clear that isolated results such as those presented by Miyake, Sugimura and Tsubota, and Moore, and indeed by ourselves, must be treated with caution. Extrapolation to world-wide generalities and ocean chemistry can hardly be justified in view of the un- certainties that can arise from numerous sources, which might include methodology in sampling and analysis, environmental geology, meteorology and regional effects of agri- cultural processes.We thank Dr. N. H. Gale of Oxford University for carrying out the uranium measure- ments. Moore suggests that his results may be high. 1. 2. 3. 4. 6. 6. 7 . 8. 9. 10. REFERENCES Bate, G. L., Volchok, H. L., and Kulp, J. L., Rev. Scient. Instrum., 1964, 25, 163. Walton, A., Kologrivov, R., and Kulp, J. L., Hlth Phys., 1959, 1, 409. Lucas, H. F., Rev. Scient. Instrum., 1967, 28, 680. -, in Adams, J. A. S., and Lowder, W. M., EdiWs, “The Natural Radiation Environment,” Di Ferrante, E. R., Gourski, E., and Boulenger, R., in Adams, J. A. S., and Lowder, W. M., Editors, “Radioactivity in Drinking Water in the United Kingdom,” H.M. Stationery Office, London, 1964. “Radioactivity in Drinking Water in the United Kingdom,” H.M. Stationery Office, London, 1966. Gale, N. H., “Radioactive Dating and Methods of Low Level Counting,” Publication International Miyake, Y., Sugimura, Y., and Tsubota, H., in Adams, J. A. S., and Lowder, W. M., Editors, Moore, W. S., Earth Planet. Sci. Letters, 1967, 2, 231. Received July 24th, 1968 University of Chicago, 1964, p. 316. op. cit., 1964, p. 353. Atomic Energy Agency, Vienna, 1967, pp. 431-462. op. cit., 1964, p. 219.

ISSN:0003-2654    

DOI:10.1039/AN9699400015

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

20 Analyst, January, 1969, Vol. 94, j@. 20-25 Some Observations on Potassium Dichromate as an Oxidimetric Standard BY R. BELCHER, C. L. CHAKRABARTI AND W. I. STEPHEN (Department of Chemistry, The Ufiiversity, P.O. Box 363, Birmingham 16) The stoicheiometry of the reaction of iron(I1) with chromium(V1) and the suitability of potassium dichromate as a primary standard have been examined. To differentiate the effects of possible non-stoicheiometry and impurities in the samples, several different sources of chromium(V1) and three different standardisation schemes, based on the use of arsenic trioxide, sodium oxalate and potassium iodate, have been used. It is shown that the stoicheiometry of the reactions is satisfactory and that suitable grades of potassium dichromate are excellent primary standards.SOLUTIONS of potassium dichromate have been used widely as oxidimetric titrants for more than 100 years, but there are several aspects of the behaviour of this well known oxidant that have yet to be explained. Although potassium dichromate can, in many of its applica- tions as an oxidimetric titrant, be advantageously replaced by cerium(1V) sulphate, the changeover is necessarily slow and potassium dichromate will probably continue to be used for many years. Most analytical chemistry textbooks consider the detennination of iron(I1) as the main use of dichromate solutions in oxidimetric titrimetry, but this oxidation is probably more readily carried out with a cerium(1V) solution as titrant, and dichromate need not be retained for this purpose alone.However, even if potassium dichromate is eventually abandoned as a titrant, the dichromate - iron(I1) reaction will continue to be of analytical importance in the determination of chromium as chromium(V1) after oxidation d t h , for example, concentrated perchloric acid. The apparently anomalous behaviour of certain redox indicators in the iron(I1) - di- chromate reaction has been explained during the last few y e a r ~ . ~ J , ~ However, certain problems remain that have only been partly answered; several investigators have indicated that the iron(I1) - dichromate reaction is not stoicheiometric and the purity of potassium dichromate has been q~estioned.4~~~~ The iron(I1) - dichromate reaction follows a different course according to the direction of the titration and the equivalence-point potential is also different; when dichromate is added to iron(II), the potential rise in the immediate neighbourhood of the equivalence point is 0.85 to 1-00 volt in M sulphuric acid, whereas in the reverse titration the corresponding fall in potential is from 1-20 to 0-85 volt.' This partly explains why ferroin sulphate, with a formal reduction potential of 1.06 volts in M sulphuric acid, functions as an indicator only in the latter titration.At acid concentrations greater than M, ferroin sulphate functions well when dichromate is added to iron(I1) solutions, because the equivalence potential is now raised and the indicator potential is lowered. This curious behaviour of potassium dichromate, studied first by Forbes and Bartlett in 1913, still remains largely unexplained, although several suggestions have been offered to account for the anomalies observed in potentiometric studies of the &chromate-iron reaction.Laitinens suggests that surface oxidation of the platinum indicating electrode causes the variation in observed potentials, and it seems quite likely that a surface effect does occur, which influences the potentiometric results. This, however, cannot be the only contributive factor to the anomalous potentials, because ferroin sulphate indicates the end- point of the dichromate - iron(I1) titration, but not that of the iron(I1) - dichromate titration. 0 SAC and the authors.BELCHER, CHAKRABARTI AND STEPHEN 21 This shows that even in the absence of the indicating electrode, the equivalence point occurs at an appreciably higher potential than that of the reverse titration. Kolthoff and F ~ r m a n , ~ WestheimerlO and Laitinen8 have reviewed the problem in considerable detail but offer no conclusive answers.Some preliminary work carried out on the problem in this Department11 has shown the need for an intensive investigation. The peculiarity in the behaviour of dichromate solutions as oxidants does not affect the quantitative nature of the reaction with iron(I1) , Laitinen having reported the successful determination of as little as 1 pg of chromium in 100 ml of solution by an amperometric titration. The availability of potassium dichromate in a state of high purity and the extra- ordinary stability of its aqueous solutions are practical factors that strongly favour the retention of dichromate as an oxidimetric reagent.Nevertheless, the doubts cast both on the stoicheiometry of the iron(I1) - dichromate reaction and the purity of potassium dichromate itself have prompted the present investi- gation,l2 which reports an examination of these associated problems. When doubts exist about both the purity of materials and the stoicheiometry of their reaction, differentiation of the two effects may be difficult, particularly when, as in the present instance, no dependable primary standards are available for direct standardisation procedures to be applied. However, it was considered that the use of different samples of chromium(V1) and of several different routes for the standardisation should allow the various factors to be elucidated.Accordingly, six varieties of chromium(V1) were acquired or prepared: “Pure for Volu- metric Standardisation” potassium dichromate of unstated purity, AnaIaR potassium dichromate, thrice crystallised AnalaR potassium dichromate, potassium dichromate pre- pared from the AnalaR chromate salt, potassium dichromate re-prepared from AnalaR potassium dichromate and AnalaR potassium chromate. Three different schemes of standardisation were chosen for the study of these materials- A. B. C. As,O, -+ Ce(IV) --+ Fe(=) --+ Cr(VI). Oxalate + Ce(IV) + Fe‘II) + Cr(V1). Iodate -+ thiosulphate --f CYV1). Methods that involve the use of potassium permanganate were avoided so that there would be none of the uncertainties caused by the intermediate reactions inherent with this reagent.The schemes were chosen so that three of the most widely accepted and unquestioned primary standards for oxidation - reduction reactions could be used. To ensure maximal precision in the results, all titrations were carried out with a weight burette.” According to Incze,l4 this titration technique used at constant temperature gives an extremely high precision, which is, of course, much better than that of any possible technique of end-point detection. CHOICE OF METHODS OF END-POINT DETECTION- Both Eppley and Vosburgh4 and Willard and Gibsons reported that positive errors of 0.2 to 0.5 per cent. are obtained when iron(I1) is titrated with dichromate solutions; these workers used potentiometric detection of end-points.Fumes9 reported similar results when visual indicators were used. However, in the present work, visual indicators were preferred as the end-points obtainable with the best oxidation - reduction indicators are visible to fractions of a drop of 0.1 N titrants; moreover, visual titrations can be done more quickly than potentiometric titrations, so that any possibility of atmospheric oxidation of iron(I1) is reduced, and, finally, any effect of the electrode potential8 is eliminated. For the titration of iron(I1) with dichrornate solution, the best visual indicator is un- doubtedly the disodium salt of diphenylbenzidinesulphonic acid, which retains the virtues of the better known diphenylamine compounds while avoiding any problems associated with the formation of intermediate products.ls For the titration of iron(1I) with cerium(1V) solution, ferroin was used as a saturated solution of its perchlorate salt.This is about 0-001 M at room temperature and enables the theoretical indicator blank to be easily calculated. Indicator blanks were determined with great care to ensure accuracy in the final results, and dilute indicator solutions were used so that exactly the same amount could be added for each titration.22 REAGENTS- Potassium dichromate-Samples of P.V.S. grade and AnalaR grade (99-9 per cent.) were obtained from Hopkin & Williams Ltd. A further sample of AnalaR grade was thrice re- crystallised from water. AnalaR grade potassium chromate (99.9 per cent., Hopkin & Williams Ltd.) was used in some tests, and a sample of dichromate was prepared from the chromate by acidification of a concentrated solution with hydrochloric acid and recrystal- lisation from water.This procedure, rather than the more direct neutralisation with chromic acid, was used because of the difficulty in obtaining pure chromic acid and the risk of pro- ducing a non-stoicheiometric material containing an excess of chromic acid. A further sample of potassium dichromate was prepared by consecutively making a concentrated solution alkaline, acidifying and recrystallising from water. The last two samples gave negative tests for chloride. All the samples were pulverised and dried at 140" to 150" C to constant weight before use. Primary standards wed-Arsenic trioxide and sodium oxalate were obtained from the U.S.National Bureau of Standards. The stated purities were 10040 per cent. (Batch 83b) and 99-95 per cent. (Batch 40g), respectively. Both of these samples were dried at 105" C to constant weight before use. Potassium iodate (P.V.S., Hopkin 81 Williams Ltd.) was dried at 120" C to constant weight. Cerium(1V) solutim-Prepare a 0-01 M cerium(1V) sulphate solution in M sulphuric acid from cerium(1V) hydroxide as described by Smith and Diehl.ls Irm(I1) solutim-Prepare a 0.1 M ammonium iron(I1) sulphate solution in M sulphuric acid from analytical-reagent grade material. Sodium thiosu&hate soZutio-F'repare an aqueous 0-1 N solution and add a little sodium carbonate in the usual way. All solutions were prepared with freshly boiled, glass-distilled water and all the reagents used were of analytical-reagent grade.INDICATORS- Diphenylbenzidinesulphonic acid (disodium salt) was used as an aqueous 0405 M solution; 0-50 ml was added for each titration. A saturated aqueous solution of ferroin perchlorate (040105 M at 25" C) was used in other titrations; 1.00 ml was added for each titration. The indicator blank for ferroin, determined by direct titration with 0401 M cerium(1V) solution, was equivalent to 0.01 ml of 0.1 M cerium(1v) per 1 nd of indicator solution, which is essentially the theoretical value. As the indicator action of diphenylbenzidine depends on an induced reaction, direct titration is unsatisfactory for establishing the blank value, which was, therefore, established by titrating 1040 and 2040-ml aliquots of iron(I1) solution with 0.01 N dichromate solution and then calculating.A blank of 0.01 ml of 0.1 N dichromate was found. BELCHER, CHAKRABARTI AND STEPHEN: SOME OBSERVATIONS ON [Autalyst, Vol. 94 EXPERIMENTAL APPARATUS- N.P.L. Grade A graduated glassware was used for all volume measurements involved in this work. All solutions of potassium dichromate were prepared in a bath maintained at 20" C by a thermostat. All titrations were carried out with weight burettes as described by LaMer and Fried- man.13 To minimise drop errors, the bulk of the titrant was added from the weight burette, and the titrations were then completed by the addition of a 10-fold diluted solution from a 10-ml Grade A burette. TITRATION PROCEDURES- Arsmic(l1I) with cerium(1'v)-Prepare a 0.1 N arsenic trioxide solution by gently heating 4 - W g of the trioxide with log of a,tihydrous sodium carbonate and 5Oml of water.When dissolution is complete, neutralise the solution by gradual addition of sulphuric acid (1 + 5), cool, transfer to a 1-litre graduated flask and dilute to volume with water in a bath maintained at constant heat by a thermostat.January, 19691 POTASSIUM DICHROMATE AS AN OXIDIMETRIC STANDARD 23 To 30 ml of this solution, add 40 ml of M sulphuric acid, 3 drops of 0.01 M osmium tetroxide in 0-1 M sulphuric acid and ferroin solution. Titrate with 0-1 M cerium(1V) solution. Cerium(1V) with Zron(11)-To 30 ml of 0.1 M iron(I1) solution, add ferroin solution and titrate with 0.1 M cerium(1V) solution. Irort(I1) with chromium(V1)-To 30 ml of 0.1 M iron(I1) solution add 100 ml of M sulphuric acid, 5 ml of syrupy phosphoric acid and diphenylbenzidine solution.Titrate with 0-1 N dichromate solution. Over-titration can be avoided by first titrating to a dark green shade, then adding the phosphoric acid and continuing the titration to the green-to-violet end-point ; for optimal precision, the phosphoric acid must always be added at the same stage of the titration. Oxalate with cerium(1V)-Weigh accurately about 0.3 g of sample, dissolve it in 100 ml of M sulphuric acid and add 50-00 ml of 0.1 M cerium(1V) solution. Heat to 40" to 50" C for 5 minutes, then allow to cool to room temperature, add ferroin solution and an excess of iron(I1) solution and titrate with 0.1 M cerium(1V) solution.Titrate an amount of iron(I1) solution equal to that used in the previous titration to establish the equivalence. Iodate with thiosuZ#hate-Prepare an aqueous 0-1 N solution of iodate, with 36668g of potassium iodate per litre. To 30 ml of this solution, add 2 g of potassium iodide and 10 ml of M hydrochloric acid, and titrate with 0-1 N thiosulphate solution, adding Thyodene as indicator near the end-point. It is important to add the iodide before the acid, in order to avoid possible iodine losses. Chromium(V1) with thioszll'hate-To a mixture of 2 g of potassium iodide, 20 ml of M hydrochloric acid and 50 ml of water add 30 ml of 0-1 N chromium(V1) solution, stopper the flask, mix well and leave in the dark for 10 minutes. Titrate with 0-1 N thiosulphate solution, adding indicator as above.RESULTS AND DISCUSSION STANDARDISATION BASED ON ARSENIC TRIOXIDE- The stoicheiometry of the titration of iron(I1) with cerium(1V) solution1' is well estab- lished, as is the precision of the titration in M sulphuric acid when ferroin is used as indicator.l8Js The stoicheiometry of the arsenic(II1) - cerium(1V) reaction was checked over the range 0 to 65 ml of 0.1 N solution; the same equivalence of the reactants was obtained over the whole range within an experimental error of 0.03 per cent. The method of Smith and Diehl,l6 based on the use of osmium tetroxide as catalystJm was used. Similarly, the equivalence of the iron(I1) - chromium(V1) reaction in M sulphuric acid in the presence of phosphoric acid was checked over the range 0 to 65 ml of 0-1 M iron(I1) solution; again the same equivalence was obtained over the entire range within an experimental error of 0-035 per cent.STANDARDISATION BASED ON SODIUM OXALATE- The method of Willard and Young1' was used for the titration of oxalate with cerium(1V). To maintain constancy in the technique of end-point detection, an excess of iron(I1) solution was added after the oxidation of oxalate was complete, and this excess was then titrated with cerium(1V) solution. The remainder of the reaction chain was the same as that used with arsenic (111). STANDARDISATION BASED ON POTASSIUM IODATE- Although no problems seem to be associated with the iodate - iodine - thiosulphate system, there is a great deal of contradiction in the literature on the reaction of dichromate with iodide." The acidity, the excess of iodide and the time of standing for this reaction are important if precise stoicheiometric results are to be achieved.In the present work, the method of Vosburgh,B who carefully established the optimal conditions for precise titrations, was used; the medium was 0-2 M in hydrochloric acid, and 2 g of potassium iodide (100 ml) were added. In confirmation of Vosburgh's results, it was found that lower concentrations of acid and higher concentrations of iodide led to erratic results.24 BELCHER, CHAKRABARTI AND STEPHEN: SOME OBSERVATIONS ON [AIza&St, VOl. 94 COMPARISON .OF RESULTS OBTAINED- The results obtained for five different samples containing chromium(V1) by the three standardisation schemes are shown in Table I.In general, the agreement between the results for the same potassium dichromate sample is excellent. The precision of scheme A is dis- tinctly superior to the others and it is probably the most reliable for standardisation purposes; the precision of the results reflects the straightforward nature of all the reactions involved in the scheme. It is clear that any of the first three samples can be used as a primary standard; the average recoveries for these samples by each of the three schemes are 100-01 (scheme A), 99.36 (scheme B) and 100*00 per cent, (scheme C ) . It is doubtful if real signifi- cance can be attached to the slightly low result obtained by scheme B. TABLE I THE STANDARDISATION OF CHROMIUM (VI) SAMPLES BY VARIOUS METHODS Sample of chromium(V1) KaCr,O,, AnalaR grade' .. &Cr,O,, AnalaR grade recrystallisdd 3 t m e s .. .. .. .. K&O, prepared from K,CrO, . . Average deviation from means . . Number of results . . .. .. K&07, P.V.S. .. .. KJrO, . . .. .. .. .. Mean &CrSO, found,* per cent. f L \ Scheme A Scheme B Scheme C Average 100.03 (0.02) 100*00 (0.07) 99.99 (0-025) 100*01 100.01 (0.005) 99-93 (0.07) 99-94 (0-025) 99.36 100~00 (0.01) 99-96 (0.03) 100.06 (0.05) 100.01 99-81 (0.02) 99-96 (0.01) 99.98 (0.015) 99.91 99-55 (0.015) 100.08 (0.03) 100.01 (0.02) 99-88 (0-013) (0.04) (0.027) 23 21 33 * The average deviations from the means are given in brackets. The excellent results obtained on these samples by the iodimetric scheme C confirm the statements of many earlier workers21 that the iodimetric reactions are stoicheiometric under proper conditions. If the reaction of iron(I1) with dichromate were not stoicheiometdc, the results obtained by schemes A and B should be consistently and considerably different from those achieved by scheme C.It is obvious that there is no significant difference, and accordingly it can be stated with certainty that the iron(I1) - dichromate reaction is stoicheio- metric under the conditions tested and over a considerable range of iron(I1) concentrations (see above). The sample of dichromate prepared from potassium chromate is less satisfactory as a primary standard than the commercial samples of dichromate; this may be caused by irnpuri- ties present in the potassium chromate starting material, which is only guaranteed at the 99.5 per cent.level. When potassium chromate is used, the results obtained by scheme A are consistent with those suggested by the purity of the material; the results obtained by the other schemes indicate satisfactory purity and the inconsistency in this series of results is difficult to understand. Potassium dichromate re-prepared from the analytical-reagent grade material, in an attempt to remove possible chromium(V1) oxide impurities, showed low and inconsistent recoveries.la The differences between the results obtained by the present schemes and those reported earlieI-4,6,6 can probably be explained on the basis of the reaction chains used in the earlier work. Eppley and Vosburgh4 based their standardisation on sodium oxalate via titration with permanganate, as did Furness.6 However, it has been s ~ o w ~ ~ ~ , ~ that standardisation of permanganate solutions against oxalate can readily lead to errors of +0*2 to 0.4 per cent.because of the complicated nature of the reactions involved. This error would be carried through the standardisation schemes and could be at least a partial cause of the confusion regarding the stoicheiometry of the iron(I1) - chromium(V1) reaction. REFERENCES 1. 2. 3. 4. 5. 6. Belcher, R., Rees, D. I., and Stephen, W. I., Chtim. Analyt.. 1959, 397. Belcher, R., Brazier, J. N., and Stephen, W. I., Talanta, 1966, 12, 778. --- , Ibid., 1965. 12, 963. Eppiey, M:, and Vosburgh, W. G., J . Amer. Chem. S O ~ . , 1922, 44, 2148. Willard, H. H., and Gibson, R. C., Ind. Engng Chem. Analyt. Edn, 1931, 3, 88. Fumes, W., Analyst, 1950, 75, 2.January, 1969J POTASSIUM DICHROMATE AS AN OXIDIMETRIC STANDARD 25 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Smith, G. F., and Brandt, W. W., Analyt. Chem., 1949, 21, 948. Laitinen, H. A., “Chemical Analysis,” McGraw-Hill Book Co. Inc., New York, Toronto and London, 1960. Kolthoff, I. M., and Furman, N., “Potentiometric Titrations. A Theoretical and Practical Treatise,” J. Wiley and Sons, New York, 1931. Westheimer, F. H., Chem. Rev., 1949, 45, 419. Brazier, J. N., M.Sc. Thesis, University of Birmingham, 1966. Chakrabarti, C. L., M.Sc. Thesis, University of Birmingham, 1960. LaMer, V. K., and Friedman, H. B., Ind. Engng Chem. Analyt. Edn, 1930,2, 54. Incze, G., 2. analyt. Chem., 1915, 54, 406. Sarver, L. A., and von Fischer, W., Ind. Engng Chem. Analyt. Edn, 1935, 7 , 271. Smith, G. F., and Diehl, H., Taluntu, 1959, 2, 382. Willard, H. H., and Young, P., J. Amer. Chem. SOC., 1933, 55, 3260. Walden, G. H., Hammett, L. P., and Chapman, R. P., Ibid., 1931, 53, 3908. J I , Ibid., 1933, 55, 2649. Gleu, K., 2. analyt. Chem., 1933, 95, 305. Kolthoff, I. M., and Belcher, R., “Volumetric Analysis,” Volume 111, Interscience Publishers Vosburgh, W. G., J. Amer. Chem. SOC., 1922, 44, 2120. Fowler, R. M., and Bright, H. A., J. Res. Natn BUY. Stand., 1935, 15, 493. Kolthoff, I. M., Laitinen, H. A., and Lingane, J. J., J. Amev. Chem. SOC., 1937, 59, 429. --- Inc., New York and London, 1957. Received August 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400020

