摘要:

AUGUST, 1969 THE ANALYST Vol. 94, No. I I21 Techniques in Gas Chromatography Part II*. Developments in the van Deemter Rate Theory of Column Performance A Review? BY E. A. WALKER AND J. F. PALFRAMAN (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) A broad outline of developments that have taken place in the van Deemter theory are reviewed. Authors are generally agreed upon the justi- fications for inclusion of a velocity-dependent term expressing resistance to mass transfer in the gas phase. That the eddy diffusion term is independent of gas velocity is strongly challenged by Giddings, who has advanced a theory of “coupled eddy diffusion,” in which he argues that the effect of column-packing geometry on efficiency is modified by the effect of the gas mass-transfer term.Except at very high pressures experimental support for this is a subject of con- troversy and a dichotomy remains to be resolved. ALTHOUGH this series of short reviews was concerned primarily with techniques in gas chromatography, as emphasised by Halaszl the importance of theory to the analyst must not be overlooked. Efficiency and resolution are the principal parameters with which the analyst is concerned. Resolution is governed largely by choice of stationary phase and temperature conditions in considering the ease with which two compounds can be separated, but resolution done is not sufficient if low efficiency results in overlapping peaks. It is not surprising, therefore, that those factors which contribute to peak broadening have continued to attract the attention of theoretical chemists.The rate theory of van Deemter, Zuiderweg and Klinkenbergz still remains the cornerstone of modern gas-chromatographic theory. In its original form it expressed the theoretical plate height, H , in terms of a number of column parameters as follows- .. .. ’ ’ (1) 2 D 8d12 kG u n2Db (1 + H = 2Adp + Yg + - where A = a constant that depends only on the geometry of the particle packing, d, = diameter of the solid support particles, y = a constant that accounts for the effect of tortuosity of the gas flow path, D, = interdiffusional coefficient of the solute in the gas phase, dl = thickness of the liquid film on the solid support, Dl = interdiffusional coefficient of the solute in the liquid phase, k = the partition ratio expressing the equilibrium ratio of the amounts of sample in stationary and gas phases and 21 = the average linear flow-rate of the gas.For convenience equation (1) is usually written in the form- in which A is known as the eddy diffusion term, B the longitudinal diffusion term and C the coefficient of resistance to mass transfer in the liquid phase. A plot of H as a function of the linear gas flow-rate gives the familiar hyperbolic curve from which experimental values * Part I of this series appeared in the Analyst, 1967, 92, 71. t Reprints of this paper will be available shortly. For details see Summaries in advertisement pages. 0 SAC; Crown Copyright Reserved. H = A + BIG + Cii .. .. .. ’ * (2) 609610 WALKER AND PALFRAMAN : TECHNIQUES IN [Analyst, Vol.94 of A , B and C have been derived. Qualitatively this expression was satisfactory, but quanti- tative determination of the values of A , B and C exposed anomalies that led several authors to re-examine the fundamental approach to the theory. Among the first was Jones who considered, in contradiction to the original views of van Deemter, Zuidenveg and Klinkenberg2 that resistance to mass transfer in the gas phase played a significant d e , and subsequent experimental work has supported this view., to 12 The success of Golay's theoretical treatment of open tubular coIumns,13 in which consideration was given to mass transfer through the gas, gives further support for inclusion of an appropriate term. Justification of a gas-transfer term was based on the argument that although gas diffusion occurs about lo4 times faster than liquid diffusion, the distances involved in the motion of solute molecules in the gas phase compared with those in the thin liquid film on the solid support are correspondingly large, with the result that the relaxation times become comparable.Jones1* claimed that his arguments were supported by the fact that a reduction in the liquid- phase loading to less than 10 per cent. decreases the liquid-phase mass-transfer term by a factor of 10. He considered the behaviour of a single solute molecule travelling through the column, residing alternately in gas and liquid phase. Three factors could affect its relative rate of progress: its residence time in either phase could be greater or less than the average; its velocity in the gas phase may differ from the average distribution either by its own diffusion along the stream path; or its presence in a stream whose instantaneous velocity differs from that of the average across the cross-section of the column.Variation in velocity would in turn affect the number of transitions taking place between the two phases. HETP was then considered on a statistical basis, and a correlation factor introduced to account for mutual interaction of longitudinal diffusion and mass transfer in the gas phase. Thus, his gas-phase contributions arose from two sources: diffusion in the moving gas and resistance to transfer from stagnant gas contained in the porous support. His final equation was- C k dfii C2k2 d@ (1 + k)2 Dl (1 + k)2'x H = A + B / i i + L a-+- .... * * (3) C3diii + 2 p ( C 2 C 3 ) ' k . d,d, l + k Dg +- D, where C, and C2 are geometrical constants, C, the gas velocity distribution coefficient and dg is the diffusion path length of the solute molecules in the gas between phase transitions. If C, is put equal to - , the third term in equation (3) is identical with the mass-transfer term in the van Deemter equation and can be written as C# the liquid mass-transfer term; similarly, the fourth term may be written as CgZ, representing a gas mass-transfer term, where C1 and Cg are the liquid and gas mass-transfer coefficients, respectively. The last term in the equation is the statistical correlation term, where p is the correlation coefficient (unfortunately, difficult to evaluate).The Jones equation was used in the work of Dal Nogare and Chiu,* who used it to study the effect of particle support size on efficiency and resolution and on the ratio of free gas void to liquid-phase volumes for both large and small values of k. They found considerable support for the inclusion of gas mass-transfer and velocity distribution terms in the rate equation. Indirect support also was found for the correlation term. Further support was given by the work of Knox and McLaren16 from studies of peak width variation of non-sorbed materials in both packed glass bead and open tubular columns. Their results indicated that the gas mass-transfer process was about 400 times as slow in the packed, as in the tubular columns.Kieselbach3s4 who set out to evaluate an earlier form of Jones' equation, particularly with respect to the relative contributions of the gas and liquid diffusion processes in packed columns, found his work overtaken by events when JoneP pointed out the inherent weak- nesses in this equation. Nevertheless, the results of Kieselbach's carefully detailed investi- gation, in which he attempted to isolate individually the various terms of the equation by choice of experimental conditions, was important support for the inclusion of a gas-diffusion term in an HETP equation for packed columns. He found at the same time evidence for interdependence of the C1 and Cg terms as indi- cated by variation of the latter with liquid concentration. The C1 term was found to be dominant for early peaks, whereas the Cg term played the more significant r6le in the late 8 7r2August, 19691 GAS CHROMATOGRAPHY.PART I1 61 1 peaks. This was corroborated by Perrett and Purnell16~17 also noticed an interdependence of H with temperature under these conditions; they concluded that Cg terms would arise not only from inhomogeneities in flow, but also in solvent distribution.1Q Cg has, however, been found to be independent of the ~ o h t e . ~ s ~ ~ Ford, Loyd and Ayers20 found Cg proportional to both LIB and k, so that Cg increases as k increases, thus confirming predic- tions already made by other~.~s*,l~ Kieselbach21 suggested that a more satisfactory fit for his experimental results might be obtained by appropriate modification of the Golay equation.13 A similar view has also been taken by Bohemen and Purnell,6 who studied the elution of acetone and benzene by both hydrogen and nitrogen from columns of differing particle size by using a rate equation of the form- H = A + B/G + Cg + CIGf/po .... . - (4) where f is the James and Martin22 pressure correction factor. They assumed that for theoretical purposes a packed column could be regarded as a bundle of non-uniform capillaries and that the gas-transfer term could be written in a Golay form thus- .. .. .. * * (5) x [l + 6k + l l k z ] d ~ 24 (1 + k)2 Dg c g = where x is a proportionating constant. Here again, the authors took great care in their experimental approach by considering extra column processes such as the shape of the sample feed band and adsorption by the support, which were not accounted for in their equation. They considered that failure by many authors to obtain concordant results could be explained by failure to consider these details in planning experimental work.Other authors have also attempted to use Golay’s theory of open tubular columns to derive an HETP equation for packed ~olumns.2~s~4 On the other hand, Giddings26 has argued that this theoretical approach contributes a term 10 to 500 times too small and disagrees with the premise that a packed column can be treated as a bunch of capillaries. Jones’ treatment has also been challenged on the grounds that diffusion processes within the individual support particles could account for not more than 10 per cent. of the observed plate height’lg and velocity variation for not more than 2 per cent.1s~19~25 Fig.1. Statistical model of a single-channel packing structure : a, spherical packing; b, cubical packing Jones’ treatment of mass transfer in the gas phase was based on an idealised statistical model of a single-channel packing structure, which Giddings regarded as too far removed from reality for interpretation of any significant departure from theory. The latter then developed an interpretation based on direct observations made on packing structures. He found26927s28 that in spite of the random nature of the packing, two modes would predominate: a close or spherical packing, as shown in Fig. 1 (a) and an open or cubic packing, Fig. 1 (b). In close spherical packing, the interparticulate channels are short and narrow, while in cubical packing they are long, persistent, and observed to occur at intervals, separated by approxi- mately twice the particle diameter. Such differences would give rise to a variety of localised stream velocities. To construct a theory he, therefore, chose a simple repeating geometrical- pattern model which did in fact bear a resemblance to real packing structures. This consisted of a cylindrical annulus of closely packed particles, through which the gas flow would be612 WALKER AND PALFRAMAN : TECHNIQUES IN [ ~ n a l ~ , s t , 1701.94 restricted, surrounding an internal hollow cylinder where flow would be unrestricted. He then calculated the total contributions to plate height as a sum of contributions from the two regions as- dp2G .... .. - (6) C g = w y - .. U g where w is a constant. The form of this equation can G01ay~~ to express the effect of column inhomogeneities. that of Golay in giving the constant w a definite value, be written- and for a non-porous support- 0.20 w = 0.63 - - be identified with one derived by Giddings' equation differed from which for a porous support could .. .. ' * (7) .. .. .. .. . . (8). He claimed good agreement between experimentally found values of w for porous supports and those found empirically. Perrett and PurnelP in examining this equation found contra- dictory evidence. At high values of k (+m), w should reach a maximum of 0.5, and although Giddings suggested a possible error of a factor of 2 these authors found values in excess of the upper limit of unity.Examination of plots of Cg against (1 + k)-1 for several solutes demonstrated that although the plots were linear as expected they were not in any way concordant nor in agreement with the theoretical plot. They therefore suggested that the flow process, as outlined by Giddings, was only one of the contributory factors to the Cg term. As w was shown to be dependent also on the solvent support ratio, Cg would be expected to do likewise, and they suggested that a more complete description of gas mass transfer might be given by an equation of the form- /3d;2J CgG = [(Om5 - 0.2/( 1 + k ) + f 3 (V,)] - .. .. " (9) Dg where f 3 ( Vl) represented a maldistribution function associated with the liquid phase. Although an additional term to account for gas-phase mass transfer is now accepted, the significance of A , the eddy diffusion term, is less certain.In some of the work so far reviewed A was often ignored on the grounds that its con- tribution to HETP was relatively small. Giddings30 pointed out that reported values, although small, were frequently anomalous. On the grounds that a satisfactory theory of column mechanism must offer an explanation of the deviations found in A , he rejected the classical concept of eddy diffusion as a flow-independent term. In a number of papersJ31 to 37 he de- veloped an alternative approach based on an argument that eddy diffusion arose from irregularities in the carrier-gas stream, which could occur in diffusion of gases through a porous bed. In the interparticulate voids the stream velocity would persist over a distance of approximately d , (the particle diameter), after which a new velocity would be acquired.Secondly, intercommunication by diffusion of molecules could occur between adjacent streams of different local velocity. On this basis, by using random walk theory, he derived an expression for the HETP contribution from eddy diffusion as- .. .. .. 