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

26 Analyst, January, 1969, Vol. 94, $$. 26-31 The Spectrof luorimetric Determination of Europium and Terbium BY R. BELCHER, R. PERRY AND W. I. STEPHEN (Department of Chemistry, The University, P.O. Box 363, Birmingham 16) The fluorescent properties of the t r i s chelates of europium and various @diketones have been examined. Fluorinated ligands give better sensitivity than unfluorinated ligands, but the best sensitivity is found with 2-thenoyl- trifluoroacetone in dimethylformamide. A rapid method for the determina- tion of 10-7 to 10-8 M europium(II1) is discussed. Both europium and terbium can be readily determined in the lo-* to M range by measurement of the fluorescence of the complexes formed with dimethylformamide alone. DURING a study of the tetrakis chelates of certain rare earth elements with several fluorinated /3-diketones and alkali metals,l it became apparent that the fluorescence of some of the rare earth complexes was sufficiently marked to warrant an independent study of this characteristic as a means for their determination.The fluorescence of several rare earth compounds is well established, but this fact has been used only to a limited extent analytically, perhaps because of the weak fluorescence provided in aqueous solutions. By forming the carbonato-complexes in concentrated potassium carbonate solutions, Taketatsu, Carey and Banks2 obtained useful increases in the fluorescence of europium and terbium, which led to the determination of both elements at concentrations down to 4.0 and 0.3 pg per ml, respec- tively. Alberti and Massucci3 have used the tungstate complexes of samarium, europium terbium and dysprosium for spectrofluorimetric determinations in the range 10-1 to 10-3 pg per ml, but there are considerable problems from mutual enhancement in rare earth mixtures.Considerable work has been done by Poluektov, Kononenko et aZ. on the fluorescent ternary complexes of europium, samarium and terbium, particularly those which contain 1,lO-phenanthroline as one of the ligand molecules. For example, benzene solutions of diphenanthroline trisalicylate complexes of europium(II1) and terbium( 111) have been used in the combined extraction and fluorimetric determination of these metals at concentrations down to 0-05 pg per ml.4 A somewhat similar procedure requiring 1,lO-phenanthroline and 2-thenoyltrifluoroacetone has been developed for the extraction and fluorimetric determination of europium and samari~m,~ and has been applied to tributyl phosphate extracts6 Direct fluorimetry of the ethanolic solutions of the europium and terbium chelates of dibenzoylmethane, after cooling in liquid nitrogen or oxygen, has been used to determine the two elements in monazite'; the sensitivity is poor (0.01 per cent.), but can be improved to lo4 to 10-6 per cent.by the formation of the ternary complex with 1,lO-phenanthroline. Ballard and Edwards have discussed procedures for europium and samarium based on extraction of the 2-thenoyltrifluoroacetone - trioctylphosphine oxide complexes ; the sensitivities appear to be about 10 ng of europium per ml and 2 pg of samarium per ml at room temperature.8 In the present work, it has been shown that a considerable increase in the sensitivity of the spectrofluorimetric determination of europium with 2-thenoyltrifluoroacetone (TTA) can be achieved by carrying out the reaction in a medium of dimethylformamide or acetonitrile.Various other p-diketones have been evaluated for the same purpose, but none is superior to 2-thenoyltrifluoroacetone. The use of these solvents also simplifies the procedure because the metal ion is rapidly and quantitatively converted into its p-diketonate at controlled pH. The solvents themselves form fluorescent complexes with europium and with terbium, which can be used in a rapid direct procedure for the determination of these metals at concentrations down to about 15pg per ml.0 SAC and the authors.BELCHER, PERRY AND STEPHEN 27 EFFECT OF SUBSTITUTION IN THE /?-DIKETONE- It has been shown that the substituents in the /3-diketone molecule have a considerable effect on the volatilities of their rare earth chelatesl and on the extractability of the complexes of the europium /3-diketonate with tributyl pho~phate.~ Accordingly, the effect of substitution on the intensity of fluorescence was examined. The limits of detection, i.e. the limiting concentration at which fluorescence under ultraviolet light is just visible compared with a TABLE I FOR EUROPIUM(III) EFFECT OF /3-DIKETONE SUBSTITUTION ON THE LIMIT OF DETECTION Limit of Range of linearity, Excitation Ligand detection M wavelength* Acetylacetone .. .. .. . . 1.5mgperml - 380 Dipivaloylmethane . . .. .. . . 0.16 mg per ml - 390 Dibenzoylmethane . . .. .. .. 15 pgperml 10-4 to 10-2 390 Tduoroacetylacetone . . .. . . 0.75 pg per ml 5 x to 6 x 366 Trifluoroacetylpivaloylmethane . . . . 0.15 p g per ml 10-6 to 7 x 10-6 296 Pentafluoropropionylpivaloylmethane . . 16 ng per ml 10-7 to 10-6 310 Heptafluorobutanoylpivaloylmethane . . 16 ng per ml 10-7 to 10-6 326 Benzoyltrifluoroacetone . . .. . . 16ngperml 10-7 to 10-6 320 Hexafluoroacetylacetone . . .. . . 15ngperml 10-7 to 10-5 336 2-Thenoyltrifluoroacetone . . .. .. 0-76ngperml 6 x 10-oto6 x 390 * The optimal excitation wavelengths depend to some extent on concentration. blank, obtained for europium(II1) in dimethylformamide media are shown in Table I.It can be seen that fluorine-containing /?-diketones provide considerably greater sensitivity than those containing no fluorine, and that 2-thenoylt~uoroacetone provides a sensitivity almost 102-fold better than those obtained with the best fluorinated pivaloyl or benzoyl derivatives. All the present work was done with the tris rather than the tetrakis chelates, because it was considered that dissociation of the tetrakis chelate would occur in dimethylformamide or acetonitrile at high dilutions. Moreover, the quantitative formation of the tetrakis chelate would involve the introduction of moderate amounts of monovalent cations, which would have a quenching effect on the fluorescence. h c W U C .- In .- 8 x !2 s W - LL 1 615 690 Wavelength, nm Fig. 1. Emission spectrum of M europium tris( S-thenoyltri- fluoroacetone) in dimethylform- amide or acetonitrile : excitation wavelength, 390 nm I I I I 590 615 655 Wavelength, nm Fig.2. Emission spectrum of M europium tris(trifluoroacety1- pivdoylmethane) in dimethyl- formamide or acetonitrile: excita- tion wavelength, 295 nm SPECTRA OF THE FLUORESCENT RARE EARTH /3-DIKETONATES- The fluorescence spectra of all the europium(II1) /?-diketonates in acetonitrile or di- methylformamide show strong emission over a narrow band at 615 nm. Typical spectra are shown in Figs. 1 and 2. At pH 7.5, only the europium(II1) and samarium(II1) - tris(2- thenoyltrifluoroacetone) chelates show fluorescence. It can be seen from Fig. 3 that the28 BELCHER, PERRY AND STEPHEN THE SPECTROFLUORIMETRIG [A?Za&St, VOl.94 fluorescence of the samarium chelate is comparatively weak. No chelate is formed between terbium and 2-thenoyltrifluoroacetone, although there is fluorescent chelate formation with some other #I-diketones. For example, the spectrum of the terbium - tris(trifluoroacety1- pivaloylmethane) chelate in acetonitrile, on excitation at 360nm, shows a major peak at 545 nm and a minor peak at 490 nm, but the fluorescence is weak compared with that of the europium chelate, the limit of detection being about 10 pg of terbium per ml. I -1 Wavelength, nm Fig. 3. Emission spectrum of lo-* M samarium tris(2-thenoyltri- fluoroacetone) in acetonitrile : ex- citation wavelength, 400 nm During these studies, it became apparent that simple dissolution of europium or terbium salts in acetonitrile or dimethylformamide could form the basis of a reasonably sensitive spectrofluorimetric determination of the two metals.Fluorimetric titration of a standard ethanolic europium nitrate solution containing 80 mg of europium with dimethylformamide indicated that a complex containing europium and dimethylformamide in a ratio of 1:6 was formed. EXPERIMENTAL REAGENTS- Acetylacetone (pentane-2,4-dione), dibenzoylmethane, hexafluoroacetylacetone (1,1,1 , 6,S,S-hexafluoropentane-2,4-dione), benzoyltrifluoroacetone (l-pheny1-4,4,4-trifluorobutane- 1,3-dione) and t henoyltrifluoroacet one (I -t henoyl-4,4,4-t ri fluorobut me- 1,3-dione) were obtained commercially. Acetylacetone and hexafluoroacetylacetone were purified by distil- lation in vacuo and from phosphorus pentoxide, respectively.Thenoyltrifluoroacetone was purified by vacuum sublimation. Dibenzoylmethane and benzoyltrifluoroacetone were recrystallised from benzene - light petroleum mixtures. Trifluoroacetylacetone (l,l,l-trifluoropentane-2,4-dione) was prepared by condensing ethyl tdluoroacetate with acetone in the presence of sodium hydride; the crude material was purified by fractional distillation (boiling range 105" to 106" C). The various pivaloyl derivatives were prepared by condensing the appropriate ester with pinacolone in the presence of sodium hydride, and the crude products were purified by distillation in vacuo. The derivatives prepared and the starting esters were as follows: dipivaloylmethane (2,2,6,6-tetramethylheptane-3,5-dione) from methyl pivalate; trifiuoro- acetylpivaloylmethane (1 , l,l-trifluoro-5,5-dimethylhexane-Z,4-dione) from ethyl trifiuoro- acetate ; pentafiuoropropionylpivdoylmethane (1 , 1 ,1,2,2-pentafiuoro-6,6-dimethylheptane- 3,5-dione) from ethyl pentafiuoropropionate ; and heptafluorobutanoylpivaloylmethane ( 1,l , 1,2,2,3,3-hep t afluoro-7,7-dime thyloc t ane-4,6-dione) from e thy1 hepta.fiuorobut yrate .Dimethylformamide and acetonitrile (laboratory-reagent grade) were used without further purification. The ethanol was of spectroscopic grade. The rare earth oxides used were obtained from Koch-Light Ltd., and were of 3N grade.January, 19691 DETERMINATION OF EUROPIUM AND TERBIUM 29 APPARATUS- The fluorescence spectra and measurements were made with a Farrand spectrophoto- fluorimeter in 10-mm silica cells.GENERAL ANALYTICAL PROCEDURES- M-Dissolve the sample of europium oxide in 2ml of a mixture of concentrated, analytical-reagent grade nitric and hydrochloric acids (10 + 1) and evaporate almost to dryness; if a solution is available, evaporate this almost to dryness. Dissolve the residue in dimethylformamide and dilute to volume in a 26-ml calibrated flask. Take a suitable aliquot of this solution, neutralise it to pH 7.5, measured on a pH meter, with aqueous M ammonia solution and add an excess of 2-thenoyltrifluoroacetone as a 1 per cent. w/v solution in dimethylformamide. Dilute the mixture to volume with dimethylformamide in a suitable calibrated flask, so that the final solution is lo-' to M in europium.Measure the fluorescence intensity with excitation and emission wavelengths of 390 and 615 nm, respectively. Method B for concentrations of eurofiium and for terbium greater than 1 x M- Dissolve the sample as described above and evaporate the solution almost to dryness. Dissolve the residue in dimethylformamide and dilute to volume in a 25-ml calibrated flask. Measure the intensity of fluorescence as follows: for terbium, use excitation and emission wavelengths of 370 and M5nm, respectively, and for europium use corresponding wavelengths of 390 and 616nm. Method A for covtcentrations of europium less thart 1 x RESULTS AND DISCUSSION METHOD BASED ON /I-DIKETONATE FORMATION- All of the P-diketones examined gave linear calibration graphs for appropriate ranges above the limit of detection.The ranges are shown in Table I. Fig. 4 shows the graph obtained for europium(II1) - tris(2-thenoyltduoroacetone) in dimethylformamide media ; acetonitrile media gave a slightly poorer sensitivity than dimethylformamide. Self-quenching of the L l l l l l l l l l [Eu (ITA),] x 10 Fig. 4. Calibration graphs for the europium tris (2-thenoyltrifluoroacetone) method : excita- tion wavelength, 390 nm: A, in dimethyl- formamide; and B, in dimethylformamide in the presence of lo-' M samarium(II1) fluorescence tends to occur at concentrations about 102 to 103 M above the limits of detection with all the p-diketones examined. The calibration graph for europium(II1) - tris(2-thenoyl- trifluoroacetone) is linear up to about 104 M europium(II1).In contrast, the graph for the80 BELCHER, PERRY AND STEPHEN THE SPECTROFLUORIMETRIC [AfldJGt, VOl. 94 corresponding trifluoroacetylpivaloylmethane complex is linear only over the range 10" to 7 x 104 M europium(III), self-quenching occurring at lo6 M concentrations. TABLE I1 DETERMINATION OF EUROPIUM WITH 2-THENOYLTRIFLUOROACETONE Europium present, ng 14.8 7.4 1.4 16.7 7.8 1.6 Europium found, ng 13.8 7.1 1.3 14.9 7.1 1.6 Error, - 1.0 Dimethylf ormamide - 0.3 Dimethylformamide -0.1 Dimethylformamide - 0.8 Acetonitrile - 0.7 Acetonitrile 0.0 Acetonitrile ng Medium Some results obtained for nanogram amounts of europium are shown in Table 11. The fluokescence is stable for about 1 hour, and then gradually fades over 5 to 6 hours. In some instances, it might be more convenient to use one of the less sensitive fl-diketones, but the high sensitivity of the 2-thenoyltrifluoroacetone method can be advantageous when foreign ions are present because the great dilution needed for the fluorimetric measurement can often result in dilution of the accompanying ions to a level at which they no longer interfere either by their own fluorescent properties or their quenching propensities. The interference of samarium(II1) in the determination of europium is of interest. Although samarium - tris(2-thenoyltrifluoroacetone) fluoresces in the red region (Fig.3) , the emission peaks at 600 and 646 nm are comparatively small and can readily be resolved from the peak of the europium chelate at 616nm. Accordingly, any interference is caused by straightforward quenching effects; thus there is no interference when an equivalent amount of samarium is present, but greater than 10-fold amounts must be corrected for by prepara- tion of suitable standard solutions.A graph prepared for europium in the presence of M samarium(II1) is shown in Fig. 4. According to Fig. 3, samarium(II1) could itself be determined by means of its complex with 2-thenoyltrifluoroacetone in concentrations above M. The relative sensitivities for the europium and samarium chelates are thus in approximately the same ratio as those found by Ballard and Edwards8 in the presence of trioctylphosphine oxide. METHODS BASED ON DIRECT MEASUREMENT IN DIMETHYLFORMAMIDE MEDIA- When europium is present in concentrations greater than 15 pg per ml, determination by direct measurement of the fluorescence in dimethylformamide or acetonitrile is simple, and the same procedure can also be used for the determination of terbium.For europium(II1) the calibration graphs in both dimethylformamide and acetonitrile media are linear over the range lo4 to 1 0 - 2 ~ ; considerably better sensitivity is available in the dimethylformamide media. For terbium(II1) the sensitivity of the measurement is slightly less than that obtained for europium, but calibration graphs are again linear over the range M in dimethyl- formamide. The graphs obtained for terbium in acetonitrile media are inferior in linearity and show less sensitivity. These methods are comparable in sensitivity with the fluorescence obtained with the chloro complexes of samarium, europium, gadolinium, terbium and dysprosium, as indicated by Parker.10 TABLE I11 DETERMINATION OF EUROPIUM AND TERBIUM IN DIMETHYLFORMAMIDE MEDIA to Europium present, mg 1-48 0.74 0.1s - - 0.76 0.08 Terbium present, mg - 1-53 0.76 0.16 0.74 0-07 Europium found, mg 1.60 0.76 0.16 - - 0.69 0.07 Terbium found, mg - 1.66 0.72 0.16 0.76 0.07January, 19691 DETERMINATION OF EUROPIUM AND TERBIUM 31 Because of the large difference in the emission wavelengths for europium and terbium, a simultaneous determination of the two ions is possible; some results are shown in Table 111.The increasing commercial importance of europium makes the provision of simple, direct methods for its determination a matter of some importance. It is to be expected that the procedures described in this paper may provide the basis for various applications in the several technical processes in which europium compounds are used. We thank Dr. N. Crawford of the Physiological Chemistry Department for the loan of the spectrofluorimeter, and one of us (R.P.) thanks the Science Research Council for a s tuden tship . 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. REFERENCES Belcher, R., Majer, J. R., Perry, R., and Stephen, W. I., J . Inorg. Nucl. Chem., in the preas. Taketatsu, T., Carey, M. A., and Banks, C. V., Tuluntu, 1966, 13, 1081. Alberti, G., and Massucci, M. A., Andyt. Chem., 1966, 38, 214. Kononenko, L. I., Lauer, R. S., and Poluektov, N. S., Zh. Anulit. Khim., 1963, 18, 1468. Kononenko, L. I., Poluektov, N. S., and Nikonova, M. P., Zuv. Lab., 1964,30, 779. Melent’eva, E. V., Tishchenko, M. A., Vitkun, R. A., Kononenko, L. I., and Poluektov, N. S., Khomenko, V. S., Kuznetsova, V. V., and Pekarskaya, L. A., Zh. PrikZ. Spektrosk., 1967, 6, 117. Ballard, R. E., and Edwards, J. W., in Shallis, P. W., Editor, “Proceedings of the SAC Conference, Sekine, T., and Ono, M., Bull. Chem. SOC. Japan, 1966, 38, 2087. Parker, C. A., “Photoluminescence of Solutions,” Elsevier Publishing Company, Amsterdam, 1968, Received August 16th, 1968 Referut. Zh., Khim, AGD. 1966, Abs. No. llG81. Nottingham, 1966,” W. Heffer and Sons Ltd., Cambridge, 1966, p. 328. p. 492.

ISSN:0003-2654    

DOI:10.1039/AN9699400026

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

32 Analyst, January, 1969, Vol. 94, $9. 32-38 An Analytical and Kinetic Study of the Bromate Oxidation of Vanadium (IV) in Acid Medium* BY C. W. FULLER AND J. M. OTTAWAY (Depadment of Pure and Afifilied Chemistvy, University of Strathclyde, Cathedral Street, Glasgow C.l.) A kinetic and analytical study of the reaction of vanadiurn(1V) with The kinetics were studied in perchloric potassium bromate is reported. acid media and the reaction was found to obey the rate equation --d [V(IV)l - kOK1 cV(IV)l L-BrO*-I dt - 1 + K, [BrO,-] A value for k,, of 7.50 minute-' was obtained at 20" C and an ionic strength of 0.10 M. Values for K, of 88.8 hole-1 and for the activation energy of 19 kcal. mole-l were obtained, and a possible mechanism is proposed for the reaction. The determination of vanadium(1V) by photometric titration with potassium bromate is feasible a t concentrations of vanadium(1V) down to 2 x An accuracy of f0-5 per cent. is obtained for concentrations of vanadium(1V) between 6 and 10 x 1 0 - 8 ~ .Under these acid conditions bromate acts as a five- electron oxidant and bromine is a primary product of the reaction. WHEN determined by an oxidation - reduction process, vanadium(1V) is usually titrated with cerium(1V) or manganese(VI1). In accurate work it is useful to have available a standardisation technique that involves the direct use of a primary standard, and for various reasons neither of the above titrants is satisfactory as a standard. In addition, other adverse effects of the cerium(1V) - vanadium(1V) titration have been investigated by Rechnitz and Ra0.l Vanadium(V) is usually determined by reduction with an iron(I1) ~olution~ps~~; the latter, being unstable to air oxidation: must be prepared or standardised immediately before use.It is surprising that potassium bromate, an accepted primary standard,* has not been previously suggested as a standard oxidant for vanadium( IV). However, potassium bromate has been used as an oxidant for vanadium in acid solutions. Willard and Young6 in 1928 used bromate to oxidise vanadium(1V) by an excess method, removing the excess of bromate by boiling in the presence of ammonium sulphate and hydrochloric acid. TomiEek, Stodolova and Hehans titrated bromate potentiometrically with vanadyl acetate in glacial acetic acid at 60" to 80" C. Under these conditions bromate acted as a six-electron oxidant.Apart from the elevated temperatures required to establish the potentials of the vanadium(V) - vanadium( IV) couple and the inconvenience of using anhydrous glacial acetic acid, the greatest difficulty in their method arises from the instability of vanadium(1V) solutions at pH values above 3, when air oxidation is important,' particularly at these high tempera- tures. They also titrated bromine, among many other substances, with vanadyl acetate under similar conditions. Erdey and MAzor8 have reported the titration of vanadium(I1) in sulphuric acid medium with potassium bromate and concluded that the vanadium was oxidised only to vanadium(1V). After the oxidation of vanadium(I1) to vanadium(III), bromine is liberated in solution by the reaction of bromate with bromide and thereafter vanadium(II1) is oxidised to vanadium(1V).Yatsimirskii and Kalininag have described a kinetic investigation of the bromate - vanadium(1V) reaction, in which they state that bromate is reduced to bromide by vanadium(1V) in the hydrogen-ion concentration range 0.3 to 0-001 M. There thus appeared to be two conflicting opinions on the extent of the bromate - vanadium(1V) reaction, (i) that bromate reacts with vanadium(1V) to give vanadium(V) and bromide as product^^^^ and (iz) that bromate does not react with vana- dium(IV).a The present paper describes a kinetic and analytical study of the vanadium(1V) - bromate reaction and attempts to elucidate the nature and usefulness of the reaction.M and at sulphuric acid concentrations above 0.06 M. * Paper presented at the Second SAC Conference 1968, Nottingham. 0 SAC and the authors.FULLER AND OTTAWAY 33 EXPERIMENTAL REAGENTS- Potassizlm bromate, 0.1313 M-Analytical-reagent grade potassium bromate was dried at 120" C for 2 hours, then the required weight was dissolved in water and diluted. Vanadizm(IV) , 0.2017 M-The required amount of vanadyl sulphate was dissolved in water containing sufficient sulphuric acid to make the final solution, after dilution, 0.052 M in sulphuric acid. . Szllphuric acid, 5.200 M-Analytical-reagent grade sulphuric acid was diluted about four times and then standardised against sodium carbonate, with methyl orange as indicator. Perchloric acid, 1.014 M-Perchloric acid, 60 per cent.w/w, was diluted about ten times and then standardised against sodium carbonate, with methyl orange as indicator. PHOTOMETRIC TITRATIONS- Photometric titrations of vanadium(1V) with potassium bromate were carried out with an Evans Electroselenium Ltd. titrator coupled with an Evans Electroselenium Ltd. Unigalvo. Filter number 609 (which has a band pass at wavelengths greater than 660 nm and is supplied with the instrument) was used. At this wavelength only vanadium(1V) absorbs. Titrations were performed in 50-ml capacity cells, which have an approximate path length of 2.6 cm. The titration results were verified by standardisation of both the vanadium(1V) and potassium bromate solutions against a standard ammonium iron(I1) sulphate solution as reference.Potassium bromate was titrated directly against the iron(I1) solution* by using potentiometric end-point detection.1° The vanadium(1V) was titrated potentiometrically in the usual manner with ammonium cerium(1V) sulphate: which was itself standardised against the same iron(I1) solution,4 potentiometrically. The titration results were found to be internally and theoretically consistent. The titrant volume for the bromate - iron(I1) titration agreed with that required to give consistency with the concentrations of the solutions obtained by direct weighings. In these titrations the ammonium iron(I1) sulphate solution was used within 3 to 4 hours of preparation, to avoid any possible air oxidation. The bromate titration of vanadium(1V) was investigated over the sulphuric acid concen- tration range 0.002 M to 5.00 M.The normal mode of titration was to add 0.1-ml increments of titrant then allow a 3 to 4-minute interval before recording the absorbance of the solution. The effects of larger and smaller titrant increments and time intervals were studied, but were found to be of no great advantage, the saving in titration time having to be set against the loss of accuracy, and vice versa. The titration time was usually in the region of 30 to 45 minutes. The smaller titrant increments and longer time intervals were only used in the initial standardisation of the vanadium( IV) solution, when greater accuracy was required. All titrations were performed at room temperature. KINETIC STUDIES- Kinetic investigations were carried out spectrophotometrically, with a Hitachi - Perkin- Elmer 137 spectrophotometer to which was attached a Honeywell Electronik 15 strip-chart recorder, to give a continuous record of the reaction.All solutions were maintained thermo- statically for 30 minutes before an experiment, and a 1-cm cell was kept ready for use in the thermostatically controlled cell compartment of the spectrophotometer. The choice of initial reactant concentrations was ruled to a great extent by the rate of reaction, the instrument requirements and the experimental technique. It was found possible to follow reactions with half-lives as low as 20 seconds and still obtain results that could be interpreted with accuracy. These requirements fixed a suitable vanadium(1V) concentration at 104 M, and allowed bromate and hydrogen ions to be present in concentra- tions at least five times in excess to maintain pseudo first-order kinetic conditions.At a wavelength of 270nm and the concentrations outlined above, vanadium(V) was found to absorb while vanadium(IV), perchlorate and bromate did not. Concentrations of potassium bromate in the range 7 to 20 x 10" M and perchloric acid in the range 0.02 to 0.40 M were studied at 20-0 4 0.1" C, while the initial vanadium(1V) concentration was kept at 1.009 x 1 0 " ~ and the ionic strength was maintained by using either sodium or lithium perchlorate. The effect of ionic strength variations from 0.05 to 1-00 M and the temperature dependence of the reaction from 12' to 27" C were also investigated.34 KINETICS- From the kinetic experiments performed under different bromate and perchloric acid concentrations, the bromate oxidation of vanadium(1V) was found to follow the kinetic rate law FULLER AND OTTAWAY: AN ANALYTICAL AND KINETIC STUDY OF [Analyst, Vol.94 RESULTS AND DISCUSSION .. . . (1). d [v(v)i - d [ v ( ~ v ) i = k , [~ro,-i ~V(IV)I dt dt 1 + k , [BrO,-I The values of k, and k, were 666lmole-l minute-, and 88*8lmole-l, respectively, at 90" C and an ionic strength of 0.10 M. The most probable reaction path is K1 BrO,- + V02+ =: Br0,-V02+ fast .. .. * (2) slow . . .. (3) BrO; + V02+ -+ products fast .. .. . . (4). .. .. (5) k0 Br0,-V02+ -+ V02+ + BrO; The rate-determining step is then the decomposition of the Br08-VOZ+ species. Rate = k, [Br0,-V02+] .. .. .. [BrO,-V02+] K1 = [Bfi*-] [ V 0 2 + ] - .... If [v03++lT and @3r0,-lT represent the total concentrations of these species present then, as bromate is present in considerable excess, but [BrOs-] [BrO,-)T wO2+lT = [VOZ+] + [Br0,-V02+] .. .. . . (7). Re-arranging (6) and (7) and substituting in equation (6) gives By comparison with the experimental rate equation it is seen that k , corresponds to the equilibrium constant K1 for the formation of the Br0,-V02+ species, while k , corre- sponds to the product of the first-order rate constant k, and the equilibrium constant K,. This allows a value of 7-50 minute-l to be obtained for k,. The temperature dependence of this rate constant gave a straight line when log,dz, was plotted against the reciprocal of the absolute temperature in the range 285" to 300" K.An activation energy of 19 & 1 kcal. mole-l and an associated frequency factor of 1 to 3 x 1013 s-l were obtained at an ionic strength of 0.050 M. The intermediate Br0,-V02+ is postulated as an actual species and not merely as an active intermediate. There are three possible structures for such a species- BrO,-. VO*+ i 0 V I 0-Br-0 I b 1 b Br o/ '0 I I1 I11 The reaction proceeds by transfer of an oxygen atom from bromate to vanadium. Structure I involves bonding of the bromine to vanadium and would incorporate the 4p orbitals on the bromine and the vacant 3d orbitals of vanadium. Oxygen transfer could occur through %bonding existing between 2p orbitals of oxygen on the bromate and vacant 3d orbitals on the vanadium. Structure 11, which is thought to be the most probable, would involve bonding between the 2p orbitals of one oxygen on the bromate and the vacant 3d orbitals of vanadium.Oxygen transfer is then simply a matter of breaking the bromine-oxygen bond.January, 19691 THE BROMATE OXIDATION OF VANADIUM(IV) IN ACID MEDIUM 35 Structure I11 describes the possibility of the species existing merely as an ion pair, with oxygen transfer occurring possibly through a combination of structures I and 11. Oxygen transfer has been established in an analogous reactionll ,l2 together with the postulation of a charge-transfer type of intermediate. Although no satisfactory accuracy can be associated with the value for the frequency factor, it is interesting that the value obtained falls in the range of bromine - oxygen vibrational frequencies for bromate (1-07,1-26, 2-42, and 2.51 x 1013 s-1).13 This is to be expected if the rate-determining step is the uni- molecular decomposition proposed.TABLE I EFFECT OF HYDROGEN-ION CONCENTRATION AND IONIC STRENGTH ON THE RATE OF THE VANADIUM(IV) - BROMATE REACTION 1.009 x 10-4 M 1.313 X 10-'M All [v(Iv)l t-KBrO,l concentrations in moles 1-' [HC10,1 [LiClO,] 0.0203 - 0.0507 - 0.1014 - 0.2028 - 0.4066 - 0.0203 - 0.0507 0.088 0.0507 0.176 0.0507 0.629 0.0507 - 0-0507 - [ NaCIO,] ko. minute-] - 9.76 - 8.15 - 7-55 - 6.34 - 5-27 0-386 6.55 - 7.54 - 7-13 - 6-44 0.664 6-15 0.940 5-63 Both hydrogen-ion and ionic strength effects were observed (Table I), but the former can be ascribed to a special type of the latter. The salt effect is slight and does not approach the usual Bronsted - Bjerrum dependence.The rate constant was found to decrease with increasing ionic strength. There appeared to be slight variations depending on the use of lithium or sodium perchlorates, but these variations fell within experimental deviations, and were thought to be of no great significance. The greatest effect was observed when the perchloric acid concentration was varied, but maintaining the ionic strength constant had the effect of reducing the variations. The ionic strength effect is attributed to some small effect on the bromate-vanadium(1V) equilibrium, with the increased effect caused by perchloric acid either being a factor of the small size of the hydrogen ion or the effect that the hydrogen-ion concentration has on the bromate species that are present Br0,- + H+ + HBrO, .. .. .. * * (9) HBrO, + H+ + HJ3rO,+ . . .. .. . . (10). HBrO, + V02+ s Br0,-V02+ + H+ .. .. . . (11). Previously only equilibrium (2) was considered, but another possibility is K, This equilibrium would suffer from a different ionic strength effect from that observed for equilibrium (2). This mechanism gives rise to the rate equation corresponding to equation (8) and is again independent of hydrogen-ion concentration. Thus, although the experimental value of k, was ascribed as equivalent to K,, it is more likely to be a combination of con- tributions by K, and K,. By a simple companson it is seen that the kinetic results presented here vary from those of Yatsimirskii and K a l i ~ h a , ~ and while any justification of one set of results rather than the other is difficult, there does appear to be a definite ionic strength effect on the reaction (Table I), although Yatsimirskii and Kalinina state that no such effect exists.The experimentally observed rate equation described here also seems to fit the reaction mechanism more closely. REACTION PRODUCTS- of the bromate - vanadium(1V) reaction. As mentioned earlier, there is some confusion in the literature as to the course and nature Potentiometric titration of vanadium(1V) by36 FULLER AND OTTAWAY: AN ANALYTICAL AND KINETIC STUDY OF [Analyst, VOl. 94 bromate under these conditions was found to be impossible, as bromine produced in the course of the reaction effectively controlled the potential of the platinum indicator electrode from the beginning of the titration and no change in the potential was observed at the end-point.However, it is an obvious extension of the normal photometric titration procedure for complexometric titrations, in which a small equilibrium constant causes curvature in the titration graph around the end-point, to applications involving a slow reaction when curvature again occurs but is now caused by the reaction becoming too slow in the region of the end- point. Provided the reaction is not so slow as to cause complete curvature of the graph, it should be possible to obtain an accurate end-point by extrapolation of the initial and final linear portions of the graph, when the reaction becomes more or less complete in a short Ol I I I I I I I 0 0.200 0.400 0.600 0.800 I~OOO 1.200 1-400 Volume of 0.1284M bromate added, mi Fig.1. Photometric titrations of 2 ml of 0.2017 M vanadium(1V) with 0.1284~ potassium bromate, in a 1.04 M sulphuric acid medium. The titration volume was 60 ml and the equivalence volume 0.628 ml time because of the presence of an excess of one or other of the reagents. Fig. 1 shows a typical photometric titration of vanadium( IV) under the conditions outlined earlier. The photometric titrations showed that bromate acts as a five-electron oxidant in its reaction with vanadium(1V) at sulphuric acid concentrations from 5 . 0 0 ~ down to 0 . 0 6 ~ . 2Br0,- + 10V02+ + 4H20 --+ 10V02+ + Br, + 8H+ Below this region the reaction begins to change towards bromate acting as a six-electron oxidant.Concentration of sulphuric acid, M Fig. 2. Dependence of the extent of the reduction of bromate by vanadium (IV) on the sulphuric acid concentration The variation of the extent of the reduction of bromate in its reaction with vanadium(1V) at various sulphuric acid concentrations is shown in Fig. 2. This shows that the reaction is quantitative over a wide range of acid concentrations and was shown to be useful inJanuary, 19691 THE BROMATE OXIDATION OF VANADIUM(IV) IN ACID MEDIUM 37 determining vanadium(1V) at concentrations down to 2 x 1 0 - 3 ~ . The accuracy in deter- minations of 6 to 10 x M vanadium(1V) solutions was k0.5 per cent., with the accuracy of any titration being mainly a function of the extrapolation procedure involved in the tech- nique used to obtain the end-point.This titration procedure has all the advantages and disadvantages associated with photometric titrations, and the method is thought to be at least as useful as the accepted cerium(1V) or manganese(VI1) titrations of vanadium(IV), with the added advantage that potassium bromate is a convenient primary standard. The analytical results that have been described raise two interesting questions that remain to be explained. The first is whether bromine is a primary or a secondary reaction product, and the second, why the reaction course changes at sulphuric acid concentrations below 0-06 M. There are three possible reaction courses- (i) (a) Br03- + 6V02+ + 3H20 --+ 6VO,+ + 6H+ + Br- then, either (b) Br0,- + 5Br- + 6H+ -+ 3Br, + 3H20 or (c) 2VO,+ + 2Br- + 4H+ -+ 2V02+ + Br, + 2H20 Br0,- + 5V02+ + 2H20 -+ 5V02+ + 4Hf + Br (ii) The possibility of reaction course (i) (a) followed by (i) (c) can be ignored, as the oxidation of bromide by vanadium(V) is slow, even at high acid concentrations.Course (i) (a) and (i) (b) would appear to give the simplest answer to the two problems, i.e., bromate changes from a five to six-electron oxidant through a breakdown of step (i) (c) at lower acidities. This arises because, while step (i) (a) is virtually independent of hydrogen ion, step (i) (b) is dependent on a hydrogen-ion concentration squared term,14 i.e., as the hydrogen-ion concen- tration is lowered, step (i) (b) becomes too slow to be effective and the over-all reaction remains at the bromide stage.However, if one compares the rates of the bromate oxidation of vanadium(1V) and the bromate oxidation of bromide,14 one finds that even in M hydrogen ion the vanadium(1V) oxidation is three times faster, assuming equal concentrations of vanadium(1V) and bromide for the calculation. During the titration, however, the bromide, if present at all, would have a concentration at least two orders of magnitude smaller than the vanadium(1V) concentration, and this factor of three can be seen as a gross under- estimation of the difference in the competing rates. It seems safe to assume, therefore, that any bromate in solution would react preferentially with vanadium(1V) rather than any bromide that might be present, under these conditions. These assumptions were confirmed by the titration of solutions containing about equal concentrations of vanadium(1V) and bromide.Simultaneous titration of the two occurred, while mechanisms (i) (a) and (b) would necessitate the titration of bromide prior to the titration of vanadium(1V). It is possible to conclude, therefore, that bromine must be a primary and not a secondary reaction product, and the reaction proceeds by mechanism (ii). It is possible to arrive at a satisfactory conclusion for the second problem by considering the variation in redox potential of the vanadium(V) - vanadium(1V) couple in varying sulphuric acid media. Hart and PartingtonlS measured these variations, (Fig. 3), observing a fairly rapid fall in potential between 1.0 and 0.1 M sulphuric acid followed by a slight levelling out in the value before falling again at about 5 x M sulphuric acid, the potential having fallen to a value of about 700 mV versus N.H.E.at 10-2 M. As the potential of the vanadium(V) - vanadium(1V) couple is about 820 mV at 5 x M sulphuric acid, it would seem thermodynamically feasible that vanadium( IV) should be oxidised by bromine under these conditions. This was confirmed by adding bromine to vanadium(1V) in 10-1 M sulphuric acid when no oxidation was observed, while in 1 0 - 2 ~ sulphuric acid oxidation of the vanadium(1V) was apparent. These facts are in agreement with results obtained by TomiEek, Stodolova and Hehxqg who titrated both bromine and bromate with vanadium(1V) in acetic acid medium. In the presence of mercury( 11) the titration proceeds quantitatively, with bromate being reduced to bromide, i.e., acting as a six-electron oxidant.This effect is caused by the forma- tion of stable complexes between mercury(I1) and bromide,ls with the result that the potential 2Br ---+ Br,38 FULLER AND OTTAWAY of the bromine - bromide couple is increased to a calculated value of 1.69 V. Hence oxidation of vanadium(1V) by bromin; again occurs. I lo-’ 10-2 Concentration of sulphuric acid, M Fig. 3. Variation of the V(V)/V(IV) redox potential with respect to the normal hydrogen electrode as a function of sul- phuric acid concentration (from results published by Hart and Partington’s) The bromate oxidation of vanadium(1V) is considered to proceed by the formation of a bromate - vanadium(1V) species, the decomposition of which is the rate-determining step.This slow step involves oxygen transfer from bromate to vanadium(1V) to produce vana- dium(V) and BrO;. The lower unstable oxides of bromine react rapidly with vanadium(1V) to give free bromine at sulphuric acid concentrations above 6 x 1 0 - 2 ~ and bromide at lower acid concentrations. The change is caused by the reduction of the vanadium (V)- vanadium(1V) redox potential to a level at which bromine can effectively oxidise van- adium (IV) . Vanadium(1V) reacts directiy with potassium bromate and at a sufficiently rapid rate for direct photometric titrations to be carried out. The addition of bromide as an accelerating agent is not required in this sytem and is in fact deleterious as it reacts with bromate to produce bromine, which is not consumed by the vanadium(1V).This is in contrast to other common bromate determinations17 in which bromate acts as a six-electron oxidant, and the presence of bromide is essential for titrations to be carried out. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. REFERENCES Rechnitz, G. A., and Rao, G. N., Analyt. Chem., 1967, 39, 1192. Sarver, L. A., and Kolthoff, I. M., J. Amer. Chem. Soc., 1931, 53, 2906. Syrokomskii, V. S., and Antropov, L. J., Zav. Lab., 1940, 9, 818. Kolthoff , I. M., and Belcher, R., “Volumetric Analysis,” Volume 111, Interscience Publishers Inc., Willard, H. H., and Young, P., Ind. Engng Chem., 1928, 20, 764. TomiEek, O., Stodolovi, A., and Hefman, M., ChemicRd Listy, 1353, 47, 616. Selbin, J., Chem. Rev., 1966, 65, 163. Erdey, L., and Mizor, L., Acta Chim. Hung., 1963,3, 469. Yatsimirskii, K. B., and Kalinina, V. E., Im. Vjjssh. Ucheb Zaved Khim. i Khim. Tekhnol., 1966, Ottaway, J. M., and Bishop, E., in Shallis, P. W., Editor, “Proceedings of the SAC Conference, Kern, D. M. H., and Gordon, G., Theory Struct. Complex Compds., Papers Symposium Wroclaw, Gordon, G., and Kern, D. M. H., Inorg. Chem., 1964, 3, 1066. Nakamoto, K., “Infrared Spectra of Inorganic and Coordination Compounds,*’ J. Wiley & Sons, Skrabal, A., and Schreiner, H., Mh. Chem., 1934,65, 213. Hart, A. B.. and Partington, J. R., J. Chem. Soc., 1940, 1532. Laitinen, H. A., “Chemical Analysis.” McGraw-Hill Book Co. Inc., New York, Toronto and Ottaway, J. M., and Bishop, E., Analytica Chim. Acta, 1966, 33, 163. Received August loth, 1968 New York and London, 1967. 8 (3), 378. Nottingham, 1966,” W. Heffer and Sons Ltd., Cambridge, 1966, p. 316. Poland, 1962, 666 (Pub. 1964); Chem. Abstr., 1966, 63, 12668f. Inc., New York, 1963, p. 87. London, 1960, p. 430.