2hdp .. 1 + 4xDg/ii/32ap H = where tip is the length of the stream path step, /3d, the length of a step in the random walk and J is the average gas velocity; h and /I were considered to be of the order of unity and characteristic of the packing. At higher velocities, H would be independent of the actual velocity and this would give the classical van Deemter contribution of 2hd,.At low velocities, H would be dependent on ti. Giddings described the process as coupled eddy diffusion. In the rate equation eddy diffusion could then be written as- .. .. .. . . (11). 1 1 , 1 H = 2hd, In the light of his theory, in 1962 GiddingsB comprehensively reviewed the extensive experimental information available, particularly with regard to the anomalous values of A ,August, 19691 GAS CHROMATOGRAPHY. PART I1 613 which included both ~ero5JO,~2,3~,~,~~ and values, and of the tendency for h to increase with particle size.21,28,44,46,46 Even taking into account inherent uncertainty in results, limitations in the graphical methods used to determine parameters and experimental limitations such as finite sample size, Giddings firmly considered clear evidence remained for his contention that eddy diffusion varied with flow-rate in the manner shown by equation (10).Thus at high flow-rates, i.e., as ii -+ co, A -+ 2hdp, as in classical theory, but as ii -+ 0, A -+ 0. The appearance of negative A values, which were physically meaningless, could be explained by failure to take into account the effects of high pressure drop across the column. In carefully controlled experiments with uncoated glass bead columns he attempted to verify his theoretical arguments. According to classical theory, the plot of H against 1 4 G should give a straight line, intercept 2hdp, whereas according to coupling theory, the line would pass through the origin. Unfortunately, the evidence was not conclusive, but although the intercept was not always zero, it was nevertheless small and much less than dp.This small deviation Giddings attributed mainly to the assumption that the CgG term in the equation could be ignored. In later paper~26,27,~* Giddings extended his coupled eddy diffusion term to that shown in equation (12). 1 I H eddy die. = C .. .. .. . . (12) i 1 , 1 where the value Cgi is the total plate height contribution from channel size variation. Then collecting all the known contributions to plate height variations (total number i) in summation terms, equation (13) was proposed as a realistic equation. + + C. C1iG + C& + Ht, . . .. .. 1 1 H = r , i 1 I + C , l a In this equation the first term replaces the classical eddy diffusion terms. The second term is identical with that of classical theory.The third, representing all known contributions to liquid mass transfer, accounts for such factors as variation of liquid-film thickness and the formation of pools of liquid at the points of contact of the solid support particles. The fourth term is included to account for gas diffusion effects that would not affect the coupling effect, such as diffusion of stagnant gas in the pores of the solid support, and the last term accounts for trans-column effects such as coiling. All these effects could, in general terms, be calculated from independent data, and the equation was used to predict plate height values for a simple glass bead column to within 20 to 50 per cent. of the experimental value. Giddings also s ~ g g e s t e d 2 ~ ~ ~ 7 ~ ~ ~ that the equation should be applicable to liquid chromatography.Giddings’ conception of a coupled eddy diffusion term has been criticised by Perrett and Purnel1,lg who found constant values for eddy diffusion6 and little evidence that, at high gas velocities, plate height became velocity independent.16919 Similarly, Knox and McLaren15 found no experimental evidence of coupling. On the other hand, Harper and H a m m ~ n d ~ ~ obtained experimental results at low gas flows that were more satisfactorily explained by use of the coupling equation than by classical theory. More convincing evidence in support of coupled eddy diffusion comes from investigations at high inlet pressures of the order of 2500 p.s.iSo Under these conditions pressure correction terms become particularly necessary. On the assumption that under these conditions the gas may be treated as ideal, Giddings showed that by classical theory the effective plate height would be given by- .... .. (14) 9 b H = -A + - + ~ g P f + Clpi . . 8 P? whereas from coupling theory the following equation would be obtained- where p1 (inlet pressure) > outlet pressure, b, cg and c1 are parameters related to B, Cg and C1, respectively, and n replaces i in equation (10) to avoid confusion. By using these equations614 WALKER AND PALFRAMAN : TECHNIQUES IN [Analyst, VOl. 94 to calculate the rate of generation of the theoretical plates (N/t), from the classical form, the value of N/t should approach a limiting value, whereas the coupling form leads to a pressure-independent limit. Experimental plots of N/t against $i were consistent with the coupling theory.Among other approaches that have been made to plate height theory, Karamba et uZ.51s52 discussed the kinetic r6le of diffusion and pressure drop in gas - liquid chromatography on the basis that the differential coefficient in the gas phase is inversely proportional to pressure; then, provided flow-rate is maintained, an equation for plate height was derived which encompassed the van Deemter equation as a limiting value. These authors concluded that for high carrier-gas velocities the equation for plate height should be written in the fonn- . . (16). De Clerk, Smuts and Pretorius53 recently made a significant departure from the more customary approach by using non-equilibrium statistical mechanics to derive a general chromatographic model.Although a general differential equation was obtained no experimental evidence was included in the paper. Although the differing views on the theory of packed columns are not yet completely resolved, under normal operating conditions , the simplest modifications of the van Deemter equation give a.n account of the physical factors determining HETP, which is probably sufficient to satisfy the analytical chemist.64 That this is still a n approximation is evident from the extensive work of Giddings, and it is possible that greater use of high pressure chromatography will increase the knowledge and understanding of column processes. We wish to thank the Government Chemist for permission to publish this paper. 1. 2. 3.4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. REFERENCES Halasz, I., Gas Chromat., 1966, 4, 8. van Deemter, J. J., Zuiderweg, F. J., and Klinkenberg. A., Chem. Engng Sci., 1956,5, 271. Kieselbach, R., Analyt. Chem., 1960, 32, 880. -, Ibid., 1961, 33, 23. Bohemen, J., and Purnell, J. H., in Desty, D. H., Editor, “Gas Chromatography 1958,” Butter- -- , J . Chem. SOC., 1961, 360, 2630. Bethka, R. M., and Adam, F. S., AnaZyt. Chem., 1961, 33, 832. Dal Nogare, S., and Chiu, J., Ibid., 1962, 34, 890. Giddings, J. C., and Robinson, R. H., Ibid., 1962, 34, 885. Norem, S. D., Ibid., 1962, 34, 40. Lloyd, R. J., Ayers, B. O., and Karasek, F. W., Ibid., 1960, 32, 618. Giddings, J. C., Stewart, G.H., Seager, S. L., and Stucki, L. R., Ibid., 1960, 32, 867. Golay, M. J. E., in Desty, D. H., Editor, op. cit., pp. 36 and 65. Jones, W. L., Analyt. Chem., 1961, 33, 829. Knox, J. H., and McLaren, L., Ibid., 1963, 35, 449. Perrett, R. H., and Purnell, J. H., Ibid., 1962, 34, 1336. Perrett, R. H., Ibid., 1965, 37, 1346. Fowlis, I. A.. Maggs, R. J., and Young, T. E., Nature, 1964, 201, 606. Perrett, R. H., and Purnell, J. H., Analyt. Chem., 1963, 35, 430. Ford, D. D., Loyd, R. S., and Ayers, B. O., Ibid., 1963, 35, 426 Kieselbach, R., Ibid., 1961, 33, 806. James, A. T., and Martin, A. J. P., Biochem. J., 1952, 50, 679. Desty, D. H., Goldup, A., and Whyman, B. H. F., J . Inst. Petrol., 1959, 45, 287. Giddings, J. C., J. Chromat., 1961, 5, 46. - , Analyt. Chem., 1962, 34, 1186.-, Ibid., 1963, 35, 353. -, Ibid., 1963, 35, 439. -, Ibid., 1963, 35, 1338. Golay, M., in Noebels, H. J., Editor, “Gas Chromatography,” Academic Press, New York, 1961, Giddings, J. C., Nature, 1959, 184, 357. -, J . Chem. Phys., 1959,31, 1462. -, J . Chromat., 1961, 5, 46. -, Nature, 1960, 188, 847. -, AnaZyt. Chem., 1963, 35, 1338. -- , J . Chromat., 1961, 5, 61. worths Scientific Publications, London, 1958, p. 6. p. 5. -, Ibid., 1961, 33, 962. -, Ibid., 1962, 34, 468.August, 19691 GAS CHROMATOGRAPHY. PART I1 615 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. -, Analyt. Chem., 1962, 34, 885. Ayers, B. O., Loyd, R. S., and Ford, D. D., Ibid., 1961, 33, 987. Brennan, D., and Kemball, C., J. Inst. Petroc, 1958, 44, 14. Desty, D. H., and Goldup, A., in Scott, R. P. W., Editor, “Gas Chromatography 1960,” Butter- worths Scientific Publications Ltd., London, 1960, p. 162. Littlewood, A. B.. in Desty, D. H., Editor, op. cit., p. 35. Purnell, J. H., Ann. N.Y. Acad. Sci., 1959, 72, 592. Desty, D. H., Godfrey, F. M., and Harbourn, C. L. A., in Desty, D. H., Editor, Op. cit., p. 200. Geuekauf, E., in Desty, D. H., Editor, “Vapour Phase Chromatography,” Butterworth Scientific Rinjders, G. W. A., in Desty, D. H., Editor, “Gas Chromatography 1958,” 09. cit.. p. 18. Giddings, J. C., Analyt. Chem.. 1963, 35, 2216. Harper, J. M., and Hammond, E. G., Analyt. Chem., 1965,37, 486. Mayers, M. N., and Giddings, J. C.. Ibid., 1965, 37, 1453. Karamba, T., and Ohzeki, K., J. Chromat., 1966, 21, 383. Karamba, T., Ohzeki, K., and Saitoh, K., Ibid., 1967, 27, 33. De Clerk, K., Smuts, T. W., and Pretonus, V., Separation Sci., 1966, 1, 443. Hargrove, G. L., and Sawyer, D. T., Analyt. Chem., 1967, 39, 945. Publications, London; Academic Press, New York, 1967, p. 29. -, J . Chromat., 1964, 13, 301. Received October 30th, 1968 Accepted January 6th, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400609

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