ISSN:0003-2654    

DOI:10.1039/AN9699400032

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Amdyst, January, 1969, Vol. 94, $$. 39-42 39 The Colorimetric Determination of Calcium in Silicate Minerals BY H. G. C. KING AND G. PRUDEN (Rothamsted Experimental Station, Harfienden, Herts.) Calcium in silicate minerals is determined in alkaline solution with glyoxal bis(2-hydroxy anil). At the pH of the reaction, aluminium, present as sodium aluminate, does not interfere, and co-precipitation of calcium with magnesium and iron is prevented by adding mannitol. Fifty micrograms of calcium can be determined in the presence of 3000pg of magnesium and 2000 pg of iron, the calibration being linear between 0 and 80 pg of calcium. IN 1957 Bayerl synthesised the Schiff's base glyoxal bis(2-hydroxy anil), (GBHA, I). I Shortly afterwards Goldstein and Stark-Mayer2 found that the compound was a sensitive reagent for the detection of calcium, with which it forms a red complex; it was subsequently used by Goldsteins as an indicator in the quantitative determination of calcium by ethyfene- diamine tetra-acetic acid.The reagent has been little used to determine calcium in natural materials, apart from the determination of the calcium content of rain and sea water: of soil extracts6 and of serum and plasma.6 Other w ~ r k e r s ~ ~ ~ have studied the reaction conditions with solutions of pure calcium salts. Essentially the calcium - GBHA complex is extracted from an alkaline solution with chloroform, and the chloroform solution clarified by spinning it in a centrifuge. Umland and Meckenstock' found that the addition of 10 to 15 per cent. of methanol clarified the chloroform solution and stabilised the colour, but they also found that the complex, with added methanol, was incompletely extracted, recovery being 15 per cent.Addition of primary alcohols of increasing chain length removed an increasing amount of the complex from the aqueous phase; 99 per cent. is extracted by hexanol - chloroform. In the absence of alcohol, as in Williams and Wilson's method,* the complex is stable for only 16 minutes after the addition of sodium hydroxide and sodium carbonate to the mixture. In determining calcium in silicate minerals, there is the possibility of interference from magnesium and iron. The colour of the calcium - GBHA complex is developed within the pH range 12 to 12.6, under which conditions a significant amount of calcium is co-precipitated with magnesium and iron.At this pH aluminium exists as aluminate and consequently does not interfere. In the gravimetric determination of small amounts of calcium in magnesite and fused magnesia, Hazel and Eglofs found that mannitol prevented the co-precipitation of calcium with magnesium hydroxide. We determine calcium by adding GBHA to a cooled solution of the mineral digest containing mannitol, and sufficient sodium hydroxide to give a pH of 12.6. The calcium - GBHA complex is extracted at once with a mixture of equal volumes of chloroform and isopentyl alcohol, the latter being as effective as hexan01.~ The extract is spun in a centrifuge and the optical density of the solution measured at 531 nm against a reagent blank.The wavelength of maximum optical density depends on the organic solvents present and the 0 SAC and the authors.40 KING AND PRUDEN : THE COLORIMETRIC [Analyst, Vol. 94 relative proportions of the other reagents. Other workers report different spectral maxima: Peaslee,S Kuczerpas and Williams and Wilson: 535 nm; and K e ~ x , ~ 520 nm. Fig. 1 shows the visible absorption spectrum of the calcium complex in chloroform - isopentyl alcohol. A linear calibration is obtained with calcium standards from 0 to 80 pg of calcium (Fig. 2). 01 I I I \ I 400 450 500 550 600 Wavelength, nm Fig. 1. Absorption spectrum of the 40 pg calcium - GBHA complex in chloroform - isopentyl alcohol Calcium, pg Fig. 2. Calibration for calcium in the presence of 3000 pg of magnesium and 2000 pg of iron, with (0-0-0) and without (0-0-0) added mannitol EXPERIMENTAL PREPARATION OF MINERAL DIGEST- The solution is prepared by a method similar to that of Meyrowitzlo Weigh accurately about 50 mg of silicate rock powder into a 30-ml platinum crucible and moisten the sample with a few drops of water.Add 1 ml of concentrated hydrofluoric acid and 2 ml of 18 N sulphuric acid, and digest the mixture on a sand-bath until fumes of sulphur trioxide first appear. Cool the crucible, add 5 ml of water, warm to ensure complete dissolution, cool and transfer the solution quantitatively to a 100-ml calibrated flask. Add 9 ml of 9 N sulphuric acid to the flask, mix well and dilute to 100 ml to give a solution 1-2 N with respect to sulphuric acid. REAGENTS- Acid blank soZzction4ulphuric acid, 1.2 N.Calcium standard solzction-Dissolve 1 g of pure calcium carbonate in a small excess of hydrochloric acid and dilute to 1 litre with water to give a solution that contains 4-00 pg of calcium per ml. Dilute aliquots of this solution to give working standards containing 1Opg of calcium per ml. Sodium hydroxide solution, 30 per cent. w/v in water-A method for preparing calcium-free sodium hydroxide is given by Kuczerpa.lf Glyoxal bis (2-hydroxy and) reagent-Prepared by Bayer 'sl method. Re-sublimed o-amino- phenol and aqueous glyoxal are condensed at 80" C. The product (91 per cent. yield) is re- crystallised from methanol; Williams and Wilson8 report a convenient method for subliming o-aminophenol. The reagent is used at a concentration of 0-5 per cent.w/v in ethanol. Extracting solvent-A mixture of equal volumes of chloroform and isopentyl alcohol. MannitoZ solution, 10 per cent. w/v, aqueozcs. DETERMINATION OF CALCIUM- From the 1-2 N sulphuric acid digest of the mineral, transfer by pipette aliquots, con- taining not more than 80pg of calcium, into 126-ml cylindrical separating funnels. Add sufficient 1.2 N sulphuric acid to give a volume of 20 ml, then add 10 ml of mannitol solution, mix thoroughly and allow to stand for 30 minutes. Add 5 ml of 30 per cent. w/v sodium hydroxide solution, mix and allow to cool to room temperature. To each funnel in turn add 1 ml of GBHA reagent, mix the contents thoroughly and immediately add 20 ml of chloroform - isopentyl alcohol mixture. Shake it for 30 seconds.Collect 10 to 15 ml of the lower phase containing the calcium - GBHA complex, after discarding the first small portion,January, 19691 DETERMINATION OF CALCIUM IN SILICATE MINERALS 41 and clarify the solution by centrifuging briefly at 4000 r.p.m. Measure the optical density of the clear solution, without undue delay, in l-cm cells at 531 nm. The reagent blank is prepared as above, with 20 ml of 1-2 N sulphuric acid in place of the test solution. Determine the calcium in the aliquot by reference to a calibration graph of calcium standards, in 1.2 N sulphuric acid, from 0 to 80 pg of calcium. DISCUSSION Earlier workers have shown that glyoxal bis(2-hydroxy a d ) is a reagent sufficiently sensitive to determine ultramicro amounts of calcium. The order in which the alkali and the GBHA are added varies with different workers.In our method it is essential to add the sodium hydroxide first and cool the solution, allowing enough time for the heat of neutralisation of the digest to disperse, before adding the GBHA. Goldstein and Stark-Mayer2 found that barium and strontium form coloured complexes with GBHA, but these can be destroyed by adding sodium carbonate to the test solution. Umland and Meckenstock? studied the cations that interfere in the determination of calcium, and showed that interfering ions were inhibited by sodium carbonate and potassium cyanide in the presence of sodium potassium tartrate. It is unlikely that silicate minerals would contain significant amounts of most of the metals cited. The addition of sodium carbonate to the sodium hydroxide to destroy GBHA complexes of barium, strontium and other trace metal^^,^,^ is unnecessary, so also is the need to buffer the The pH of the reaction mixture is easily reproducible, and small variations in the strength of the acid digest, resulting from small losses of sulphur trioxide during fuming, are too small to affect significantly the final pH.The stability of the calcium - GBHA complex in chloroform is increased by the presence of alcohol. Without alcohol it is necessary to read the colour of the complex 10 or 15 minutes after adding the alkali to the calcium - GBHA soluti~n.~,~ Ken? found that in ethanol- butanol the colour is stable for 30 minutes after development. Kuczerpa,a by first keeping the reaction solution in 1-2 N sodium hydroxide at 0" C for 10 to 60 minutes, increased the stable period of the chloroform solution of the complex to 1 hour.We recommend measuring the colour without undue delay after centrifuging the chloroform - isopentyl alcohol solution, although the colour is stable for at least 30 minutes after adding the reagent to the alkaline solution. The colour at the surface begins to fade after keeping the solution in a spectro- photometer cell for 1 hour, so that it is necessary to limit the number of determinations in a batch to pennit the procedure to be completed within about 30 minutes. TABLE I COMPARISON OF THE PERCENTAGE OF CALCIUM OXIDE FOUND IN STANDARD SILICATE ROCK SAMPLES BY THE GBHA METHOD, AND BY OTHER METHODS Calcium oxide, per cent.References GBHA Conventional Rapid and rapid Sample method method1* method1* methods) L I 1 (conventional 16 ;i; 16 Granite ((3-1) . . .. .. 1.32 - Diabase (W-1) . . .. . . 11-19 - 10.92 Granite (G-2) . . .. .. 1.91 1-96 Granodiorite (GSP-1) . . .. 1.97 2-06 Basalt (BCR-1) . . .. .. 6.64 6.9 1 Dunite (DTS-1). . .. . . 0.09 0.02 0.07 Peridotite (PCC-1) . . .. 0.39 0.42 0.47 Granite (GA) . . .. .. 2.69 - 2.48 Granite (GH) . . . . . . 0.74 - 0.68 17 Basalt (BR) .. .. . . 14.02 - 13.89 Andesite (AGV-1) . . . . 4.67 4.90 4.8 The precision of the colorimetric method has been tested by determining the standard deviation of the calcium contents of six replicate aliquots of the acid digest from three standard minerals containing widely differing amounts of calcium. The values found were: granite (G-1), mean calcium oxide, 1.34 per cent.; standard deviation, 0.02: basalt (BCR-I), mean calcium oxide, 6.56 per cent.; standard deviation, 0.08: diabase (W-1), mean calcium oxide, 11.19 per cent.; standard deviation, 0.09.Loss of unprotected calcium by co-precipitation with magnesium and iron is great (Fig. 2). Addition of mannitol to the acid digest before adding alkali prevents the co-precipitation42 KING AND PRUDEN of 50 pg of calcium when 3000 pg of magnesium and 2000 pg of iron are both present. The same amount of protected calcium is unaffected in the presence of 6 0 0 0 ~ of magnesium alone. The method ensures that calcium can be determined in most silicate minerals, even when the magnesium content is large, e.g., in an olivine (dunite, DTS-1, Table I, 30 per cent.of magnesiumlz), or the iron content is large, e.g., in a mica Piotite, 16 per cent. of iron (G. Pruden, unpublished work)]. When exceptionally large amounts of magnesium or iron are present, the method is sensitive enough to allow a small aliquot of the acid digest to be taken; the amounts of magnesium or iron can then be decreased to values within the above limits. The tolerance of protected calcium to magnesium and iron far exceeds that reported by other worker~,~,~,~,6 who did not add mannitol to the test solution. The proposed method is suitable for determining calcium in all samples containing silicate minerals, for example, soils and sedimentary and other rocks, and the results compare well with those obtained by other methods. Table 1 shows the calcium oxide contents of some standard rock samples. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. REFERENCES Bayer, E., Chem. Bey., 1957, 90, 2326. Goldstein, D., and Stark-Mayer, C., Analytica Chim. Acta, 1968, 19, 437. Goldstein, D., Ibid., 1959, 21, 339. Ken; J. R. W., Analyst, 1960, 85, 867. Peaslee, D. E., Soil Sci., 1964, 97, 248. Kuczerpa, A. V., Andyt. Chem., 1968, 40, 681. Umland, F., and Meckenstock. K.-U., 2. unalyt. Chem., 1960, 176, 96. Williams, K. T., and Wilson, J. R., Analyt. Chem., 1961, 33, 244. Hazel, W. M., and Eglof, W. K., Ind. Engng Chem. Anulyt. Edn, 1946, 18, 759. Meyrowitz, R., Amer. Miner, 1964, 49, 769. Kuczerpa, A. V., Analyt. Chem., 1967, 39, 1197. King, H. G. C., and Pruden, G., Analyst, 1967, 92, 83. Peck, L. C.. “Systematic Analysis of Silicates,” U.S. Geol. Suw. Bull., 1964, 1170. Shapiro, L., and Brannock, W. W., “Rapid Analysis of Silicate, Carbonate and Phosphate Rocks,” Ingamells, C. O., and Suhr, N. H., Geochim. Cosmochim. Acta, 1963,27, 897. Flanagan, F. J., Ibid., 1967,31, 289. Roubault, M., de la Roche, H., and Govindaraju, K., Sciences Tewe, 1966, 11, 108. Ibid., 1962, llP4-A. Received August ZOth, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400039

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, January, 1969, Vol. 94, @. 43-48 43 The Determination of Sulphur Dioxide with Rosaniline Dyes BY H. G. C . KING AND G. PRUDEN (Rothamsted Experimental Station, Harpenden, Herts.) The compositions of commercial rosaniline hydrochloride (magenta) and pararosaniline hydrochloride samples have been re-investigated to improve the colorimetric determination of sulphur dioxide. The dyes, purified in small amounts by paper chromatography, each give a linear calibration up to 80 pg of sulphur dioxide, with the hydrochloric acid - formaldehyde reaction at pH 1.1. Of the two dyes, pararosaniline hydrochloride is preferred because of its smaller reagent blank value. Impurities can be readily removed from pararosaniline base by recrystallisation, and this reagent is, therefore, recommended for routine work.IN seeking methods to improve the colorimetric finish to the determination of sulphur dioxide from sulphur in soils by Bloomfield's vanadium pentoxide method,l we have re-investigated the relative merits of pararosaniline hydrochloride (Ia) and rosaniline hydrochloride (Ib, magenta or basic fuchsin) as colorimetric reagents in the sulphite-hydrochloric acid- formaldehyde reaction. c1- NH2 la (R=H), Pararosaniline hydrochloride I b (R=CH,), Rosaniline hydrochloride (magenta) Acid-bleached magenta, first used by Steigmann2 as a quantitative reagent for bisulphite, was later used by Urone, Boggs and Noyes3 to determine sulphur dioxide in air. The use of magenta was largely discontinued about 6 years after its introduction as a quantitative reagent, because of an alleged lack of purity, usually ascribed to admixed paarosaniline hydro- chloride and to other, unspecified, compounds.It is known that pararosaniline hydrochloride may be formed during the manufacture of magenta, although the cause of erratic results, which some authors attribute to the impurities, has not been shown conclusively to be the result of the use of mixed magenta - pararosaniline hydrochloride reagents. The substitution of pararosaniline hydrochloride for magenta was first made by West and Gaeke? who did not discuss the purity of the new reagent. Later workers,6s6 however, drew attention to the existence of a variety of commercial pararosaniline reagents of varying dye potency. Pate, Lodge and Wartburg6 assayed spectrophotometrically eighteen samples of pararosaniline hydrochloride and basic fuchsin, and accepted only those showing the spectral maximum of the purest pararosaniline then available.More recently Scaringelli, Saltzman and Frey7 made a detailed study of the conditions affecting the determination of sulphur dioxide colorirnetrically, and used a counter-current distribution procedure to purify pararosaniline hydrochloride. We have examined various samples of pararosaniline hydrochloride and magenta by paper chromatography, and have further obtained sufficient