616 Analyst, August, 1969, Vol. 94, pp. 616-624 Ligand-exchange Chromatography on Thin Layers and Columns of Natural and Substituted Celluloses* BY R. A. A. MUZZARELLI, A. FERRERO MARTELLI AND 0. TUBERTINI (Ciamician Chemical Institute, University of Bologna, Via Selmi 2, Bologna 40126, Italy) The rates of adsorption of antimony, cobalt, mercury and silver on cellulose, diethylaminoethyl and p-aminobenzyl celluloses and cellulose phos- phate from diethyl ether have been measured. Columns were prepared with antimony and cobalt pre-treated celluloses and the chromatographic behaviour of dimethylamine, trimethylamine, aniline and ethylenediamine studied. It was shown that these metal-loaded celluloses can be used in ligand-exchange chromatography with organic solvents. Antimony pre-treated celluloses can also be used with aqueous solutions.X-ray diffraction spectra show that the technique proposed is undoubtedly ligand-exchange chromatography and not reversed-phase extraction. LIGAND-EXCHANGE chromatography is based on the principle that a molecule or ion, which is part of a complex fixed on a support, can be released because a different molecule or ion enters to form a more stable complex, or because the complex collapses when the medium is altered. This definition is broader and more comprehensive than the only other one given by other workers,l with which the latest results do not conform.2t08 The interaction of amines with celluloses has been investigated by several workers who have not used chr~matography.~ to l3 However, chromatography of amines on metal-ion pre-treated celluloses has been mentioned in a few instan~es,8p~~p~~pl~ but a thorough investi- gation is still needed.The adsorption of metal ions on cellulose in both organic and aqueous solvents has been studied,17 to 22 but not with amines attached to the polymer. Ligand-exchange chromatography should not be confused with reversed-phase extraction chromatography, which involves cellulose impregnated with large amounts of amine. How- ever, it was concluded17 that, a t low loading of amine on cellulose, the cellulosic support interacts with the metal ions, and in order to confirm this it has been noted that the R, values for amine or tributyl phosphate impregnated silica-gel layers are higher than for impregnated cellulose layers,lsp23 but it is not clear whether the amine molecule is an intermediate in that interaction. Other studies with metal ions on tri-octylamine impregnated layers omitted an unambiguous evaluation of the interaction between the stationary phase and the metal ions.24 A review on reversed-phase extraction chromatography is to be published.25 Argentation chromatography and related techniques are not ligand-exchange chromato- graphy because the complexing agent is merely added to silica or alumina and is not bound.2s This work has been carried out to demonstrate that ligand-exchange chromatography on cellulose can be performed in both organic and aqueous solvents, and to elucidate the r6le played by cellulose in the retention of ligands.X-ray diffraction was preferred as an analytical tool because cellulose has been studied by many workers in this way.27 EXPERIMENTAL SOLUTIONS- solutions without further treatment.Analytical-reagent grade chemicals and solvents were used for the preparation of all Antimony trichloride solutions in diethyl ether, 174 and 700 pg 1-l. Mercury chloride solution i n diethyl ether, 200 pg 1-l. * Paper presented at the Second SAC Conference 1968, Nottingham. 0 SAC and the authors.MUZZARELLI, FERRERO MARTELLI AND TUBERTINI 617 Cobalt nitrate hexahydrate solution in diethyl ether, 240 pg 1-l. Silver nitrate solution in methanol + diethyl ether (20 + 80), 200 pg I-l. Dimethylamine and trimethylamine, 5 mmolar solutions-Prepare by diluting with diethyl ether 1 ml of a 33 per cent. amine solution in ethanol; 2 ml of this solution were added to 48 ml of ether for each experiment.Aniline and ethylenediamine, 5 mmolar solutions in diethyl ether. CELLULOSES- The cellulose fibrous powders used for column chromatography and X-ray diffraction were Whatman C F l l natural (CF), DEll diethylaminoethyl (DE), P11 phosphate (P) and Bio-Rad PAB $-aminobenzyl (PAB). For thin layers, microgranular Schleicher and Schull Selectacel celluloses were used. RADIOISOTOPES AND LABELLED COMPOUNDS- Antimony-1 25, cobalt-60, mercury-203 and silver-1 l0m-These were supplied by the Radiochemical Centre, Amersham, England: 100 pl of the aqueous solutions were diluted to 10ml with 90 per cent. fuming nitric acid, and 100-pl aliquots containing nanogram amounts of the element were used for each experiment.Ethylene-l,2-1QC diamine dihydrochloride, dimethyPC amine hydrochloride, trimethyPC amine hydrochloride and aniline-14C ethanol solutions-These were supplied by the New England Nuclear Corp., Boston, Mass., U.S.A. They were diluted with ethanol and 50-ml aliquots were used for each experiment. L-Tryptophan (methylene-W) and ~~-phenylalanine-3-~~C-These were obtained from the Radiochemical Centre, Amersham, England. RADIOACTIVITY MEASUREMENTS- The solutions were collected in 20-ml plastic bottles; the cellulose powders were filtered on a medium porosity filter and their radioactivity measured with a Laben 512-channel y-ray spectrometer. A factor for geometry correction was determined to compare the radio- activity of the bottles and of the Gooch filter with a standard.Liquid scintillation measure- ments were made with a Mark I Nuclear Chicago system and with a Selo counter; for a 0-4-ml sample, 6 ml of ethyleneglycol monomethyl ether and 10 ml of toluene containing 50 mg 1-1 of 1,4-bis(5-phenyloxazo1-2-yl)benzene and 4 g 1-1 of 2,5-diphenyloxazole were used. Thin layers were scanned by using a Berthold - Desaga thin-layer scanner, equipped with a gas-flow detector connected to a counting rate meter with automatic integrator and chart recorder. X-RAY DIFFRACTION- anti-cathode was used. The spectra were recorded with a Geiger counter. GAS CHROMATOGRAPHY- A Rigaku - Denki diffractometer positioned against an X-ray tube working with a copper A Carlo Erba Fractovap G, with thermal-conductivity detector, was used.THIN-LAYER AND COLUMN CHROMATOGRAPHY- The basic Desaga apparatus was used, including a Brenner - Niederwieser developing chamber operated at 20" C. The layers were 0.25 mm thick. Columns, 15 cm long and 1 cm diameter, were operated under pressure with flow-rates of 1 ml minute-1 with organic solvents, and 0.2 ml minute-l with aqueous solvents, at 20" or 5" C. RESULTS AND DISCUSSION ADSORPTION OF METAL IONS ON MINE PRE-TREATED CELLULOSES- The rates of adsorption of amines on celluloses were measured and found to be very fast; the adsorption of metal ions on celluloses with adsorbed amines was also measured. The results are presented in Table I. The general trend is that adsorbed amines increase the collection of metal ions on celluloses, except for mercury which is not adsorbed.The peculiar behaviour of mercury618 MUZZARELLI et d. : LIGAND-EXCHANGE CHROMATOGRAPHY ON THIN [AutabSt, VOl. 94 TABLE I ADSORPTION OF METAL IONS ON AMINE PRE-TREATED CELLULOSES (pg PER g ) AFTER SHAKING 0.1 g OF CELLULOSE IN 60 ml OF ETHER CONTAINING 12 pg PER 60 ml CF DE r \ I > Untreated EDA TMA DMA Untreated EDA TMA DMA A A Cobalt . . .. 2 15 20 33 3 66 43 90 Silver . . .. 67 83 82 74 90 94 89 94 Antimony .. .. 20 77 70 70 100 100 100 90 Mercury . . .. b.d.1. b.d.1. P PAB A I > r A \ Untreated EDA TMA DMA Untreated EDA TMA DME Cobalt . . .. 5 40 31 57 4 18 26 26 Silver . . .. 70 40 33 37 74 91 85 91 Antimony .. .. 85 100 90 100 27 100 100 80 Mercury . . .. b.d.1. 3 3 3 3 Re-treatment: 1 hour for amines; shaking: 1 hour for Co and Hg; 4.5 hours for Ag and Sb.Each value is the average of a t least three measurements. Recision : f 8 per cent. on ten measurements. b.d.1. : Beyond detection limits. EDA is ethylenediamine, TMA is trimethylamine and DMA is dimethylamine. TABLE I1 Rp VALUES OF CARRIER-FREE RADIOISOTOPES ON A 0.25 mm THICK LAYER OF NATURAL CELLULOSE AT 20" c A B Mercury . . . . 1.00 n.m. Silver . . .. . . 0-37 0.20 Cobalt .. . . 0.32 0.13 Antimony .. .. n.m. n.m. A is natural cellulose; developed with methanol - diethyl ether (20 + 80) containing 40 mg of ammonium thiocyanate per 100ml. B is natural cellulose pre-treated with ethylenediamine - diethyl ether (1 + 1) solution for 1 hour; developed with 6 mmolar ethylene- diamine in diethyl ether. n.m. No movement. t I 10 20 30 20 Fig.1. X-ray diffraction patterns of commer- cial cellulose phosphate : A, alone; B, pre-treated with ethylenediamine; C, pre-treated with silver; and D, pre-treated with ethylenediamine and silverAugust, 19691 619 has already been noted,2*929 and it should be added that ethylenediamine pre-treated celluloses are unable to collect any mercury, although they collect silver, cobalt and antimony under the same conditions. However, a low amount of ethylenediamine is fixed (about 40 pmoles g-l) ; a larger amount of ethylenediamine on cellulose should modify the behaviour of mercury towards adsorption on celluloses. Thin-layer chromatography on celluloses pre-treated with a solution of ethylenediamine in diethyl ether (1 + 1) gives evidence that the migration of mercury and other cations is retarded (Table 11).The presence of amine on cellulose phosphate apparently prevents the collection of silver (Table I). Thin-layer chromatography confirms that silver is less strongly bound to ethylene- diamine pre-treated P cellulose than to the untreated P cellulose; in fact, silver can migrate slightly on the former. This has also been investigated by X-ray diffraction. The spectra in Fig. 1 show that the treatment with ethylenediamine sensibly modifies the spectrum of P cellulose, from which certain peaks disappear and others become more evident; they are sodium dihydrogen orthophosphate peaks. The ethylenediamine and silver pre-treated P cellulose spectrum exhibits lowered orthophosphate group peaks. P cellulose is disturbed by ethylenediamine to such an extent that phosphate is released into diethyl ether (as confirmed by precipitation with molybdate) and the subsequent amount of silver collected is lower, in agreement with results in Table 111.TABLE I11 COLLECTION OF AMINES ON METAL-ION PRE-TREATED CELLULOSES MOLES g-l) AFTER SHAKING FOR 1 HOUR 0-1 g OF CELLULOSE IN 50 ml OF ETHER WITH 12 MOLES PER 50 ml LAYERS AND COLUMNS OF NATURAL AND SUBSTITUTED CELLULOSES CF DE A I h \ I \ Sb % 40 Sb Untreated Co 25 70 25 65 70 65 Dimethylamine . . 29 55 70 60 Ethylenediamine . . completely adsorbed completely adsorbed Aniline . . .. 6 2 0 17 14 9 0 8 % 49 Untreated Co Trimethylamine . . 49 49 P PAB A r A -l I \ Sb Sb Untreated Co 65 75 71 55 44 100 Untreated Co Trimethylamine . . 40 58 Dimethylamine .. 106 55 101 101 Ethylenediamine . . completely adsorbed completely adsorbed -4niline . . .. 9 0 0 0 0 9 13 0 % 59 % 69 Each value is the average of a t least three measurements. Precision: &8 per cent. on ten measurements. Ethylenediamine does not substantially modify the DE cellulose spectrum. Cobalt pre-treated DE cellulose, on the contrary, gives rise to peaks shifted towards smaller angles (about 15' less), thus giving evidence of a modification of the 0,0,2, l,O,i and l,O,l crystal planes (Table IV). TABLE IV X-RAY DIFFRACTION DATA OF METAL-ION PRE-TREATED CELLULOSES : 28 Untreated 22" 40' 16" 24' 14" 50' 22" 37' 20" 03' 16" 27' 14" 47' 29" 05' 23" 45' 22" 26' 20" 00' 16" 40' + Dimethyl- amine + Cobalt 22" 40' 22" 22' 16" 20' 16" 12' 14" 42' 14" 34' 22" 30' 22" 21' 20" 00' 19" 52' 16" 24' 16" 12' 14" 37' 14" 30' 29" 00' 28" 46' 23" 40' 23" 30' 22" 20' - 20" 00' 16" 40' 16" 25' Independent of order.- + Cobalt + Antimony + dimethyl- + &methyl- amine* + Antimony amine 22O 33' 22" 42' 22" 40' 16" 30' 16" 26' 16" 23' 14" 40' 14" 47' 14" 49' 22" 35' 22" 38' 22" 36' 20" 00' 20" 00' 20" 00' 16" 12' 16" 30' 16" 26' 14" 30' 14" 47' 14" 48' 29" 03' 39" 03' 29" 00' 23" 50' 23" 40' 23" 40' - 22" 26' - - 20" 00' - 16" 42' 16" 40' 16" 38'620 MUZZARELLI al. : LIGAND-EXCHANGE CHROMATOGRAPHY ON THIN [ArtalySt, VOl. 94 ADSORPTION OF AMINES ON METAL-ION PRE-TREATED CELLULOSES- The rates of adsorption of metal ions were measured in order to prepare celluloses with a known amount of adsorbed metal ion; results are shown in Figs. 2, 3 and 4 (for cobalt, see reference 14).It can be seen that antimony is adsorbed in a very short time, and that the amount of antimony per gram of cellulose is much larger than the amount of the other ions adsorbed. The metal-ion pre-treated celluloses were used to collect amines, for which they exhibited a greater capacity than untreated celluloses, because the adsorbed metal ion is able to form amine complexes under the conditions indicated. The results are presented in Table 111. M W CL L 5 0 . Y I ? 