of each of the pure dyes by 0 SAC and the authors.4 4 KING AND PRUDEN: THE DETERMINATION OF [AIzalyst, VOl. 94 chromatography on thick paper to compare their usefulness in the colorimetric determination of sulphur dioxide.Standard sulphite solutions were stabilised as the disulphitomercurate ion: [Hg(SO,),]2-, and the colour of the complex with the acid-bleached dyes was deter- mined at pH 1-1 after reaction with formaldehyde. With either dye a linear calibration was found from 0 to 80 pg of sulphur dioxide, but the reagent blank (E = 0.024) for para- rosaniline hydrochloride at the wavelength of maximum optical density (572 nm) was about one tenth of that (E = 0.196) of magenta at the same wavelength. Although none of the magenta samples we examined contained pararosaniline hydro- chloride as an impurity, it is possible that there are such samples, the purification of which would be difficult on a large scale. Also, because of its small reagent blank pararosaniline hydrochloride would be preferred as a reagent for determining sulphur dioxide.However, the preparation of the pure dye on a small scale by paper chromatography is unsuitable for routine work, especially if the dye contains a large proportion of impurity. We find that pararosaniline base (11)- NH2 I I, Pararosaniline base prepared by treating an acid-bleached solution of the hydrochloride with sodium hydroxide, can be substituted with advantage for the parent dye. The base can be made on as large a scale as convenient and, when recrystallised from aqueous methanol, is a reproducibly pure and stable reagent. Variability in the quality of the pararosaniline hydrochloride is over- come by removing impurities when the base is recrystallised. Rosaniline base, although it can be prepared and recrystallised in the same way as pararosaniline base, suffers from the disadvantage of its parent dye in giving the same high blank value in the determination. EXPERIMENTAL PARAROSANILINE HYDROCHLORIDE- After a chromatographic and spectrophotometric examination of various samples, one that contained few impurities (Hopkin & Williams Ltd., Catalogue No.6496.6) was selected for the preparation of small amounts of the pure dye by preparative paper chromatography and for making pararosaniline base. MAGENTA- Samples, some many years old, from various sources were examined. No sample was found to contain pararosaniline hydrochloride as an impurity, although one, labelled “Basic fuchsine,” consisted almost entirely of pararosaniline hydrochloride, with no magenta.In- organic salts were absent from every dye sample, but all of the samples contained more or less organic impurity. Small amounts of the pure dye could readily be isolated by paper chromatography . A sample of magenta, “For the preparation of Schiff’s reagent’’ (British Drug Houses Ltd.), was selected for preparative paper chromatography. PAPER CHROMATOGRAPHY- From the many combinations of solvents tested, the simplest solvent that would separate pararosaniline hydrochloride and magenta, giving appreciable RF values, was 60 per cent. METHODSJanuary, 19691 SULPHUR DIOXIDE WITH ROSANILINE DYES 45 aqueous ethanol. Solutions of pararosaniline hydrochloride and magenta (0.5 per cent. w/v in methanol) were applied to Whatman No. 3 chromatographic paper in the form of bands about 2 cm long, by using a fine glass capillary.The use of thick paper, rather than the thinner Whatman No. 1 and No. 2 grades, helps to keep the bands as compact as possible during the run. At a temperature of 22" C the solvent front travelled 30 cm (descending) in 64 hours, after which time the bands had increased to about three times their starting width. RF values, to the front edge of the bands, were 0.78 for pararosaniline hydrochloride and 0.83 for magenta. Although mixtures of the dyes could be distinguished by their colours, definite separation of the bands was made difficult by their broadening during the run. Each dye contained a violet-coloured mobile impurity, RF value 0-70. Magenta samples also contained a second violet impurity that ran immediately in front of the main dye band.Each dye left at the origin a blue - violet band that could not subsequently be eluted. The mobile bands could all be eluted with methanol. For preparative work the dyes were banded on to several Whatman No. 3 papers, 25 em wide, cut to a point at their lower edge. The main dye bands were collected in beakers placed under the papers during prolonged elution with the chromatographic solvent. SPECTROPHOTOMETRY- Measurements of spectra in the visible region were made in 1-cm cells with a Hilger & Watts Spectrochem Mark 2 spectrophotometer. Purified pararosaniline hydrochloride and magenta were dissolved in water or ethanol and diluted to give a final concentration of 0.5 mg in 100 ml of solvent.Pararosaniline base, in aqueous suspension, was made just acid with hydrochloric acid to regenerate the hydrochloride, then diluted with water to give the same final concentration. A solution of the base in methanol, in which it is sparingly soluble was diluted to give a final concentration of 0.5 mg per 100 ml of methanol. 0 Wavelength, nm Fig. 1. Visible spectra of A, pararosaniline hydrochloride in water ; B, magenta in water; and C, para- rosaniline base in ethanol The wavelength of maximum optical density of pararosaniline hydrochloride was 536 nm in water and M5nm in ethanol. For magenta, the maxima were 540nm in water and 546 to 547 nm in ethanol. Pxarosaniline base, which is insoluble in water, showed a maxi- mum at 541 nm in ethanol. Fig. 1 shows the spectra of the dyes in water and of the base in ethanol.46 KING AND PRUDEN: THE DETERMINATION OF [Afldyst, VOl.94 The violet impurity, RF value 0-70, showed a maximum at 639 nm for water, and at fj49nm for ethanol. The fast-running impurity in magenta had maxima at MOnm for water and 649 nm for ethanol. None of the impurities, in butanol, showed a maximum as high as 660 nm, reported by Scaringelli, Saltzman and Frey.7 PREPARATION OF PARAROSANILINE BASE- One gram of pararosaniline hydrochloride was dissolved in 250 ml of 2.6 N hydrochloric acid. After standing for 2 hours the bleached solution was filtered through a Whatman No. 1 filter-paper. The base was precipitated as pale magenta-coloured shiny plates by adding a slight excess of 2 . 5 ~ sodium hydroxide solution to the filtrate, and the precipitate was collected on a sintered-glass plate, porosity 3; the filtrate was colourless.The crude base was washed thoroughly with water to remove sodium chloride and excess of sodium hydroxide, and was then recrystallised by dissolving in 70 ml of methanol at the boiling-point, adding 300 ml of water at 80" C and allowing the solution to cool at room temperature. The re- crystallised product, 64 per cent. of the weight of starting material, blackened in the range 200" to 205" C and decomposed at 286" C. Scaringelli, Saltzman and Frey7 prepared the base by recrystallisation from water as an intermediate in their subsequent preparation of pararosaniline hydrochloride. We found that recrystallisation from water gave a yield of only 12 per cent.Recrystallisation from aqueous methanol undoubtedly leads to some loss of product, which is retained in the filtrate, but impurities that are soluble in the solvent are completely removed, as shown by a chromatographic check on the hydrochloride prepared from the pure base. COLORIMETRIC DETERMINATION OF SULPHUR DIOXIDE- Bleached 9ararosaniline reagent solutiort-One gram of pararosaniline base is dissolved in 60ml of concentrated hydrochloric acid and the solution diluted to 1 litre with water. We have not had the opportunity to study the stability of the reagent solution over a long period, but have kept a solution unchanged for 6 weeks in the dark; the reagent solution of Scaringelli, Saltzman and Frey' is reported to be stable for at least 9 months. Potassium tetrachZoromercurate solutim-A solution of 27.2 g of mercury( 11) chloride and 14-9g of potassium chloride in water is diluted to 1 litre.Formaldehyde solution, 0.2 per cent.-Five millilitres of 40 per cent. formaldehyde solution are diluted to 1 litre with water. The solution should be freshly prepared. Standard sulphite solution-A solution of 0-4 g of anhydrous sodium sulphite in 500 ml of water is prepared, corresponding to between 300 and 400pg of sulphur dioxide per ml. The concentration of sulphite is determined iodimetrically. One millilitre of 0-1 N iodine solution is equivalent to 3.2033 mg of sulphur dioxide. Immediately after analysis the sulphite solution is stabilised by dilution with potassium tetrachloromercurate solution.Ten millilitres of sulphite solution diluted to 600 ml with the tetrachloromercurate give a solution corresponding to between 6 and 8 pg of sulphur dioxide per ml. Wavelength, nm Fig. 2. Visible spectrum of the sulphur dioxide (60 pg) - pararosaniline hydrochloride - formaldehyde complexJanuary, 19691 SULPHUR DIOXIDE WITH ROSANILINE DYES 47 CALIBRATION- Aliquots of the dilute sulphite solution, corresponding to not more than 80 pg of sulphur dioxide, are transferred to 50-ml graduated flasks, and 10 ml each of bleached pararosaniline reagent and formaldehyde are added. The solutions are diluted to 50 ml (final pH 1-1) and are then allowed to stand for 20 minutes before measuring the optical densities, in 1-cm glass cells, at 672 nm against a reagent blank.Fig. 2 shows the visible absorption spectrum for the developed colour, and Fig. 3 the calibration graph. Sulphur dioxide, mg Fig. 3. Calibration for sulphur dioxide with A, para- rosaniline hydrochloride; and B, magenta, against respective reagent blanks DISCUSSION In the various dye samples we examined, neither the amounts nor the nature of the impurities were such as to give variable results in the colorimetric determination of sulphur dioxide. It is difficult to compare our results directly with those of other workers, who used reagents of varying quality from many sources but, unless objection is made to the higher reagent-blank value of magenta, there seems to be no reason why either of the rosaniline dyes should not be used. Although none of our magenta samples was contaminated with para- rosaniline hydrochloride, we found that one sample, labelled “Magenta,” was pararosaniline hydrochloride, and that one sample of the latter was magenta.* If pararosaniline hydrochloride contains a large amount of impurities, its optical density will be much less than Ei& = 2600 at 536 nm (water), but small amounts of the pure com- pound can be obtained by preparative paper chromatography. Column chromatography on cellulose is unsatisfactory because the dye is adsorbed strongly on the cellulose; this does not happen on paper chromatograms unless the paper is dried thoroughly after a run and the chromatogram is then re-run.For gram-scale amounts of the reagent when only impure samples are available, or as a routine measure, it is preferable to use pararosWne base, which can be prepared in good yield as a pure, crystalline product.The base gives an extremely pale acid-bleached reagent solution and as good a calibration as the pure hydrochloride. The impurities in pararosaniline hydrochloride and magenta can be readily detected by the chromatographic system described above. The small difference in the structures of the dyes makes it more difficult to detect a mixture, but this can be done if the bands of dye applied to the starting line of the chromatogram are kept as narrow as possible. It is useful, but not always necessary, just to dry the chromatogram and re-run it in the same solvent to improve the separation. Much information about possible contamination of magenta with pararosaniline hydro- chloride can be gained by a study of the spectral maxima of mixtures of the dyes.The spectral assay by Pate, Lodge and Wartburg6 showed that it is possible to distinguish between the dyes; the grouped samples having a maximum optical density at 543 nm as para- rosaniline hydrochloride, and those with maximum at 549 nm were grouped as basic fuchsin * Large blank values were found by Bloomfield in the colorimetric finish to his determination of sulphur in soils1 (Dr. C. Bloomfield, personal communication). His “pararosaniline hydrochloride” was, in fact, magenta.48 KING AND PRUDEN (magenta). Pate, Lodge and Wartburg do not state specifically the nature of the solvent in which measurements were made, and we have found no solvent for which the maximum optical density of magenta is as high as 549 nm. Earlier workers failed to point out the importance in spectral assay of the lower optical density of magenta compared with that of an equal concentration of pararosaniline hydro- chloride.Magenta that contains more than 10 per cent. of pararosaniline hydrochloride shows a decrease in its wavelength of maximum optical density, accompanied by an increase in its original optical density. For a spectral assay it is necessary first to prepare small amounts of the pure dyes chromatographically. Table I shows the change in optical density and maximum wavelength of mixtures of the dyes in water. TABLE I CHANGES IN SPECTRAL MAXIMUM AND OPTICAL DENSITY OF MIXTURES OF MAGENTA 06mg IN 1ood OF WATER AND PARAROSANILINE HYDROCHLORIDE AT A TOTAL CONCENTRATION OF Pararosaniline, % 0 10 20 30 40 60 60 70 80 90 100 Emax.(nm) . . 640 640 639 639 638 638 637 637 637 636 636 Magenta, % , . 100 90 80 70 60 60 40 30 20 10 0 0.889 0.909 0.960 0.960 0.984 1.107 1.177 1.180 1.193 1.293 1-303 Variations in the conditions for the colorimetric determination of sulphur dioxide have been studied by Scaringelli, Saltzman and Frey,' who measured the developed colour at one of two pH values, 1.6 (Method A), Amax. = 548nm, giving a large reagent blank value, or at 1.2 (Method B), Amax. = 575 nm, giving a smaller reagent blank, the change from the higher to the lower pH value being made by adding 3 N orthophosphoric acid. Under our conditions, with 100 mg of pararosaniline base in 1 litre of 6 per cent.v/v hydrochloric acid, the pH of the reaction mixture is 1.1 and the wavelength of maximum optical density is 572 nm. We suggest that, to eliminate variations in the determination of sulphur dioxide when impure commercial pararosaniline hydrochloride has been used, recrystallised pararosaniline base should be substituted. A considerable improvement can be made at once to the determination of sulphur in soils (measured as sulphur dioxide) to improve the original calibration of 2 to 25 pg of sulphur, linear in the range 0 to 10pg.l The determination of sulphur dioxide from other sources, e.g., from air, or in foods and drinks, may require modifications to the initial stages of the method to inactivate interfering substances, e.g., heavy metals, nitrogen dioxide and ozone. These modifications are fully discussed by Scaringelli, Saltzman and Frey,' but the final colorimetric determination of sulphur dioxide is unaltered. It has been suggested8 that p-aminoazobenzene is a superior reagent to pararosaniline hydrochloride. We cannot confirm this ; the reagent blank for the 9-aminoazobenzene reagent has twice the optical density of that of magenta, and the calibration is linear only up to 60 pg of sulphur dioxide. REFERENCES 1. 2. 3. 4. 6. 6. 7. 8. Bloomfield, C., Analyst, 1962, 87, 686. Steigmann, A., Analyt. Chem., 1950,22, 492. Urone, P., Boggs, W. E., and Noyes, C. M., Ibid., 1961, 23, 1617. West, P. W., and Gaeke, G. C., Ibid., 1956, 28, 1816. Pate, J. B., Lodge, J. P., and Wartburg, A. F., Ibid., 1962, 34, 1660. Barabas, S., and Kaminski, J., Ibid., 1963, 35, 1702. Scaringelli, F. P., Saltzman, B. E., and Frey, S. A., Ibid., 1967, 39, 1709. Kniseley, S. J., and Throop, L. J., Ibid., 1966, 38, 1270. Received August lst, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400043