2w t Time, hours Time, hours Fig. 2. Adsorption rates of antimony from 50 ml Fig. 3. Adsorption rates of antimony from 50 ml of diethyl ether containing 35 pg-atom on four celluloses of diethyl ether containing 8-7 pg-atom on four celluloses /-E " I I 2 Time, hours , Fig.4. Adsorption rates of silver from 50ml of diethyl ether containing 10 pg-atom on four celluloses It will be observed that ethylenediamine is completely adsorbed, i.e., that 1 g of any cellulose collects the 120 pmoles of amine available in solution. The amount of aniline adsorbed, however, is below 15 per cent. of that available, and often no adsorption occurs at all. This means that aniline and presumably other amines are not adsorbed on celluloses from &ethyl ether, and that the metal ions are not in a suitable condition for forming com- plexes. Aniline complexes of metal ions have been reported.30August, 19691 621 Cases of particular interest in Table I11 are the following: 1 g of natural cellulose adsorbs 29 pmoles of dimethylamine, but 1 g of the same cellulose with adsorbed cobalt or antimony collects 55 and 63 pmoles, respectively; DE cellulose adsorbs 25 pmoles g-l of dimethylamine and this increases to 65 pmoles g-1 for cobalt and antimony pre-treated celluloses.Columns in diethyl ether have been prepared to check the correspondence between these values and the chromatographic behaviour of dimethylamine. Fig. 5 shows the elution curves obtained; dimethylamine can be washed from plain cellulose with diethyl ether, but it is necessary to use ethanol for eluting dimethylamine from cobalt pre-treated CF cellulose. When curves obtained with CF cellulose are compared with those obtained with DE cellulose, a retardation is evident because of the more active r81e that DE cellulose plays in adsorption of dimethyl- amine, in agreement with results in Table 111.These results were confirmed by gas chromato- graphy: with a 1-111 column filled with cellulose, trimethylamine passed through first followed by dimethylamine, but they were unable to pass through columns filled with cobalt pre- treated cellulose, even at 70" C. LAYERS AND COLUMNS OF NATURAL AND SUBSTITUTED CELLULOSES 0 c er .- E L Q P Ether E'tlher Ethanol Either Ethanol Volume, mi Fig. 5. Chromatographic behaviour of dimethylamine on (a) natural CF cellulose ; (b) CF pre-treated with cobalt; (c) CF pre-treated with antimony; (d) DE cellulose; ( e ) DE pre-treated with cobalt; and (f) DE pre-treated with antimony. Columns were 16 x 1 cm in ether: columns containing cobalt were pre-washed with ethanol and re-condi- tioned with ether A sharp separation of aniline from trimethylamine and dimethylamine on antimony pre-treated DE cellulose is presented in Fig.6. These amines are also separated from ethylenediamine, which is retained on the column. Another example of the possibilities of the metal-ion pre-treated celluloses is given in Table V. It appears that tryptophan and phenylalanine have higher thin-layer chromato- graphic R, values on antimony pre-treated cellulose than on untreated cellulose. The difference between the R, values is not large enough for a conventional separation to be made and two-dimensional chromatography was required under these condition^.^^ However, after overnight development in the Brenner - Niederwieser chamber, these two amino-acids622 MUZZARELLI &! UZ.: LIGAND-EXCHANGE CHROMATOGRAPHY ON THIN [Astulyst, Vol. 94 Volume, ml Fig. 6. Chromatographic behaviour of A, aniline; B, trimethyl- amine ; and C, dimethylamine on antimony pre-treated diethyl- aminoethyl cellulose DE: column 16 x 1 cm in ether separate on cellulose and on antimony pre-treated cellulose, and the spots travel further on the latter. This may be caused by electrostatic repulsion as well as steric hindrance; it has already been reported that sodium ions prevent albumin from adsorbing on cellulose.32 TABLE V (For solvents and conditions, see reference 30) RF VALUES FOR TRYPTOPHAN AND PHENYLALANINE ON CELLULOSE Cellulose r-A-, Antimony pre-treated reported found cellulose Tryptophan ..Phen ylal anin e .. .. 34 30 .. .. 39 40 39 43 With cold water as the eluent, there is no leakage of antimony from columns filled with CF, DE or P cellulose pre-treated with antimony trichloride in ether and soaked in cold water before preparing the column to avoid clogging consequent to excessive swelling. How- ever, when ammonia solutions were used, some antimony appeared in the effluent of the CF and DE celluloses. PAB cellulose was not tried because it contains a soluble fraction. Antimony pre-treated P cellulose is suitable for ligand-exchange chromatography in ammonia solution up to N concentrations. The chromatographic behaviour of aniline is presented in Fig. 7, to show that antimony pre-treated P cellulose in N ammonia solution effectively retained and concentrated in a narrow band the aniline added in 5 ml of N ammonia solution at 5" C.Untreated P cellulose, however, allowed aniline to pass through after only one free column volume, in a very broad band. s Other amines tried were much more strongly adsorbed than aniline. CONCLUSIONS The adsorption results and X-ray diffraction spectra show that metal ions interact strongly with celldoses even at very low concentrations; the products obtained have a marked capacity for collecting organic ligands. Therefore, metal-ion pre-treated celluloses can be used successfully in ligand-exchange chromatography in organic solvents. Antimony pre-treated celldoses are also suitable for ligand-exchange chromatography in aqueous solutions, especially antimony pre-treated cellulose phosphate in ammonia solutions.This was found when studying the properties of four metal ions only, and the possibility that other ions may be found to be held strongly enough on celldoses to enable chromatography to be carried out on them should not be excluded. In any case, metal-ionAugust, 19691 LAYERS AND COLUMNS OF NATURAL AND SUBSTITUTED CELLULOSES 623 Volume, ml Fig. 7. Chromatographic behaviour of aniline in water at 5” C. Columns were 20 x 1 cm: free column volume, 8ml. The elution was carried out with N ammonia solution on A, a cellulose phos- phate column; and B, an antimony pre- treated cellulose phosphate column pre-treated celluloses seem promising and they might be the subject of further research. With the proper choice of cellulose, metal loading and eluting agent, the analytical applications appear virtually unlimited.Adsorption of antimony is not accompanied by a great modification of the X-ray patterns; there is, however, a slight increase in the angular values to show that antimony made the crystal planes draw nearer. This does not occur with cobalt and silver for the three celluloses considered. If this is found to be typical of antimony in a broader context of study involving several metal ions, it could be correlated to the peculiar behaviour of antimony in chromatography on celluloses. Among the metal ions studied1’ antimony is most strongly held on celluloses. When celluloses are used for ligand-exchange chromatography, adsorption of amines is caused not only by the complexation of the metal ion, but also by the interaction of the amine with the cellulose.This is a difference between celluloses and resins. No relationship between the gram atoms of metal and the moles of amine adsorbed was found. Not only the chelating bi-functional amines, such as ethylenediamine, produce variations in the X-ray spectra but also mono-functional and small molecules, such as dimethylamine. Their interactions become important when their very low concentrations are taken into account. The amine pre-treated celluloses, because of their large capacity, may have applications in the fast removal of metal ions from organic solvents. It appears from this work that, besides the direct adsorption of metal ions on the cellulosic support, amines are intermediates of the interaction of metal ions with cellulose.We thank Professors G. Semerano, A. Breccia and R. Zannetti for advice and continued interest in this work. The financial assistance of the National Research Council of Italy is gratefully acknowledged. REFERENCES 1. 2. 3. Carunchio, V., and Grassini Strazza, G., in Lederer, M., Editor, “Chromatographic Reviews,” Cummings, T., and Korkisch, J., Talanta, 1967, 14, 1185. Helfferich, F., J . Amer. Chem. SOC., 1962, 84, 3242. Volume 8, Elsevier Publishing Company, Amsterdam, London and New York, 1966, p. 286.624 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. MUZZARELLI, FERRERO MARTELLI AND TUBERTINI Walton, H. F., and Latterell, J. J., in West, P. W., MacDonald, A.M. G., and West, T. S., Editors, “Analytical Chemistry,” Elsevier Publishing Company, Amsterdam, London and New York, 1963, p. 356. Latterell, J. J., and Walton, H. F., Analytica Chim. Acta, 1965, 32, 101. Goldstein, G., Analyt. Biochem., 1967, 20, 477. Burtis, C. A., and Goldstein, G., Ibid., 1968, 23, 502. Shimomura, K., Dickson, L., and Walton, H. F., Analytica Chim. Acta, 1967, 37, 102. Creeley, J. J., Segal, L., and Loeb, L., J. Polym. Sci., 1959, 36, 205. Segal, L., and Creeley, J. J., Ibid., 1961, 50, 451. Segal, L., Ibid., 1961, 55, 395. -, Ibid., 1964, 2A, 2951. Hennige, E., Melliand TextBer., 1963, 44, 1228. Muzzarelli, R. A. A., Marcotrigiano, G., Liu, C. S., and FrCche, A., Analyt. Chem., 1967, 39, 1762. Sastri, M. N., Rao, A. P., and Sarma, A. R. K., J. Chromat., 1965, 19, 630. Graham, R. J. T., Bark, L. S., and Tinsley, D. A., Ibid., 1968, 35, 416. Muzzarelli, R. A. A., in Giddings, J.. and Keller, R. A., Editors, “Advances in Chromatography,” Volume 5, Edward Arnold (Publishers) Ltd., London; Marcel Dekker Inc, New York, 1967. Bhatti. A. H., Pakist. J. Scient. Ind. Res., 1965, 8, 252. Ishida, K., and Kuroda, R., AnaZyt. Chem., 1967, 39, 212. Kuroda, R., Kiriyama, T., and Ishida, K., Analytica Chim. Acta, 1968, 40, 305. Ishida, K., Kiriyama, T., and Kuroda, R., Ibid., 1968, 41, 537. Gagliardi, E., and Likussar, W., Mikrochim. Acta, 1965, 765. Bark, L. S., Duncan, G., and Graham, R. J. T., Analyst, 1967, 92, 31. McCormick, D., Graham, R. J. T., and Bark, L. S., an “International Symposium on Chromato- graphy and Electrophoresis,” Presses Academiques Europbennes, Bruxelles, 1968, p. 199. Cerrai, E., and Ghersini, G., to be published. Snyder, R. L., Encycl. Ind. Chem. Anal., 1966, 1, 78. Vainshtein, B. K., “Diffraction of X-Rays by Chain Molecules,” Elsevier Publishing Company, Muzzarelli, R. A. A., Talanta, 1966, 13, 809. Muzzarelli, R. A. A., and Marcotrigiano, G., Ibid., 1967, 14, 305. Ahuja, I. S., Brown, D. H., Nuttall, R. H., and Sharp, D. W. A., J. Chem. Soc., 1966, 938. Bujard, E., J. Chromat., 1966, 21, 19. Davidova, E. G., and Rachinsky, V. V., Prikl. Biokhim. Mikrobiol., 1967, 3, 341. Amsterdam, 1966. Received October loth, 1968 Accepted February 7th, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400616

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, August, 1969, Vol. 94, @. 625-627 The Rapid Determination of Water-immiscible Liquids in Polymer Latices with a Modified Dean and Stark Apparatus BY T. KENNEDY AND R. P. TAUBINGER (Imflerial Chemical Industries Limited, Plastics Division, Teesside Works, Wilton) Imperial Chemical Industries Limited, Plastics Division, Research Department, Welwyn Garden City, Herts.) Modifications made to a conventional Dean and Stark apparatus allow steam heating and permit a determination of water-immiscible liquids in polymer latices to be completed in 10 minutes. The procedure described is particularly suitable for routine operation. A DEAN and Stark distillation, similar to the I.P. method used for the determination of diluent in crank case oil,l has for some years been used in our laboratories as a routine proce- dure for the determination of free styrene or methyl methacrylate during the manufacture of butadiene copolymer latices. This procedure gives reliable and reproducible results, but is time consuming in operation.To obtain results more rapidly the distillation apparatus has been modified and, although used specifically for the determination of liquids lighter than water, it should be equally possible to incorporate these modifications into equipment intended for the determination of liquids heavier than water. Thus although our experience has been limited to the determination of monomers, such as styrene and methyl methacrylate in aqueous copolymer latices (emulsions), this modified procedure should be equally applicable to all Dean and Stark type distillations when results of analyses are required within 10 minutes.The apparatus used (Fig. 1) has two major differences from the conventional equipment : these are a cooling jacket fitted around the calibrated receiver side-arm to allow readings to be taken as soon as all the monomer has distilled over, and the use of steam to supply heat directly into the sample. This latter modification is particularly advantageous in flame- proof areas, and has the added advantage that it speeds up distillation and prevents localised superheating. The modified apparatus reduces the time required per test to 10 minutes and is convenient for multiple unit assembly. APPAFUTU S- The apparatus (Fig. 1) differs from the conventional equipment in the following ways. The calibrated receiver is enclosed in a flow-through water jacket to aid the rapid cooling of the condensate.A tap has been fitted to the lower end of the calibrated arm to facilitate emptying and washing of the receiver without dismantling the apparatus. The length of the connecting arm on the receiver has been reduced to a minimum and the angle has been adjusted to fit the B19 socket of the two-necked flask. The 1-litre distillation flask has been replaced by a 500-ml two-necked flask. The B24 socket is held vertically and is fitted with an adaptor with a B34 cone, as shown. This acts as a funnel to receive the sample and is then connected to the steam supply to deliver steam into the sample at the bottom of the flask. A large-bore tap is fitted to the bottom of the flask to permit draining after completion of the test and when washing out the unit in preparation for the next sample.The B19 socket of the flask is angled and connects to the side-arm of the receiver. An adaptor has been fitted to the top of the condenser, which permits wash water, from a mains tap, to be intro- duced at this point with a swirling motion, so that the whole of the inner surface of the condenser and the rest of the apparatus can be washed out without dismantling. PROCEDURE- be adapted for other applications. METHOD The following procedure is specific for the determination of monomers, but can, of course, 0 SAC and the authors.626 KENNEDY Set up the equipment as water and all taps closed. AND TAUBINGER : RAPID DETERMINATION OF [A?ZdJGt, VOl.94 shown in Fig. 1, with the calibrated arm of the receiver full of 14' Wash water To w a s q Fig. 1. Modified Dean and Stark apparatus Remove the B34 socket and, through the B34 cone, introduce about 1Omg of hydro- quinone to prevent further polymerisation taking place during heating, and a few drops of proprietary silicone anti-foaming agent. Now add 50.0 ml of the latex sample, replace the B34 socket and fasten it to the flask with springs. Make sure water is flowing through the condenser and receiver side-am jacket and turn on the steam. Maintain steady boiling in the flask by controlling the steam supply. A flow-rate of steam producing about 150 ml of condensate in 10 minutes is suitable. After 10 minutes stop the flow of steam, read the volume of monomer distilled over, and hence calculate the monomer content of the latex sample.Open the taps at the bottom of the receiver and flask and run the contents to waste. Through the adaptor at the top of the condenser run in water to wash out the receiver and the flask until they are clean. Close both taps, making sure some water remains in the receiver side-arm. The apparatus is now ready for the next test. RESULTS AND DISCUSSION In initial trials of the modified apparatus and procedure, latex samples free from monomer were used and known amounts of monomer were added immediately before distillation. The results obtained (Table I) show that recoveries of styrene are good, but that recoveriesof methyl methacrylate axe marginally low. These low results are caused by the slight solubility of the methyl methacvlate in water.TABLE I RECOVERY OF STYRENE AND METHYL METHACRYLATE FROM LATICES BY USING THE MODIFIED PROCEDURE Added, ml . . . . 1.0 2.0 3.0 4.0 5-0 6-0 7.0 Found, ml .. . . 1.0 1.9 3.0 4.0 4.9 6.0 7.0 Stpene- Methyl methawyZate- Added, ml .. . . 1.0 2.0 3.0 4.0 5.0 6-0 7.0 Found, ml .. . . 0.9 1.9 2.8 3.9 4.9 5-8 6.8August, 19691 WATER-IMMISCIBLE LIQUIDS IN POLYMER LATICES 627 Comparison of the modified procedure with the conventional Dean and Stark procedure on random latex samples (Table 11) shows that the modified procedure gives slightly higher results than the old procedure. This is thought to be caused by the more even heating in the modified procedure preventing localised overheating, which is suspected to have caused slight polymerisation in the conventional test even in the presence of the inhibitor.TABLE I1 COMPARISON OF THE CONVENTIONAL DEAN AND STARK PROCEDURE WITH THE MODIFIED PROCEDURE ON RANDOMLY SELECTED LATEX SAMPLES Styrene recovered, ml Conventional procedure . . . . 2.0 1-1 0-6 0.2 3.5 2.4 3.0 1.4 Modified procedure . . . . 2.0 1.1 0.7 0.3 3.5 2.5 3.1 1.4 Methyl methacrylate recovered, ml Conventional procedure . . . . 2.8 0.4 nil 4.9 2-6 0.3 4.5 2.7 Modified procedure . , . . 2.9 0.4 nil 5.0 2-7 0-3 4.5 2.9 The rate of distillation of monomer depends on the flow-rate of steam and on the monomer examined, thus methyl methacrylate distils over more rapidly than does styrene. With a flow-rate producing about 150 ml of condensate in 10 minutes it was found that both the monomers had completely distilled over within 8 minutes. A multiple assembly of four units of the modified apparatus is in routine daily use for plant control purposes and satisfactory operation is being maintained, even although the personnel operating the equipment are relatively unskilled. In our experience the modified method can be used to cover any range of concentrations provided the volume of sample taken and the capacity of the receiver are compatible. The only limitation we have found is the difficulty in obtaining homogeneous samples if the free monomer content of the latex is high. We thank Mr. J. M. Lloyd for helping with the experimental work. REFERENCE 1. IP23/64 “IP Standards for Petroleum and its Products, Part I, Methods for Analysing and Testing,” Received August lst, 1968 Accepted January 2 lst, 1969 Twenty-third Edition, The Institute of Petroleum, London.

ISSN:0003-2654    

DOI:10.1039/AN9699400625

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

628 Analyst, August, 1969, Vol. 94, pp. 628-633 The Direct Determination of Free Acrylonitrile in Aqueous Copolymer Latices Part I. Rapid Potentiometric Titration Method BY R. P. TAUBINGER (Imperial Chemical Industries Limited, Plastics Division, Research Department, Welwyn Garden City, Herts.) A rapid semi-automatic procedure for the direct determination of 0-05 to 15 per cent. of acrylonitrile monomer directly in aqueous butadiene - acrylo- nitrile latices is described. The determination is carried out without prior separation of the monomer and is based on the addition reaction of sulphite and acrylonitrile and subsequent potentiometric titration of the liberated alkali to a pre-set pH end-point. Results reproducible to within f2 per cent. relative are obtained on latex samples containing 15 per cent.of monomer, and with samples containing only 0.1 per cent. of free acrylonitrile the reproducibility is still within f 15 per cent. relative. DURING the emulsion polymerisation of batches of butadiene - acrylonitrile copolymer latices in autoclaves, it is necessary to obtain a fairly accurate measure of conversion into polymer at various stages of the process as a means of monitoring the progress of the polymerisation. Conversion can be monitored either by measurement of the increase in solids content of the latex or by measurement of the decrease in free monomer during polymerisation. The procedure used to monitor conversion must be rapid, giving results within 15 minutes, and must, therefore, be directly applicable to the latex samples without any preliminary treatment.A total solids method, successfully used for many years, has the disadvantage that in process control analysis, when results are required rapidly, only small samples that can be dried quickly can be used. This can lead to inaccuracies, particularly at conversions above 60 per cent. Further, in latices where polymerisation has been terminated, unreacted acrylonitrile is recovered by a stripping process and the level of the residual monomer during stripping must be determined. It seemed sensible, therefore, particularly as consideration was being given to the possible automation of analytical procedures, to develop a method for the determination of acrylo- nitrile monomer in latices, which had the required accuracy for the analysis of stripped latices (0 to 1 per cent.), the necessary speed and flexibility for conversion analysis (3 to 15 per cent.) and which might be a useful basis for further automation developments.The procedure reported here fulfils these requirements. EXPERIMENTAL The preferred methods for the determination of acrylonitrile monomer are reaction with an excess of dodecyl mercaptan with subsequent determination of the excess by titration,l p2 v3 and polarography.496 Neither of these techniques can be applied directly to latex samples, as mercaptan has to be used in alcoholic solution which causes coagulation of the latex and additives such as surface-active agents, which are present in latices, interfere in direct polar0 graphy of samples. Critchfield and Johnsons and Terentjev and his co-workers7~8 reported use of the sulphite addition reaction to determine acrylonitrile for assay purposes.CH2=CH-CN + H20 + Na2S0, -+ NaS03CH2-CH2-CN + NaOH. Critchfield and Johnson in their procedure added a known excess of acid with the 2 M sodium sulphite and, after 5 to 30 minutes' reaction time at 25" C, back-titrated the excess of acid with the aid of a visual indicator. In the presence of latex, visual indicator end-points 0 SAC and the author.TAUBINGER 629 are difficult to detect and, under the acidic conditions used, it was found that the latex tended to coagulate. Terentjev and his co-workers used either visual or potentiometric end-point detection in the direct titration of the alkali liberated from the added 0-5 N sodium sulphite, but the dioxan present in their reaction mixture unfortunately caused the latex to coagulate. Preliminary work showed that alkaline 1 .6 ~ sodium sulphite reagent that contained a little added emulsifier reduced the danger of coagulation of the latex, gave reasonably fast reaction and ensured that the reagent gave a positive blank titration. It was found that the liberated alkali could be titrated potentiometrically with hydrochloric acid and that the change in pH at the end-point was about 1.5 units. The end-point consistently occurred at a pH of 10.9, varying by not more than 0.2 unit from this value. I 3 I2 I I - 10 9 - - - - 1 Volume of 0.25~ hydrochloric acid Fig. 1. Typical end-points obtained when the above mixtures were titrated : 0, reagent mixture alone; reagent mixture with added acrylonitrile; m, reagent mixture with added latex Fig.1 shows that the pH of the end-point is not affected by the presence of monomer or latex. It was, therefore, not necessary to plot the full titration graph for every sample and it was decided to use an automatic titrimeter to titrate to the pre-set pH end-point. The instrument available in our laboratories was an E.I.L. Model 24 pH titrimeter, which titrates to pre-set potentiometric end-points with facility for slow titrant addition near the end-point. This instrument proved to be suitable for this application. The reaction of sodium sulphite and acrylonitrile proceeds relatively slowly and a delay of 5 minutes for reaction, before titration is commenced, was found to be necessary.To facilitate this and to simplify operation, an electric timer was fitted to the titrimeter, by- passing the “automatic” switch, so that the operator, after adding the sample to the reagent, simply activated a switch to start the clock, which after 5 minutes’ delay started the titration. When titration was complete the burette was read in the normal way. To simplify this test further, a Smith’s pattern automatic piston pipette was fitted to deliver the