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, January, 1969, Vol. 94, 99.49-53 49 The Flamephotometric Determination of Calcium in Solutions of High Sodium Content BY W. E. BLAKE, M. W. R. BRYANT AND A. WATERS (Depavtment of Chemistry, University of Technology, Loughborough) A method is described for the flame-photometric determination of 0 to 60 p.p.m. of calcium in solutions containing 4 per cent. w/w of sodium. Calcium is adsorbed from 100ml of solution on a column of sodium-form Chelex-100 chelating resin ; sodium is selectively eluted with hydrochloric acid at pH 2.4. Adsorbed calcium is eluted with 2 N hydrochloric acid and the resin washed with distilled water. The eluate and washings are diluted to 100m.l and the calcium determined on a Unicam SPQOO flame photometer. Magnesium, zinc, nickel, barium, mercury, manganese, copper and iron, present separately in concentrations of 26 p.p.m.or collectively in concen- trations of 5 p.p.m. each, do not interfere. Aluminium depresses the amount of calcium found. The method has been used to determine calcium in simulated sea water. ATOMIC-ABSORPTION spectrophotometry, flame photometry and titration with EDTA have been used to determine calcium in alkali-metal salts and in sea water. Bentley and Lee1 showed that, in the determination of calcium by atomic-absorption spectrophotometry, the amount of calcium found is dependent on the pH of the sprayed solution. Rocchiccioli and Townshend2 described the interferences produced in the atomic-absorption spectrophoto- metry of calcium by several anions and cations. Koirtyohann3 compared the use of atomic- absorption and flame-emission methods for calcium and concluded that the flame-emission instrument was more flexible in its applications. He gave the detection limit for calcium by this method as 0-1 pg per litre as compared with 2 pg per litre for atomic absorption.Sizonenko, Zolotovitskaya, Fidel’man and Bulgakova4 determined calcium in alkali halides by extraction of calcium 8-hydroxyquinolinate with butyl alcohol prior to flame- photometric determination. Emmott and Law6 determined 2pg per d of calcium in solutions containing 1.5 per cent. w/w of lithium chloride by using a flame photometer with a methanol - butyl alcohol - water mixture as solvent. Breaults separated a mixture of sodium, potassium, magnesium, calcium and strontium on a column of cation-exchange resin, but found by flame-photometric analysis that recovery of the last two metals was 50 and 59 per cent., respectively.Chelating resins have been used increasingly for concentrating complex-forming metals from solution.7~8~9~10 Callahan, Pascuel and Laill and Riley and Taylor12 reported the concentration of trace elements from sea water with Chelex-100. Olsen, Diehl, Collins and Ellestead13 determined traces of calcium in lithium chloride by titration with EDTA after concentration on this resin. van der Reyden and van Lingen14 determined calcium in sodium salts by a similar method and later van der Reyden and Polmanls showed that barium was also adsorbed. Preliminary experiments in this study were directed to finding the effect of sodium and hydrochloric acid on the emission of solutions containing 5 p.p.m.of calcium. Results (Fig. 1) showed that the presence of 100 p.p.m. of sodium led to an error of +0.5 p.p.m. in the calcium found. For solutions containing 50 p.p.m. of calcium these errors are not significant, but for solutions of low calcium content (less than 10 p.p.m.) the sodium content of the final solution should be less than 30 p.p.m. and the acid concentration less than 0.2 M. In 0 . 5 ~ hydrochloric acid the error was -0.5 p.p.m. 0 S.4C and the authors.M) BLAKE, BRYANT AND WATERS : FLAME-PHOTOMETRIC DETERMINATION [AndySt, VOl. 94 Rossetl6 gave the capacity of the sodium form of Chelex-100 as 3.7 milli-equivalents per g. The theoretical amount of resin needed to absorb calcium from 1OOml of solution of concentration 50 p.p.m.is 0-07 g. To ensure complete adsorption, more resin must be used and some will remain in the sodium form. Rosset16 reported that the sodium form of Chelex-100 was hydrolysed by water; Hering" showed that metals were released by 80, 1 Top scale: sodium, p.p.m. Lower scale: acid normality Fig. 1. Flame-photometric determination of calcium : A, sodium added: B, calcium, 6 p.p.m.; C, acid added an iminodiacetic acid resin of similar type to Chelex-100 at pH values characteristic of each metal. It was found in the work reported here that hydrochloric acid at pH 2-4 would selectively elute sodium from a column of Chelex-100 carrying that metal and calcium. For all the calcium to be retained and for sodium contamination to be reduced to a low level, column dimensions and volume of pH 2.4 eluant used must be related to the calcium content of the original solution.This is illustrated in Table I, which shows the calcium and sodium recovered from 100 ml of solutions containing 4 per cent. w/w of sodium and varying amounts of calcium, when varying column sizes and eluant volumes were used. TABLE I AND VOLUME OF PH 2.4 ELUANT VARIATION OF CALCIUM AND SODIUM FOUND WITH DIFFERENT COLUMN SIZES Volume of pH 2.4 eluant Weight of dry sodium- Column form resin, diameter, 0.6 4 g mm 0.4 0.3 0.3 4 4 3 Calcium added, p.p.m. 2.0 4.0 10.0 16.0 20.0 26.0 60.0 6.0 6.0 2.0 4.0 6.0 8.0 200 ml - Ca, Na, p.p.m. p.p.m. 4.0 160 6.6 142 11.0 110 16.2 100 21.0 47 26.6 37 39.0 10 6-2 66 0.4 2 - - 160 ml - Ca , Na, p.p.m.p.p.m. - - 11.0 160 16.2 113 21.0 64 26.6 68 61.6 109 - - - - 2.2 10 2.7 4 3.8 3 2.4 2 120 ml - Ca, Na. p.p.m. p.p.m. - - 3.0 23 4.6 26 6-8 23 8.6 26 Preliminq experiments also showed that for the flame photometer used, although sodium interference was smaller at 422 nm than at 622 nm, less background and instability of the instrument made the latter wavelength preferable.January, 19691 OF CALCIUM IN SOLUTIONS OF HIGH SODIUM CONTENT 51 EXPERIMENTAL The resin used, Chelex-100 (analytical-reagent grade Dowex A-1, wet mesh, range 20 to 50, U.S. standard), was supplied by Bio-Rad laboratories in 1966. CONDITIONING OF RESIN- Stir 20 g of resin with 15 ml of 2 N hydrochloric acid for 6 minutes, decant off the acid and wash the resin twice with 26-ml portions of distilled water. Decant off the water and stir the resin with 16 ml of 2 N sodium hydroxide solution for 5 minutes.Decant off the alkali and wash the resin twice with 26-ml portions of distilled water. Repeat the cycle five times and dry the resin at 100" C. PREPARATION OF COLUMNS- diameter as shown in Table 11. TABLE I1 COLUMN DIMENSIONS, WEIGHT OF RESIN AND VOLUME OF HYDROCHLORIC ACID (PH 2.4) ELUANT Slurry a weighed amount of dry sodium-form resin into a glass column of appropriate FOR SOLUTIONS CONTAINING 4 PER CENT. OF SODIUM AND 0 TO 50 P.P.M. OF CALCIUM Calcium content Weight of dry Column Volume of pH 2.4 of solution, sodium from resin, diameter, eluant, p.p.m. w/w g mm ml 2 to 10 0.3 3 120 10 to 26 0.6 4 200 26 to 60 0.6 4 160 TREATMENT OF SOLUTIONS- Pass 100 ml of sodium - calcium solution (neutral to litmus) through the resin column at a flow-rate of 1 ml per minute.Pass the appropriate volume of hydrochloric acid (pH 2.4) at the same flow-rate and discard the eluate. Elute the column with 5 ml of 2 N hydrochloric acid. Wash the column with 25ml of distilled water. Dilute the eluate and washings to 100 ml and determine calcium by flame photometry at 622 nm on the Unicam SP900, with instrument settings as specified by the suppliers. RESULTS TABLE I11 ANALYSIS OF SOLUTIONS CONTAINING 4 PER CENT. OF SODIUM AND CALCIUM AS SHOWN The results are given in Table 111. Calciumadded,p.p.m.. . 2.0 4.0 6.0 8.0 10.0 15.0 20-0 25.0 40.0 60-0 Calcium found, p.p.m.. . 3.0 4.6 6.8 8.6 11.0 16.2 21.0 26.5 41.0 61.6 3.2 4.9 6.7 8.8 11.0 16.2 21.0 26.6 40.8 60.6 2.7 4.6 6.6 8.6 10.8 16.6 20.6 25.6 40.6 61.0 3.2 4.9 6.3 8.3 10-7 16.0 21.2 25.2 40.3 60.6 Mean .. 3.0 4.7 6.6 8.6 10.9 16.8 20.9 26.7 40.7 50.9 Standarddeviation . . 0-24 0.21 0.22 0.16 0.16 0.68 0.30 0.67 0.32 0.48 STATISTICAL ANALYSIS OF RESULTS- of calcium and 15 to 50 p.p.m. of calcium. As indicated by the standard deviations, the results fall into two groups, 0 to 10 p.p.m. TABLE IV ANALYSIS OF VARIANCE TABLE, 2 TO 10 P.P.M. OF CALCIUM Degrees of Source of variation freedom Sum of squares Mean squares F ratio Caused by regression . . .. 1 152.49026 162.49026 F = 4.16 Deviation from linearity . . 3 0-52676 Replicate measurements . . 16 043250 Total .. .. .. .. 19 163.64950 3,16 0.17658 0.0421 6752 BLAKE , BRYANT AND WATERS : FLAME-PHOTOMETRIC DETERMINATION [AulaZySt, VOL 94 DISCUSSION Table IV shows that, for the determination of 2 to 10 p.p.m.of calcium in solutions containing 4 per cent. w/w of sodium, the F value is significant. This must be caused by the varying amount of sodium retained by the resin. The standard deviations over this range indicate that the precision of the method compares favourably with those previously described for the same determination. Over the range 16 to 50 p.p.m. of calcium, the insignificant F value (Table V) shows that the results can be fitted to a straight line through the origin when “calcium found” is plotted against “calcium added.” TABLE V ANALYSIS OF VARIANCE TABLE, 15 TO 50 P.P.M. OF CALCIUM Degrees of Source of variation freedom Sum of squares Mean squares F ratio Caused by regression .. .. 1 340 1.000 1 3401*0001 F < 1 3,15 Deviation from linearity . . 3 0.2349 0.0783 Replicate measurements . . 16 3.2026 0.2135 Total .. .. ,. .. 19 3404.4376 The effect of the presence of barium, zinc, mercury, manganese, iron, nickel, aluminium and copper on the results of the determination of 6 p.p.m. of calcium in solutions containing 2 per cent. w/w of sodium was studied. Dry sodium-form resin, 0.3 g, *as used in a 3-mm diameter column and 150ml of hydrochloric acid (pH 2.4) used to elute sodium. Iron and aluminium were precipitated on the resin and were not completely removed by 5 ml of 2 N hydrochloric acid. Aluminium seriously depressed the amount of calcium found, but the other metals, when present singly in concentrations of 25 p.p.m., had negligible effect.TABLE VI ANALYSIS OF SOLUTIONS CONTAINING 2 PER CENT. OF SODIUM, 5 P.P.M. OF CALCIUM AND 25 P.P.M. OF ADDED IMPURITY Mn cu Fe A1 6.4 6.6 2.3 6.4 5.6 2.8 f: 6.6 6.1 6.6 5.6; 6.6 6.6 Metal added Zn Ni Ba Calcium found . . 6.1 6.6 6.6 Analysis of solutions containing 2 per cent. of sodium, 5 p.p.m. of calcium and 5 p.p.m. The approximate order of selectivity of Dowex A-1 for common cations in chloride of each of the above metals (except aluminium) showed calcium found as 5.2 p.p.m. media is given’ as- Cu2+ > Fe3+ > Ni2+ > Zn2+ > Fez+ > Mn2+ > Ba2+ > Ca2+ > Na+. This suggests that the metals listed in Table VI will be retained on the resin and will appear in the final solution on which flame-photometric measurements are made.van Schouwenberg and van der Weyl8 described the enhancement of calcium emission by iron. That 25 p.p.m. of iron had no effect on calcium emission in this study must be caused by the precipitation of iron as a basic salt on the resin. DETERMINATION OF CALCIUM IN SEA WATER- Spiegler19 listed the major cations in sea water as sodium 10,661 p.p.m., magnesium 1272 p.p.m., calcium 400 p.p.m. and potassium 380 p.p.m. Before the determination of calcium in simulated sea water, up to 500 p.p.m. of magnesium were found to have negligible effects on the emission of a solution containing 5 p.p.m. of calcium. Solutions were prepared to correspond to sea water that had been diluted 50-fold. A 100-ml volume of the neutral solution was passed at a flow-rate of 1 ml per minute through a 3-xnm column containing 0.3 g of CheIex-100 in the sodium form.The presence of magnesium was shown by a dark band in the resin. In passage of the pH 2.4 eluant, this band moved slowly down the column. It was found that the volume of eluant used (150 to 200 ml) was not critical. Elution was stopped when the dark band was about 2.5 cm from the column base. The metal ions were eluted with 5 ml of 2 N hydrochloric acid, the column washed with 26 ml of distilled water, the eluate and washings diluted to 100 ml and the calcium content measured.water. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. January, 19691 OF CALCIUM IN SOLUTIONS OF HIGH SODIUM CONTENT 53 The composition of solution was calcium 10 p.p.m., sodium 200 p.p.m., potassium 10 p.p.m.and magnesium 25 p.p.m. Calcium found was 10.2, 10.1 and 10.2 p.p.m. CONCLUSIONS The method described for the flame-photometric determination of calcium in solutions of high sodium content, after concentration of the calcium on a column of Chelex-100 resin, compares favourably in simplicity, precision and accuracy with methods described previously. It can be used for the determination of calcium in technical-grade sodium salts or in sea REFERENCES Bentley, E. M., and Lee, G. F., Environ. Sci. Technol., 1967, 1, 721. Rocchiccioli, C., and Townshend, A., Analytica Chim. Acta, 1968, 41, 93. Koirtyohann, S. R., Atomic Absorption Newsletter, 1967, 6, 77. Sizonenko, N. T., Zolotovitskaya, E. S., Fidel’man, B. M., and Bulgakova, A. M., Zb. Analit. Emmott, P., and Law, G., Analyst, 1966, 91, 383. Breault, E. A., J . Ass. Off. Agric. Chem.. 1965, 48, 719. “Dowex A-1 Chelating Resin,” 1964, Dow Chemical Company, Midland, Michigan. Hering, R., 2. Chemie, Lpz., 1965, 5, 402. Pennington, L. D., and Williams, M. B., Ind. Engng Chem., 1959, 51, 769. Schmuckler, G., Talanta, 1963, 10, 745. Callahan, C., Pascuel, J. M., and Lai, M. G., USNRDL-TR-87-10, 1966, US. Naval Radiological Riley, J. P., and Taylor, D., Analytica Chim. Ada, 1968, 40, 479. Olsen, R. O., Diehl, H., Collins, P. F., and Ellestead, R. B., Tulanta, 1961, 7 , 1987. van der Reyden, A. J., and van Lingen, R. L. M., 2. analyt. Chem., 1962, 187, 241. van der Reyden, A. J., and Polman, J., Ibid., 1967, 232, 274. Rosset, R., Bull. Soc. Chim., FY., 1964, 1845. Hering, R., 2. Chemie, Lpz., 1963, 3, 30. van Schouwenberg, J. Ch., and van der Wey, A. D., Analytica Chim. Acta, 1966, 36, 243. Spiegler, K. S., “Salt Water Purification,” John Wiley & Sons Inc., New York and London, 1962, p. 10. Received June 6th, 1968 Khim., 1966, 21, 264. Defense Laboratory, San Francisco, California, 94135.