reagent mixture, and titrant was delivered from a horizontally mounted self-zeroing free piston burette. A special titration vessel was designed that could be kept permanently connected to the apparatus. This reduced the danger of breaking the combined electrode and also simplified cleaning between samples.Tap water was used to wash out the titration vessel ~ P Z situ. It was feared that variations in the pH of the latex (pH 7 to 11) might cause a change in the pH of the end-point and that the presence of acrylonitrile or latex might affect the blank titration, but neither effect was significant. Neither butadiene nor any of the other additives present in the latices examined interfered in this determination. SAMPLING- Butadiene - acrylonitrile copolymer latices are manufactured in autoclaves at pressures above atmospheric , and considerable amounts of butadiene dissolve in the aqueous phase. When a sample is removed from the autoclave this gas tends to evaporate and cause vigorous froth formation. Addition of small amounts of proprietary silicone anti-foaming agent and gentle agitation were found to reduce foaming in the collecting vessel but, before an aliquot630 TAUBINGER : DIRECT DETERMINATION OF FREE [Analyst, VOl.94 could be removed by pipette, the bulk of the butadiene had to be removed, otherwise the pipette tended to fill with bubbles. The most successful and simple means of removing this dissolved butadiene was to transfer the sample to a measuring cylinder and subject it to vigorous pumping action with a perforated disc, fitted horizontally to the end of a metal rod. This form of agitation removed the dissolved butadiene in 1 to 2 minutes. When a sample is first taken from the autoclave it is necessary to arrest polymerisation. The addition of a small amount of sodium diethyldithiocarbamate does not interfere in this determination but it effectively stops further polymerisation.Aqueous solutions of the carbamate deteriorate on standing, and the solid reagent was not effective because latex tends to coagulate round the crystals, thus preventing the reagent from becoming active. It is, therefore, preferable to dissolve a small amount of the carbamate in 1 to 2 ml of water in the sampling bottle just before sampling. Quinol must not be used to arrest polymerisation as it was found to lower the end-point pH. METHOD The arrangement of the apparatus is shown diagrammatically in Figs. 2 and 3. APPARATUS- A = Reagent reservoir B = Titrant reservoir C = Smith’s pattern automatic pipette of 50-mi capacity D = Automatic, self-zeroing free-piston burette, horizontally mounted, 50 or 100-ml capacity E = “Mini-stirrer,” speed adjusted to give maxi- mum mixing without trapping air bubbles F = Automatic titrator, the E.I.L.Model 24 auto- matic pH titrimeter is suitable; set the fastlslow changeover to 95 mV before the end-point G = Titrator tap-unit; set to give a fast flow-rate of about 60 ml minute-1 and a slow flow- rate of about I ml minute-’ H = Electric delay timer. The circuit diagram of the timer and the method of connecting to the “automatic” switch of the titrator are given in Fig. 3 I = Combined glass - calomel electrode, 9 inches long J = Tap water inlet to titration vessel K = Overflow to waste L = Tap for emptying the titration vessel to waste M = Opening for sample addition N = Tap for filling the free-piston burette Fig.2. Diagram of apparatusAugust, 19691 ACRYLONITRILE I N AQUEOUS COPOLYMER LATICES. PART I 631 I ‘ . . Leads removed from S 33 103 switch and connected to SW4/8 i Fig. 3. Circuit diagram of the timer showing method of connecting the E.I.L. Model 24 titrator. Switch “C” of timer is changed over all the time the clutch is energised. Switch “2” changes over a t selected period. Switch “M” opens after this by a time equal to one dial division. This is to prevent mechanism damage. Details are given in the Appendix REAGENTS- Hydrochloric acid, 0.25 N-Standard titrant solution. Sodium sulphite reagent, 1.6 M-Dissolve 100 g of anhydrous sodium sulphite (this dis- solves more readily than the hydrate) in water containing 5-0 ml of Teepol and 25-0 ml of 1.0 N sodium hydroxide, dilute to 1 litre and mix thoroughly.Bufer solution, pH 9.2-This can be prepared from commercially available tablets or powders. PROCEDURE- Standardisation-As the E.I.L. Model 24 pH titrimeter only covers the pH range 3 to 11 and the end-point occurs at a pH of 10.9, it is necessary to offset the dial on the meter by two units. By using pH 9.2 buffer solution set the meter to read 7.2. This extends the range of the titrimeter to cover the pH range 5 to 13. To ascertain the exact pH of the end-point, manually titrate 50ml of the reagent by adding 0 6 m l increments of titrant and plotting the graph of pH meter reading against volume of titrant added. Locate the point of inflection (end-point) of the graph and note this pH reading. Set the titrimeter to stop the titration at this reading.The hydrochloric acid is standardised by conventional procedures, and 1 ml of 0.25 N hydrochloric acid = 0.01325 g of acrylonitrile. Alternatively, a standard solution of pure monomer can be prepared (3 to 5 per cent. w/v) and a 10-ml aliquot of it can be used for standardisation under the conditions of this test. Determirting monomer in striflped latices (0 to 1 per cent.)-To the clean titration vessel add one delivery (50 ml) of sulphite reagent, start the stirrer and fill the burette. Set the function switch of the titrimeter to “mV falling” and add, by pipette, 50ml of sample.632 TAUBINGER : DIRECT DETERMINATION OF FREE [Analyst, Vol. 94 Start the timer. After 5 minutes the timer will start the titration and when the titration is complete, note the volume of titrant delivered.Carry out a blank titration omitting the sample and correct the sample titre by subtracting the blank. Hence calculate the free acrylonitrile content of the latex sample. Determining free monomer in latex during Polymerisations (3 to 15 eer certt.)-To a clean, dry 16-oz bottle add 0-5 g of sodium diethyldithiocarbamate (to short-stop polymerisation), a few drops of proprietary silicone anti-foaming agent and 2 ml of water. Collect 100 to 200 ml of latex sample; stirring and gentle swirling helps to reduce frothing and hence speeds up collection of the sample. De-gas the sample by using a perforated disc plunger. Transfer 10 ml of this de-gassed sample solution to the reaction vessel containing one delivery (50 ml) of sulphite reagent and proceed as described above.Cleaning the apparatus-The titration vessel is washed with mains tap water with the drainage tap open. When it is clean, the water is stopped, the vessel allowed to empty and the tap is then closed. RESULTS Aqueous solutions of acrylonitrile were examined by this procedure to check on the recovery of the monomer. Table I gives the results obtained on two separate solutions, analysed at various times over a period of 1 month. These results show a mean recovery of 100.0 per cent., with a standard deviation of 0-26 per cent. TABLE I RECOVERIES OF ADDED ACRYLONITRILE Solution Acrylonitrile added, g Acrylonitrile found, g Recovery, per cent. 1 2 0.4982 0.4982 0.4982 0.5347 0,5347 0.5347 0.5347 0.5347 0.5347 0.8020 1.0694 1.3367 0.499 0.496 0.498 0.535 0-534 0.535 0-534 0.534 0-537 0-804 1.075 1.342 100.2 99.6 100.0 100.0 99.9 100.0 99.9 99.9 100.4 100.3 100.5 100.4 just before testing autoclave samples Addition of known amounts of acrvlonitrile to monomer-free latex gave similar recoveries.The accuracy"of recoveries of monomer from was more difficult to establish, as the free monomer content of latices tends to fall ofl- on standing. This is presumably caused by volatilisation, oxidation , polymerisation or hydrolysis. Typical results showing this decrease are given in Table 11. TABLE I1 DECREASE IN FREE ACRYLONITRILE IN SAMPLES ON STANDING Free acrylonitrile, as determined, per cent. Sample day 1 day 2 day 4 7/1610 8-11 7-87 6.59 13/1910 3.77 3-49 3.20 19/2210 1-80 1.63 1-57 37/0710 0-17 0.14 0-13 t h \ 1/1310 14-09 13.20 9.49August, 19691 ACRYLONITRILE IN AQUEOUS COPOLYMER LATICES.PART I TABLE I11 REPEATABILITY OF THE RECOMMENDED PROCEDURE 633 Aliquot analysed, Sample ml 1 10 50 2 10 50 3 10 50 4 10 50 6 10 10 26 25 Titre, ml 11.92 47.16 16.65 70.76 11.28 43.8 1 7-00 23-15 3.34 3.42 4-49 4-40 Corrected titre, ml 8.82 44.05 13-56 67-66 8-18 40-71 3.90 20.06 0.39 0.47 1.54 1-48 Free acrylonitrile, per cent. w/v 1-21 1-21 1-86 1.86 1.12 1.12 0.64 0.65 0.054 0.066 0.086 0.080 The repeatability of the procedure when applied to latex samples taken from autoclaves can be judged from the results listed in Table 111. CONCLUSION The procedure described here is suitable for the determination of free acrylonitrile in latices during polymerisations, in the range 3 to 15 per cent., and also for the determination of residual monomer in stripped latices (0.05 to 1 per cent.).This procedure has been used in our plant laboratories with considerable success for over 2 years. It is simple to use and only sample addition and the final reading of the burette are critical, all the other operations are of the simple push-button or tap-operation type. Relatively sophisticated equipment has been used in this method because it reduces operator time in routine testing, but equally good results can be obtained by the use of more conventional apparatus, such as pH meter, beakers, pipettes and burettes. In these instances, however, it would be advisable to plot the full pH verws volume of titrant graph, as conditions would not be kept as constant as in the semi-automatic equipment described here. Appendix LIST OF COMPONENTS = Radio spares panel, neon red = Plessey plug, MK 4, 12-way, type CZ 49112 = Plessey plug, MK 4, 12-way, type CZ 49458 = Omron relay, type MK 2P, 250 V a.c.. 2 off = Plessey socket, MK 4, 12-way, type CZ 49126 = Plessey socket, MK 4, 12-way, type CZ 49469 = NSF switch, DPDT 3 A, 260 V a.c. = Radio spares switch type min push centre off = Painton rotary switch, type P.W., code AS/4P/2/2B = D. Robinson “Rodene” timer, 8000 F, 12 minutes Ll PL, SKT, SKT, SWl SW,, SW, 3 A sw, TMl REFERENCES 1. Philipp, B., and Bartels, U., Ada Chim. Hung., 1962, 32, 19. 2. Haslam, J., and Newlands, G., Analyst, 1955, 80, 50. 3. Beesing, D. W., Tyler, W. P., Kurtz, D. M., and Harrison, S. A., Analyt. Chem., 1949, 21, 1073. 4. Crompton, T. R., and Buckley, D., Analyst, 1965, 90, 76. 6. Claver, G. C., and Murphy, M. E., Analyt. Chem., 1959, 31, 1682. 6. Critchfield, F. E., and Johnson, J. B., Ibid., 1956, 28, 73. 7. Terentjev, A. P., and Obtemppsaja, S. I., Zh. Analit. Khim., 1956, 11, 638. 8. Terentjev, A. P., Obtemppsaja, S. I., and Buslanova, M. M., Zav. Lab., 1968, 24, 167. Received August 1st. 1968 Accepted February loth, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400628

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

634 A~zalyst, August, 1969, Vol. 94, @$. 634-642 The Direct Determination of Free Acrylonitrile in Aqueous Copolymer Latices Part 11. Use of a Continuous-flow Enthalpimetric Analyser Designed for the Measurement of Macro Amounts BY R. P. TAUBINGER (Imperial Chemical Industries Limited, Plastics Division, Research Department, Welwyn Garden City, Herts.) The continuous-flow enthalpimetric analyser described is particularly suitable for measurement of the large temperature changes associated with the determination of reactants at high concentrations. Sample and reagent streams are metered in constant proportions through coils kept at constant temperature in a thermostat into a mixing vessel. The temperature rise of the solution a t the exit of the mixing vessel is monitored by a thermistor bridge circuit and recorded on a strip-chart recorder.The apparatus can be used successfully even when reaction between sample and reagent is slow. This is demonstrated by the application described here for the determination of acrylonitrile monomer in aqueous polymer latices by reaction with sodium sulphite . FOR some years the free acrylonitrile content of butadiene-acrylonitrile latices has been determined in our laboratories by reaction with sodium sulphite and subsequent potentio- metric titration of the liberated alkali with hydrochloric acid.l It was noted that a considerable amount of heat is produced in this reaction and it was natural to consider measurement of this heat of reaction for the determination of the monomer when a fully automatic procedure suitable for possible on-line application was required.Attention was therefore given to the development of a continuous-flow enthalpimetric analyser to carry out this determination. Sample Ll Reagent TI Thermostat bath Temperature equilibration coils Mixer Thermlstors waste I Control unit Paper presented at Fig. 1. Flow sheet of the simplified apparatus the Second SAC Conference 1968, Nottingham. 0 SAC and the author.TAUBINGER 635 DESCRIPTION OF THE APPARATUS- The flow system developed is shown in outline form in Fig. 1. The essential features of the equipment are a peristaltic pump, which meters sodium sulphite solution and diluted latex sample, and a mixing unit and equilibration coils, maintained at constant temperature in a thermostat, from which the mixed solution flows, after thorough mixing, past a matched thermistor to waste.The temperature difference between this thermistor and a matched thermistor, immersed in a mercury pool in the thermostat bath, is measured by a Wheatstone bridge type circuit and recorded on a strip-chart recorder. G Fig. 2. Sample reservoir. Details of the apparatus are given in the text The sample reservoir-This, as shown in Fig. 2, is a glass vessel, with a capacity of about 80m1, fitted at A with a wide-bore tap to allow the vessel to be emptied through tube B to waste. The level probe has two connections, at E and F, and when the liquid breaks the contact between the two wires at L, wash water is switched on. Wash water enters the vessel at two points : through a capillary tube connection at D, which flushes the vessel from the lower end and at J, which flushes the funnel into the vessel through the B19 joint at H.The bulk of the water entering at D and J overflows at G, to waste. The flow of “flush water’’ should be regulated by the mains tap, not to splash over the reservoir funnel and to allow the ovedow at G to cope. The opening from the vessel at C leads to the sample channel of the peristaltic pump. The reagertt reservoir-This is a large aspirator filled with 20 per cent. w/v anhydrous sodium sulphite solution. The peristaltic pzcm@-This is equipped with two channels. Silicone rubber tubing is used and the pump is adjusted to give a flow-rate of 8 ml minute-l each of reagent and sample. The mixer and thermostat coil assembly-This is shown in Fig.3. The coils are each made of 7-5 feet of inch 0.d. copper tubing. At the lower end they are connected with D r h rubber captive seal connectors to the concentric tube injector, which is made of glass. The mixer assembly consists of a 1$ inch diameter brass tube, which is bolted at the lower end to a 4 inch thick Perspex disc. A watertight seal is obtained by a silicone rubber washer held between the brass tube and the Perspex disc. The Perspex disc has a central hole through which the glass injector passes, the silicone rubber washer giving a watertight seal round the glass. The mixer itself, which is also made of brass, consists of a 8 inch diameter brass tube sealed to a & inch diameter tube at the top and an insert that consists of three636 TAUBINGER : DIRECT DETERMINATION OF FREE [Analyst, Vol.94 discs which are a "push-fit" in the &inch diameter tube. These discs, which have three 1-mm orifices drilled 5rnm from the edge round the circumference, are supported by equidistant spacers on a brass rod. Fixed to the lower end of this brass rod is an inverted brass cup with three 1-mm orifices cut 2mm from the lower edge. The end of the supporting rod is threaded to facilitate removal and cleaning of the mixer. 4 \ Mercury Perspex o r .......... other plastic s i I i cork rubber Glass Brass Fig. 3. Cross-section of the mixer and coil assembly The mixer is held in position over the glass injector with four screws, threaded into the Perspex disc. Again the silicone rubber washer produces a watertight seal.The top end of the mixer, which reaches to the thermistor, is connected to the glass overflow tubes with a Quickfit screw-thread joint, to give a watertight seal. The glass overflow tube passes through a second silicone rubber washer and through a 10 inch diameter Perspex disc, to which the upper end of the 19 inch diameter brass jacket is bolted, giving a watertight seal. This produces an air-insulated space between the mixer and the outer jacket of the assembly. A small bottle of mercury, into which the second thermistor dips, is held by the 10 incl diameter Perspex disc in such a position that the bottle dips into the water of the bath. The water-bath (25" 0.2" C) with thermostat-This is lagged and covered with an asbestos lid, which supports the thermostat, a water-cooling coil and the two standard 2-2-kQ resistors, which are held inside a glass jacket in the water-bath.Fixed to the top of the lid of the water-bath is the small switching box into which the two thermistors and the two standard 2.2-kS2 resistors are plugged. This box permits the substitution of the standard resistors for the thermistors and thus allows a check on the correct functioning of the thermometric analyser control unit to be made. The asbestos lid has an 83 inch diameter hole through which the coils and mixer assembly are inserted, the assembly being supported by the lid of the bath.Fig. 4. Thermometric analyser control unit To face page 6371August, 19691 ACRYLONITRILE IN AQUEOUS COPOLYMER LATICES.PART I1 637 The measuring and recording eqzcipmenb-This is as shown in Fig. 4. The recorder is a 10-mV Honeywell-Brown strip-chart recorder with a chart speed of 8 inches hour-l. The chart-operating switch at the front of the recorder has been replaced by a three-position switch so that the recorder can be independently switched off or run under manual or automatic control. S.T.C. matched thermistors F.23 Standard resitton 2.2 KPOl %type K 191 A.K.U. Muirheld sKT.10 SKT.~ Fig. 5. The thermometric analyser control unit NOTE- TM2 switch actions: C1, changed over all the time the clutch is energised; C2, as C1; M1, changed over just before the end of the timed period; M2, changed over a t the end of the timed period; M3, changed over just after the end of the timed period.Remove all links on the plug of TM2 TM1-the uniset-S1 starts timing when an external on/off switch is switched on, and operates a change-over contact from the end of the timed period until switched off, when it re-sets automatically. Further details are given in the Appendix Contact (1) on SW3 is the manual position and contact (2) on SW3 is the auto pump position All connections taken from plug of TM2 The circuit diagram of the thermometric analyser control unit is shown in Fig. 5. The only mains input to the whole equipment assembly enters this control unit at the front in the bottom left-hand corner, and all ancillary equipment plugs into the back of the control unit. Switched socket outlets are provided at the back of the unit for the recorder638 TAUBINGER : DIRECT DETERMINATION OF FREE [Andyst, VOl.94 and thermostat and unswitched outlets for the flush water, peristaltic pump, recorder chart drive, level probe, the inputs from the thermistors - standard resistance box and the output to the recorder. In the top left-hand corner of the unit is the mains switch, flanked by a fuse and an indicator lamp. The “bridge sensitivity” selector switch is in the top right-hand corner above the coarse and fine “bridge balance” controls. These bridge balance controls are used to compensate for any imbalance in the matched thermistors, and are adjusted to give a zero reading on the chart when water (in place of sample) and sulphite reagent are pumped through the mixer. In our experience the correct bridge balance control setting varies for each pair of matched thermistors.Below the bridge balance controls is the reverse polarity switch, which allows the polarity of the recorder input to be reversed, and once the apparatus is set up it need not be used. In the centre front of the control unit is the flush period timer. This is a 30-minute adjustable timer and we have found that a water “flush period” of 7.5 minutes is adequate for cleaning out the sample reservoir and for keeping the apparatus under equilibrium conditions. Inside the control unit, reached by removing the back panel, is a second timer. This timer has a maximum period of 1 hour, and is used at its maximum to control the automatic re-setting operation of the apparatus to maintain the equipment under equilibrium conditions. In the bottom left-hand corner at the front of the control unit is a row of three switches.The right-hand switch is a three-position switch that allows the peristaltic pump to be independently switched off or run under manual or automatic control. At the left of the row is the automatic sequence switch, which when switched on initiates the automatic sequence, ie., it starts both timers, the peristaltic pump, the recorder chart drive and the flush water. After 7& minutes the automatic sequence control switches off the pump, the flush water and, Q minute later, the recorder chart drive. After 1 hour the cycle is re-set and the sequence is repeated continuously until the automatic sequence is switched off. In the middle is a push-button type switch, labelled “sample cycle.’’ When the empty reservoir is filled with sample, use of this switch allows the sample to be pumped through the analyser before the “flush water” rinses out the reservoir.Thus if the sample cycle switch is operated before the automatic sequence is switched on, the first cycle is modified so that the 1-hour timer, the recorder chart drive and the peristaltic pump only are started as before, while the flush water and flush period timer are only switched on when the level in the sample reservoir has fallen below the tip of the level probe. When the sample level falls below the tip of the probe, a circuit is broken and this activates a relay that starts the flush period timer and the flow of flush water. After the pre-set time (7a minutes) the flush water and the peristaltic pump are again switched off followed, Q minute later, by the recorder chart drive and, after 1 hour, the automatic sequence described earlier comes into operation and continues until it is switched off.The flush water is controlled by an on - off magnetic valve such as the Type R.B.G., made by Teddington Industrial Equipment Ltd., Sunbury, Middlesex. In a sample cycle, the sample is pumped for about 10 minutes before the flush water is switched on by the sample level falling below the tip of the probe, and it therefore takes about 174 minutes from the start before the pump and flush water have stopped. Should it be necessary to run samples more frequently than this, the automatic sequence could be switched off manually as soon as the sample reservoir has been washed clean by the flush water and a fresh sample cycle can then be started.START-UP PROCEDURE- Check that the pumping speed and the flush period timer are correctly set and that the water-bath is a t the correct temperature. Set the recorder and pump to automatic and switch on the analyser. Fill the sample reservoir with water and with the bridge sensitivity at 2 switch the 2-2-kQ resistors into the circuit. Depress the sample cycle switch and then switch on the automatic sequence. The recorder should now give the normal check reading, which is due to imbalance of the thermistors, and will alter when the thermistors are replaced. Having confirmed that the circuit is working, switch the thermistors back into circuit. TheAugust, 19691 ACRYLONITRILE IN AQUEOUS COPOLYMER LATICES.PART I1 639 timer will stop the pump, flush water and chart and after 1 hour the whole cycle will be repeated automatically. SERVICING- A periodic servicing is necessary to keep the apparatus running satisfactorily without breakdown. The bearings on the pump should be oiled weekly and the silicone rubber tubes checked for wear or build-up. It is advisable to change the tubes every 4 weeks, and they should always be changed in pairs. A certain amount of build-up of polymer takes place in the mixer, which should therefore be cleaned weekly. It is convenient to keep two coil - mixer assemblies and simply to change them over. Also, wipe the thermistor immersed in the sample stream with a clean tissue to remove any build-up.METHOD REAGENTS- sulphite (anhydrous). the hydrate. Sodium sulphite, 20 per cent. w/v-Make up a solution to contain 200 g 1-1 of sodium The anhydrous salt is more readily soluble and also cheaper than Teepol solution, 10 per cent. VIV. ANALYTICAL PROCEDURE- Take the sample of latex, arrest polymerisation and remove dissolved butadiene by the techniques described previous1y.l Transfer 50ml of the de-gassed latex to a 100-ml measuring cylinder, add 50ml of 10 per cent. Teepol solution and mix thoroughly. Switch off the automatic sequence, allow the sample reservoir to drain and close the tap. Place a piece of gauze over the funnel (at K, Fig. 2) and pour the diluted sample through it into the reservoir. Select the appropriate bridge sensitivity, activate the sample cycle switch and switch on the automatic sequence.Measure the step height produced, and by reference to the appropriate calibration graph determine the monomer content of the sample. CALIBRATION- The specific heat of latex is similar to that of water and standard solutions for calibration of the analyser could be made up in either medium. We have, however, found it best to prepare our standard solutions for calibration in monomer-free latex diluted with an equal volume of the 10 per cent. Teepol solution. 