ISSN:0003-2654    

DOI:10.1039/AN9699400049

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

64 Autalyst, January, 1969, VO~. 94, &5. 54-61 The Absorptiornetric Determination of Silicon in Water Part W.* Determination of Polymeric Silicic Acid BY I. R. MORRISON AND A. L. WILSON (Central Electricity Research Laboratories, Cleeve Road, Leatherhead, Surrey) A method is described for determining polymeric silicic acid in water. The sample is first digested with sodium hydroxide to convert the polymer into “reactive” silicon, which is then determined absorptiometrically. The extent to which certain other forms of “non-reactive” silicon are determined has also been established. In conjunction with methods for “reactive” and total silicon, the method is useful in helping to distinguish between the different forms of “non-reactive” silicon that may occur in samples. At concentrations of about 0.02 p.p.m.of silica, the standard deviation of analytical results was about 0.001 p.p.m. of silica. Ten samples can be analysed in about 3 hours. SILICIC ACID is appreciably soluble in steam at the temperatures and pressures used in modern high-pressure boilers. As the solubility decreases with decreasing temperature and pressure, silicic acid dissolved in the steam may precipitate from solution as the steam passes through a turbine. This phenomenon can lead to the formation of siliceous deposits in turbines, which cause important losses of efficiency. To counteract this potential problem, the con- centration of silicic acid in the steam - water circuit is controlled below specified limits. To this end, the make-up water (used to make good the losses of water and steam from the circuit) must contain very low concentrations of silicon; a typical requirement is that the silicon content should not exceed 0.02 p.p.m.of silica. De-ionisation plant is frequently used to produce make-up water of this quality. The analytical methods normally used for determining silicon in make-up water provide estimates of the “reactive”t rather than the total silicon content. However, several ~ o r k e r s ~ ~ ~ J ~ ~ ~ 6 ~ 6 ~ 7 have reported that “non-reactive” forms of silicon may be present in the make-up water from de-ionisation plant, and that such forms are at least partially converted into “reactive” (and steam-soluble) silicon in the boiler. “Non-reactive” silicon has been detected by the authors in the make-up of water of several power stations of the Central Electricity Generating Board; in this work, we used our previously published methods for “reactive”8 and total silicon.9 Thus, it is desirable that a method for total silicon be avail- able for regular testing of make-up water.The method9 previously reported was intended mainly as a means for preliminary examination of different waters, but a more precise method will be described in Part VII of this series. Several different forms of “non-reactive” silicon (e.g., quartz, clays and polymeric silicic acid) could conceivably occur in the waters of interest, and it is, therefore, desirable to attempt to develop techniques for distinguishing between such forms. Selective methods of this nature would be of value in studying the occurrence of, and techniques of eliminating, “non-reactive” silicon in de-ionisation plant.As a first stage in this work, an attempt has been made to devise a method that would determine only “reactive” silicon and polymeric silicic acid. * For details of earlier parts of this series, see reference list, p. 61. t “Reactive” silicon (mainly monomeric and dimeric silicic acid) is defined in this paper as those forms of silicon which react with ammonium molybdate in 10 minutes to form molybdosilicic acid under the conditions of the method given in Part IV of this series.* 0 SAC and the authors.MORRISON AND WILSON 56 A study of the literature indicated that polymeric silicic acid is relatively easily converted into “reactive” silicon. It seemed, therefore, that a suitable method would be to determine the “reactive” silicon content of samples of water before and after addition of a de-poly- merising agent.Several such methods have been reported, in which sodium hydroxidel0J1 (or the equiva- lent sodium carbonatelo or sodium hydrogen carbonate12) or hydrofluoric acidll *ls are used for the de-polymerisation. The available evidence” sU indicated that sodium hydroxide was less likely than hydrofluoric acid to attack other forms of “non-reactive” silicon. Accordingly, it was decided to investigate the use of sodium hydroxide. To improve the selectivity of the technique, it seemed desirable to add the minimum possible amount of sodium hydroxide to samples. Preliminary tests indicated that a final concentration of 0.01 N sodium hydroxide in the sample gave conveniently rapid and quantitative de-polymerisation of polymeric silicic acid.This alkalinity was, therefore, used in all of the tests. The results of this investigation and the recommended method of analysis are given in this paper. Although the technique proved not to be completely specific for polymeric silicic acid, the method remains of value in studying make-up waters. METHOD APPARATUS- PoZythevte bottles, 8-02 capacity-These bottles are required for the alkali-digestion stage of the procedure, and each should have a loose-fitting polythene lid. Some bottles were found to give systematically high results, especially after having been used for certain raw waters. Therefore, the bottles should be tested, as described below, before being used for the first time.The bottles should also be tested after having been used for samples of raw water or for samples of treated water that are dirty or contain the cations present in the raw water. As experience is gained concerning the tendency of a particular water to contaminate the bottles, the frequency of testing can be adjusted accord- ingly. The initial test on new bottles should suffce, provided that they are used only for andysing high purity waters (q., make-up water, condensate and feed-water). Before using the bottles for the first time, wash them well with water and then use them to carry out blank determinations (as described under Procedure) up to the end of the alkali-digestion stage. Discard the contents of the bottles, wash them well with water and test their freedom from contamination as follows.(i) Carry out a blank determination in each bottle. (ii) Add 100-ml portions of the water used for (i) to each of four 4-02 polythene bottles (see below). To each bottle, add 1 ml of N sulphuric acid, followed immediately by 1 ml of N sodium hydroxide solution. Mix the contents by swirling and continue as described under Procedure, starting with the addition of the ammonium molybdate reagent, and by using the same reagent solutions as for (i). If the optical density obtained with a bottle from (i) does not differ from the mean of the four results from (ii) by more than a significant amount, the bottle is suitable for use. If results from (i) are significantly greater than the mean from (ii), then either the bottles are causing contamination or the water or the sodium hydroxide solution, or both, contain “non-reactive” silicon.If all of the results from (i) are high by the same amount, the latter is likely, and a supply of purer water should be sought and the test repeated. Our experience is that the bottles are the most likely source of high results. I t is unlikely that the sodium hydroxide reagent will contain sufficient “non-reactive” silicon in a form capable of affecting these tests. However, if desired, this possibility can also be checked by heating the reagent solution (in its polythene bottle) in boiling water for 1 hour, and then repeating tests (i) and (ii) above. When the bottles have been shown to be satisfactory, they should be reserved solely for this analysis.PoZythne bottles, 4-02 caeacity-These bottles are required for the absorptiometric deter- mination after the alkali-digestion procedure. It has been found adequate to clean new bottles by washing them well with water. PoZythene bottles for collecting samples-Any convenient size may be used. It has been found adequate to clean new bottles by washing them well with water.66 MORRISON AND WILSON : THE ABSORPTIOMETRIC [Analyst, Vol. 94 REAGENTS- All chemicals should be of analytical-reagent grade whenever possible. Water and all reagent and standard solutions should be stored in polythene bottles. Water-The water used for all purposes should contain as little silicon as possible, and it is important to take dl possible steps to ensure that it contains no %on-reactive” silicon.Distilled water from a Manesty still has been found to be suitable; the boiling chamber of the still was cleaned regularly by removing any scale and precipitate, and washing well with distilled water. Before using any water for blank determinations and for diluting samples, first determine its “reactive” silicon content [Test (a) below], and check that its “non- reactive” silicon content is negligible [Test ( b ) ] . Test (a)-Add 100 ml of the water under test to a 4-oz polythene bottle and add 5 ml of the same water to a second 4-02 bottle. To each bottle add the molybdate and tartaric acid reagents as described under Procedure. Immediately after adding the tartaric acid, add 95 ml of the same water to the second bottle.Continue treating both solutions as de- scribed under Procedure. From the difference in the optical densities of the two solutions, calculate the apparent “reactive” silicon content of the water. This concentration must be multiplied by 1.03 to give the true concentration, but the magnitude of this correction should usually be sufficiently small to allow it to be neglected. The correction allows for the facts (i) that the difference in optical densities is caused by the silicon in 95ml of water rather than 100 ml, as used in preparing the calibration graph, and (ii) that the final volume of the solutions is 2 ml less than those used to prepare the calibration graph. Test (b)-The test is the same as that used for checking the 8-oz polythene bottles. If the mean of the results from the four best bottles is not significantly greater than the mean of the results from part (ii) of the test, the water is considered suitable for use.If a significant difference is found and appears to be caused by “non-reactive” silicon in the water, attempts to prepare purer water should be made. Sodium hydroxide solution, N . Sulphuyic acid, N. Acid@ed molybdate solution-Dissolve 89 g of amrnonium molybdate , (NHJ 6Mo,024..4H20, in about 800 ml of water at room temperature. Add 63 ml of 98 per cent. sulphunc acid cautiously to 100ml of water, with stirring, and allow the mixture to cool. Add the acid to the molybdate solution, cool to room temperature and dilute to 1 litre with water. This solution was found to be adequately stable for at least 6 months.Tartaric acid solution, 28 per cent. w/v-This reagent was found to be stable for at least 6 months. Reducing agent solution-Dissolve 2.4 g of sodium sulphite, Na$O,.’IH,O, and 0-2 g of l-amino-2-naphthol4-sulphonic acid in about 70ml of water. Add 14g of potassium metabisulphite, shake well until dissolved and dilute to 100ml with water. This solution should be freshly prepared each week. Standard silicon solution-Prepare as described in Part I1 of this series.” PROCEDURE- Collection of sameles-Collect samples in polythene bottles. If it is required to determine the concentration of certain forms of %on-reactive” silicon in the sample from the difference in results between this method and that for “reactive” silicon,8 determine the “reactive” silicon content of the sample as soon after sampling as possible (preferably within a few minutes).If the sample is to be analysed by the alkali-digestion technique alone, the analysis should be made within a few hours after sampling. Analysis of sam9Zes-Place 100 ml of the sample in an 8-02 polythene bottle, and then add 1 ml (+O.Ol ml) of N sodium hydroxide solution to the bottle. Cover the bottle with a polythene lid, and place in a boiling water bath for 10 minutes. Cool the bottle to room temperature, add 1 ml (+O-Ol ml) of N sulphuric acid and mix the contents by swirling. Transfer as much as possible of the contents of the bottle to a 4-02 polythene bottle, ignoring the few small drops of solution remaining in the 8-oz bottle. Add 205ml of the acidified molybdate solution and mix.Ten minutes (& 1 minute) later, add 2-5 ml of tartaric acid solution and mix. After a further 5 minutes (2 1 minute), add 2.0 ml of reducing-agent solution and mix. Between 20 and 60 minutes later, measure the optical density of the solution at 810 nm in Pcm cuvettes against distilled water. Let the optical density be As.January, 19691 DETERMINATION OF SILICON IN WATER. PART VI 67 Blank determinations-Analyse 100 ml of water of known “reactive” silicon content in exactly the same way as samples. Let the optical density be AB. CaZcuZation of reszclts-Calculate the optical density, AT, due to “reactive” silicon in the sample after alkali digestion, from the equation AT = As - AB + A,, where Aw is the optical density due to “reactive” silicon in the water used for the blank determination.From AT and the calibration graph, read off the concentration of “reactive” silicon in the sample after alkali digestion. Compensation for turbidity and coloztr in a sample-If it is desired to allow for any turbidity and colour present in the sample, analyse further 100-ml portions of sample and the water used for the blank determination exactly as described above, except that the tartaric acid is added before the molybdate solution. The optical densities of these two solutions should be measured immediately after the sample and blank measurements. Let the optical densities of these special sample and blank determinations be ASB and ACB, respectively. When these corrections are to be used, AT is calculated from the equation AT = A S - AB - (ASB - ACB) Preparation of ca2ibratio.n graph-To a series of 8-02 polythene bottles, add 100,99,98,97, 96 and 95 ml of water.Then add 0, 1, 2, 3, 4 and 5 ml, respectively, of a standard silieon solution containing 10 p.p.m. of silica. Analyse each solution as described under Analysis of samples. Subtract the optical density of the blank solution from that of the other solutions, and plot a graph of corrected optical densities against concentration of silicon in the initial solution. Repeat these determinations until the calibration graph is defined with the required precision. SOURCES OF ERROR- The temperature of samples and blank solutions should not be less than 15” C during the determination, or the formation of the molybdosilicate may be incomplete. The optical density of the reduced molybdosilicate decreases by about 2.4 per cent.for an increase of temperature of 10” C. If samples are likely to contain more than 0.5 p.p.m. of silica, portions should be diluted with water before analysis so that the concentration of silicon in the diluted sample does not exceed 0.5 p.p.m. of silica. The water used for diluting a sample must be from the same homogeneous batch as that used for the blank determination. If I‘ ml of the sample are diluted to 100ml with water, the equations for calculating the results become In these equations, AT represents the optical density due to silicon in the diluted sample; appropriate correction must, therefore, be made for the dilution of the sample when calculating the concentration of silicon in the original sample.This method is intended primarily for the analysis of high purity waters. If less pure waters are analysed and precipitates form in the sample during the alkali-digestion stage, low results may be obtained (see Effect of other substances). If precipitates are observed, results should be treated with caution, and it will often be desirable to re-analyse a sample diluted with sufficient water to prevent appreciable precipitation. RESULTS EFFECT OF SODIUM SULPHATE- After the treatment with sodium hydroxide and the subsequent neutralisation with sulphuric acid, samples contain an appreciable amount of sodium sdphate. Tests showed that this salt decreased the optical density obtained for a given amount of silicon by about 0.7 per cent., as compared with the optical density obtained in the absence of sodium sulphate.This small effect is allowed for by the method of calibration given under Method. RATE OF DISSOLUTION OF “NON-REACTIVE” SILICON- Solutions containing different forms of %on-reactive” silicon were prepared as described below, and then analysed as described under Method, except that the duration of the heating period was varied. The results are given in Table I.68 MORRISON AND WILSON : THE ABSORF’TIOMETRIC [AIzalyst, VOl. 94 Polymeric silicic acid-The pH value of an aqueous solution of sodium silicate (1500 p.p.m. of silica) was adjusted to 8 by adding sulphuric acid, and the solution then set aside in a polythene bottle for 2 months to allow ageing of the polymeric silicic acid.The aged solution was diluted (lo00 times) with distilled water immediately before the tests described above were applied. Clay-A suspension of bentonite clay, described previouslyg (suspension No. 2), was used after storage for about 2 years in a polythene bottle. The total silicon content of this suspension was determined to be 1700 p.p.m. of silica. This suspension was diluted (lo00 times) with distilled water immediately before the tests described above were applied. Portions of the diluted suspension were passed through Millipore filters with pore sizes of 0.45 and 0.1 pm, and the filtrates and filters analysed for their total silicon content. These tests showed that about 96 and 5 per cent. of the total silicon in the suspension passed through the coarser and finer filters, respectively. “Magrtesium silicates”-Unpublished work by the authors has shown that “non-reactive” silicon is formed when solutions of magnesium sulphate and sodium silicate are mixed under alkaline conditions.The mechanism of this effect was not established, but it appeared that at least part of the “non-reactive” silicon was formed by the silicate reacting with, or being adsorbed by, magnesium hydroxide. Such forms of “non-reactive” silicon are referred to as “magnesium silicates” in this paper. The presence of such substances is conceivable in the water emerging from certain types of de-ionisation plant, and they were, therefore, included in the present tests. To prepare the suspensions, a portion of a solution of sodium silicate (4500 p.p.m.of silica) was placed in a beaker, and an equal volume of a solution of magnesium sulphate (1500 p.p.m. of magnesium) was rapidly added to it. After a further 16 seconds, an equal volume of 0.03 N sulphuric acid was added to the mixed solution, which was then stirred thoroughly and set aside for the desired time; 1 ml of this suspension was added to lo00 ml of 0.01 N sodium hydroxide, and 100-ml portions of this solution were immediately placed in polythene bottles, which were then heated and treated further as described under Method. TABLE I RATE OF CONVERSION OF DIFFERENT FORMS OF “NON-REACTIVE” SILICON INTO “REACTIVE” FORMS Extent of conversion* into “reactive” silicon, per cent. Heating I A 1 minutes silicic acid (about 80 seconds old) (about 6 hours old) Clay time, Polymeric “Magnesium silicate” “Magnesium silicate” 6 100 88 10 100 92 16 - 96 20 99 30 32 - 97 40 102 - - - - 70 77 79 - - 86 2 2 2 3 - - * The total silicon contents of the solution analysed were polymeric silicic acid, 1.6 p.p.m.of silica; “magnesium silicate” (80 seconds), 1.6 p.p.m. of silica; “magnesium silicate” (6 hours), 1.6 p.p.m. of silica; and clay, 1.7 p.p.m. of silica. PRECISION- No special experimental design was used to determine the precision of analytical results. However, some of the other tests of the method gave information on the within-batch standard deviation when distilled water containing various concentrations of “reactive” silicon was analysed. These results are summarised in Table 11.Further, several raw and de-ionised waters were analysed in power station laboratories (by using their absorptiometers) and in the authors’ laboratory; estimates of standard deviations from these results are also given in Table 11. These latter results were not derived from a homogeneous set of values, and should, therefore, be regarded with some caution. Finally, during these tests, the same samples were analysed for “reactive” silicon, both with and without the digestion with sodium hydroxide. There was no significant difference (6 per cent. probability level) between the precisions of the results given by the two methods.January, 19691 DETERMINATION OF SILICON IN W.4TER. PART VI TABLE I1 PRECISION OF ANALYTICAL RESULTS Within-batch standard deviation, * p.p.m.of silica 69 r At concentrations between 0 and 0.1 p.p.m. of silica A t about ’ Type of sample 0.5 p.p.m. of silica Distilled water . . .. .. 0-0007 (38) 0.0008 (9) Raw and de-ionised waters . . 0.0009 (8) 0.0041 (16) *The figures in brackets are the degrees of freedom associated with each estimate of The figures quoted refer to the standard deviations of the differences standard deviation. (sample minus blank). To check that the procedure under Method gave a linear calibration graph, 100-ml portions of solutions containing known concentrations (0 to 068 p.p.m. of silica) of “reactive” silicon were analysed in duplicate. The mean results, after subtraction of the blank value, gave a straight line, the equation for which was D = 1-34 C, where D is the optical density and C the concentration of silicon (p.p.m.of silica). The greatest and average deviations of the mean results from this calibration graph were 0402 and 04)008 optical density units, respectively. EFFECT OF OTHER SUBSTANCES- Portions of distilled water (100 ml), containing 0 or 0-5 p.p.m. of silica (as sodium silicate) and other substances, were analysed as described under Method. The results are given in Table 111, which indicates that none of the substances tested is likely to cause any important interference effects when make-up water, feed-water and other similarly pure waters are analysed. TABLE I11 EFFECT OF OTHER SUBSTANCES Deviation from expected result,* p.p.m. of silica silica added,’ Concentration of substance, Substance p.p.m. 0.0 p.p.m.0-6 p.p.rn. each } <O-OOl <0*001 } & Mo(VI), V(V), Mn’+ Cu’+ 1.0 <0.001 <0-001 W(VI), Ti(IV), Sns+ Co’+ Ni’+ 1.0 <0*001 < 0.001 Zna+ 1.0 <0.001 - 0.003 CP+ 1.0 < 0.001 <0*001 AP+ 5.0 < 0.001 0.001 Ala+ 1.0 <0*001 t0.001 Fe*+ 2.0 0.001 -0.019 Fe’+ 0-5 <0*001 -0.004 Fea+ 0.2 <0.001 - 0.003 Fe‘+ 0.1 < 0.001 -0-002 Fe’+ 2.0 t0.001 -0.018 Fea+ 0.6 <O*OOl -0.005 Fe’+ 0.2 <0.001 -0.002 Fea+ 0-1 <0.001 <0.001 Ca’+ 100 <0.001 -0.003 Cas+ 50 0.001 t0.001 Cas+ 10 <0.001 <0-001 Mgs+ 100 t0.001 - 0.006 Mgs+ 60 <0.001 - 0.003 Mga+ 10 < 0.001 <0.001 Detergent t 42 0.044 0-042 Detergent t 10 0.009 0,009 * The 96 per cent. confidence limits on each result are about 0.0016 and 0.002 p.p.m. of silica for added concentrations of 0.0 and 0.6 p.p.m. of silica, respectively.t Alkylaryl sulphonate, detergent powder with no added silicate or phosphate; analysis of the ash showed that the original material contained siliceous material equivalent to 900 p.p.m. of silica.60 MORRISON AND WILSON : THE ABSORPTIOMETRIC [ A d y s t , Vol. 94 When raw and certain treated waters are to be analysed, Table I11 shows that falsely low results may be caused by some of the metallic impurities (chiefly iron and magnesium). White precipitates were observed in the solutions containing 50 or 100 p.p.m. of magnesium after the digestion with alkali. Unlike the precipitates found when certain raw waters were analysed, these white precipitates were easily dispersed by stirring and probably dissolved when the sulphuric acid was added. During the analysis of raw waters, white or brown precipitates sometimes formed when the undiluted samples were heated with sodium hydroxide.In these instances, the silicon found after alkali digestion was always less than the “reactive” silicon content of the original sample, by amounts varying between 1-5 and 26 per cent. It seems probable that this effect is caused by interaction between metallic ions or hydroxides, or both, and silicate, with the result that “non-reactive” silicon remains after the alkali digestion. It was also observed that blank determinations sometimes gave falsely high results when the alkali digestion was carried out in polythene bottles used for undiluted raw waters in the previous batch of analyses.The effect was confined to raw waters that gave precipitates during the alkali digestion. The greatest deviation was equivalent to 0.014 p.p.m. of silica, and successive blank determinations in the same bottle gave progressively smaller deviations. No satisfactory method of cleaning such contamined bottles quickly was found, but the prob- lem should be minimised by diluting raw waters before analysis, as described under Method. DISCUSSION DETERMINATION OF DIFFERENT FORMS OF “NON-REACTIVE” SILICON- The results in Table I show that the alkali-digestion procedure converted polymeric silicic acid quantitatively into a “reactive” form, even when the polymer had been allowed to age for 2 months. As increasing age decreases the rate of de-polymerisation, the method is probably suitable for de-polymerising all forms of polymeric silicic acid likely to be en- countered in de-ionisation plant. Table I also shows that a very small proportion of the silicon content of a clay was converted into “reactive” form by the alkali digestion.Thus, by analysing a sample by the methods for “reactive” silicon (before and after alkali digestion) and total silicon it should be possible to obtain reasonably reliable estimates of the relative proportions of “reactive” silicon, polymeric silicic acid and clay. Unfortunately, this simple approach is complicated by the possible presence of other forms of “non-reactive” silicon of intermediate solubility in sodium hydroxide solutions ; the “magnesium silicates” tested in this work provide one such example.The results obtained do not suggest any alternative simple analytical techniques that would allow differentiation of the various possible fonns of “non-reactive” silicon. Further investigations of the rates of dissolution of other forms of “non-reactive” silicon (e.g., quartz) were not carried out because the results showed that the alkali-digestion technique is not specific for “reactive” silicon and polymeric silicic acid. Thus, the present method, in conjunction with the methods for “reactive” and total silicon, can only be regarded as allowing an empirical classification of different forms of “non-reactive” silicon. Nevertheless, such a classification may be of considerable value in studying the source and behaviour of “non-reactive” silicon in de-ionisation plant.The results in Table I show different rates of conversion for the three forms of “non- reactive” silicon studied. Thus, further information on the nature of “non-reactive” silicon might be obtained by analysing portions of the same sample, with different periods allowed for the alkali digestion. It is possible that “non-reactive” silicon in certain waters is all converted into a “reactive” form during the alkali digestion. For such waters, the present method should facilitate plant control because it allows results to be obtained more precisely, and much more quickly and simply, than the method for total silicon. BIAS- No important sources of bias were detected when relatively pure waters (e.g., make-up water and feed-water) were analysed. Low results were obtained when precipitates formed in the sample during the alkali digestion.This effect appears to be caused by interaction between the silicate and metalJanuary, 19691 DETERMINATION OF SILICON IN WATER. PART VI 61 ions or hydroxides, with resultant formation of “non-reactive” silicon. The magnitude of the error probably depends on the nature and rate of formation of the precipitate, and the only certain way of ensuring negligible error seems to be to prevent the formation of pre- cipitates. This effect was recognised only during the analysis of samples, but suitable dilution of the sample, as now recommended under Method, should usually ensure unimportant bias, although the precision of the analytical result may be worsened. PRECISION- Earlier workeJ4 has shown that the within-batch standard deviation for determination of “reactive” silicon was about 0-0005 to 0.001 p.p.m.of silica. The results in Table I1 show that similar precision was obtained with the present method for standard solutions containing 0 to 0.1 p.p.m. of silica. It appears, therefore, that the alkali-digestion technique does not introduce any large additional sources of random error when relatively pure waters are analysed. The precision obtained is quite satisfactory for the analysis of make-up and feed-water. Table I1 also shows that satisfactory precision was obtained for solutions containing greater concentrations of silicon but relatively small concentrations of metals. The markedly worse precision found when raw waters were analysed is attributed mainly to the formation of precipitates in most of the samples. This probably caused poorer reproducibility because of the effect mentioned in the previous section.As this effect was first observed only during the analysis of the samples (many of which had not been diluted as now recommended under Method), the standard deviation quoted can almost certainly be improved, but there has not yet been any opportunity to test this supposition. The precision of analysing such samples is of interest mainly for investigational purposes; routine analytical surveillance of the performance of a de-ionisation plant would rely principally on the analysis of relatively pure waters, for which adequate precision is readily attainable. This paper is published by permission of the Central Electricity Generating Board. We thank all of our colleagues who enabled us to obtain samples of water and to carry out analyses in several power stations. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Brines, M. E., Proc. Amer. Pwr Conf., 1952, 14, 406. Dick, I. B., Ibid., 1966, 18, 661. Baker, M. M., Proc. 19th A . Wat. Conf., Engrs’ SOC., West. Penn., 1958, 88. Admiraal, J. A,, Trans. Instn Chern. Engrs, 1959,37, 84. Vyhnalek, H. J., and Banks, C. D., Proc. 21st A . Wat. Conf., Engrs’ Soc., West. Penn., 1960, 133. Solt, G. S., Eff. Wat. Treat. J., 1962, 2, 622. Howell, F. W., P w Engng, 1967,71, 62. Webber, H. M., and Wilson, A. L., Analyst, 1964, 89, 632. Momson, I. R., and Wilson, A. L., Ibid., 1963, 88, 446. Iwasaki, I., Tarutani, T., Katsura, K., and Shimojima, H.. Japan Analyst, 1963, 2, 210. Piryutko, M. M., and Shmidt, Ya. A., Izv. Akad. Nauk S.S.S.R., Otdel. Khim. Nauk, 1953, No. 4, Iwasaki, I., Katsura, T., and Tarutani, T., Bull. Chem. SOC. Japan, 1951, 24, 227. Fresenius, W., and Schneider, W., 2. analyt. Chein., 1965, 207, 16. Morrison, I. R., and Wilson, A. L., AnaZyst, 1963, 88, 100. NOTE-References 8, 9 and 14 are to Parts IV, 111 and I1 of this series, respectively. 607. Part V Received February 21st, 1968 of this series appears in Analyst, 1965, 90, 270.

ISSN:0003-2654    

DOI:10.1039/AN9699400054

出版商:RSC    

年代:1969

数据来源: RSC