8 minutes B Fig. 6. Typical recorder traces: A, monomer 2-05 per cent., sensitivity 2; B, monomer 6.23 per cent., sensitivity 1; C, monomer 16.3 per cent., sensitivity 1 Monomer in solution, per cent. Fig. 7. Calibration graphs relating step height reading to monomer content: A, bridge sensi- tivity = 1; B, bridge sensitivity = 2640 TAUBINGER : DIRECT DETERMINATION OF FREE [Artalyst, VOl.94 Make up standard solutions of acrylonitrile monomer in this diluted latex to cover the range 0 to 15 per cent. w/v. These standard solutions must be freshly prepared, as they tend to lose acrylonitrile on standing. Old standard solutions can be used, provided the concentration of monomer is first determined by the titrimetric procedure. Run these standard solutions of acrylonitrile in latex on the thermometric analyser as described under Analytical procedure. Run all solutions on sensitivity setting 1 and the solutions containing less than 4 per cent. also at setting 2. Typical recorder traces are shown in Fig.6. Measure the step height produced for each standard solution and plot the graphs of step height at settings 1 and 2 against concentration of acrylonitrile in the latex. These should be similar to Fig. 7. Prepare also an aqueous standard solution to contain 3.50 per cent. w/v of acrylonitrile and use this standard solution, which in our experience is stable for at least 2 months, to check the calibration daily. DISCUSSION The main difficulty encountered during the development work was the design of an efficient mixer. As the latex samples tend to coagulate, causing some build-up of solids on the sides of the mixer and occasionally the formation of fine solids, any orifices must be reasonably large (about 1 mm), and the mixer must be easily dismantled for cleaning.Also, as reaction is relatively slow, maximum time delay must be built in between the initial mixing and the thermistor. After much trial and error an effective mixer was developed, which repeatedly forced the liquids through restricted orifices into small expansion chambers. The original apparatus was made of glass, fitted with a vacuum jacket, but although this gave efficient mixing, temperature equilibration was slow and a marked hysteresis effect was observed. Fig. 8 shows the different calibrations obtained when the concentration of acrylo- nitrile was successively increased and successively decreased. The reasons for this hysteresis effect were the high heat capacity and poor thermal conductivity of the glass. The final design of mixer, Fig. 3, which gave rapid equilibration with no evidence of hysteresis, was therefore made of metal (brass).Acrylonitrile, per cent. Fig. 8. Hysteresis effect caused by the high heat capacity and poor conductivity of the glass mixer : x , increase in acrylonitrile ; 0, decrease in acrylonitrile As the rate of reaction between sodium sulphite and acrylonitrile is relatively slow, reaction has not reached completion when the temperature rise is measured. To keep the rate of reaction constant it is, therefore, necessary to keep the reactants at a constantAugust, 19691 ACRYLONITRILE IN AQUEOUS COPOLYMER LATICES. PART 11 641 temperature by immersing the temperature equilibration coils and the mixer in a water-bath fitted with an efficient thermostat. The range of monomer concentration in the samples examined lay between 0 to 15 per cent.w/v and, as the maximum solubility of the sulphite in water is only slightly over 20 per cent. w/v, a 1 : l metered ratio of sample to reagent limits the concentration of monomer determined to 10 per cent. w/v. To extend the range of concentrations covered beyond this, all samples were diluted in a 1 : 1 volume ratio before testing. Dilution with 10 per cent. v/v aqueous Teepol (additional emulsifier) was preferred, as the extra emulsifier helps to prevent coagulation of the latex. Latex readily coagulates and forms a skin, an effect accelerated by exposure to air. It is therefore necessary to prevent air from entering the mixer and coil assembly. A probe was therefore built into the sample reservoir (Fig.2) which, when the sample level drops below a certain point, automatically starts a flow of water that washes out the vessel and prevents air from entering the mixer. An electric timer was fitted to switch off the water, the pump and the recorder chart drive after a pre-set time (0 to 30 minutes), thus saving reagent and chart paper. When 20 per cent. w/v sodium sulphite is mixed with water a marked cooling effect is observed and it was noted that from start-up a considerable delay occurred before the apparatus reached equilibrium. A second electric timer was therefore fitted which, once every hour, switches on the pump, the wash water and the recorder for the above pre-set time. The apparatus is thus maintained at or near equilibrium. The sensitivity of the instrument has been made adjustable by providing bridge voltage selection at 0-75 V (sensitivity 2) and 0-15 V (sensitivity 1).This provides the choice of one setting for the range 0 to 4 per cent. w/v (0" to 1-75' C) of acrylonitrile and a less sensitive setting for higher concentrations (0" to 8-8" C ) . RESULTS by the titrimetric and enthalpimetric procedures. Typical results obtained on latex samples are given in Table I, which compares recoveries TABLE I COMPARISON OF RECOVERIES OF ACRYLONITRILE BY THE TITRIMETIC AND ENTHALPIMETRIC METHODS Tityimetyic- Polymerisation run Emthalpimetric- Titvimetk- Stripped latices Enthalpimetyic- monomer, per cent. . . . . 15.3 10.7 9-0 7.7 5.5 4.1 3.7 3.5 monomer, per cent. . . . . 15.4 11.2 9.0 7.3 5.8 3.9 3.5 3.4 monomer, per cent.. . . . 0.38 0.70 0.03 0.77 0.25 0.27 0.30 0.28 monomer, per cent. . . . . 0.46 0.67 0.10 0.57 0.25 0.46 0.33 0.27 It will be seen from these results that recoveries of acrylonitrile on polymerisation runs (3 to 15 per cent.) by the enthalpimetric procedure are better than *lo per cent. relative. On stripped latices (0.05 to 1 per cent.) (latices from which unpolymerised acrylonitrile has been recovered) the accuracy is somewhat less, being k0.2 per cent. absolute. The analyser has proved a useful instrument for the routine determination of free acrylonitrile in latices during polymerisations and also for the determination of residual acrylonitrile in stripped latices. Results of analyses are available within 10 minutes of sampling and about four tests can be completed in 1 hour. The main advantage of this procedure lies in the saving of operator time, which is reduced to a minimum of considerably less than 5 minutes per test.This analyser should also be suitable for other continuous-flow enthalpi- metric analysis applications, and although applied here to a slow reaction and large tem- perature rises with high concentrations, it should be equally suitable for fast reaction and smaller temperature rises.642 TAUBINGEH Appendix LIST OF COMPONENTS = 68,000-ohm, +-watt, wire-wound resistor = 435-ohm, 30-watt, wire-wound resistor = 68,000-ohm, +-watt, wire-wound resistor = 1.2-megohm. &-watt, resistor, carbon = 3.3-megohm, &-watt, resistor, carbon = 750-ohm, 3-watt, wire-wound resistors = 180-ohm, +-watt, metal oxide, 1 per cent. = 560-ohm, *-watt, metal oxide, 1 per cent. = 6000-ohm, 2-watt, variable resistor, He1 Pot = 100,000-ohm, 2-watt, variable resistor, Pot Rl % R, R, REi Re, R, R, R,, R1o VRl V% VR,, VR, = 10,000-ohm, $-watt, variable resistor, Flatpot Fl = Fuse holder, 5A LP,, LP, = Mains neon SW,, SW, = Mains S.P.S.T., 6A SW,, SW, = Switch centre off, 2 P.D.T. 6A = Switch, type MS3, 3 P.D.T. = Winkler 1 P. 5-way SKT,, SKT, = 3-pin switch sockets S KT, = 4-pin C/Socket S KT, = 3-pin C/Socket S KT, = 2-pin C/Plug SKT, = Min mains socket SKT, = 2-pin Screenector C/Socket SKT,, SKT,, = 3-pin Screenector C/Sockets SKT,, SKT,, = 3-pin Screenector C/Plugs SKT,,, SKT,, = Type SM3 C/Plugs SKT,,, SKT,, = Type SM2 C/Sockets Tl = Transformer min mains RLA = Relay, type TM06 RLB = Relay, type MR1600K RLD, RLE = Relay, type MK2P = Uniset-SL 60 min = Robinson Chronset-B Sync. Timer 36 min = Push button, 1A = Power supply, 0.6 A = O.l-pF capacitor, 400-V d.c. polyester = 220-pF capacitor, ceramic = Valve ER21A cold-cathode tube SW6 swt3 TMl TM, PBl PSl c, c2 Vl REFERENCE 1. Taubinger, R. P., Analyst, 1969, 94, 628. Received August Id, 1968 Accepted February loth, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400634

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, August, 1969, Vol. 94, $9. 643-662 643 Molecular-emission Spectroscopy in Cool Flames Part W.* The Determination of Chloride, Bromide and Iodide by Themnal-emkion Spectroscopy in the Presence of Indium Saltst BY R. M. DAGNALL, K. C. THOMPSON AND T. S. WEST (Chemistry Department, Imperial College, London, S . W.7) A sensitive and selective molecular-emission method for the determination of chloride, bromide and iodide is described, involving measurement of the intense InC1, InBr and In1 band-emission signals. A cool nitrogen - hydrogen diffusion flame is used and measurements are made at 360, 376 and 410 nm, respectively. Chloride can be determined in the presence of a large excess of bromide and iodide, bromide in the presence of a large excess of chloride and iodide, and iodide in the presence of a large excess of chloride.The use of other flame types and the gallium halide system are also described. IN previous papers in this series, we reported on the application of cool nitrogen - hydrogen flames to the determination of sulphur,l phosphorus2 and tin.3 These procedures were based on the blue emission from the S2 species for sulphur, the green emission of the H-P-0 species for phosphorus and the red emission of Sn-H for tin. Like sulphur and phosphorus, chlorine, bromine and iodine have their principal atomic-resonance lines in the vacuum ultraviolet region of the spectrum4; for chlorine these occur at 138.0, 139.0 and 139.7 nm, for bromine at 154-1 and 157.7 nm and for iodine at 178.3 and 183.0 nm. The absorption by air, flame gases and quartz optics precludes the measurement of atomic absorption and atomic fluores- cence by normal procedures in this region.For this reason we have sought to form molecular species of the halogens capable of strong emission in the cool flames needed for their formation. Indium was found to be the most suitable element for this purpose. Conventional methods for the determination of the halide ions by flame methods involve indirect procedures such as addition of a known amount of silver ions to the halide-containing solution, filtration of the silver halide precipitate and measurement of the silver concentration of the filtrate, either by flame-emission or by atomic-absorption ~pectroscopy.~ However, this method does not discriminate between chloride, bromide and iodide.Gilbert6 gives a method for chlorine based on the emission of InCl in a separated air - hydrogen flame; a piece of copper tubing was coated with indium metal and suspended between the primary and secondary combustion zones of a separated air - hydrogen flame. This arrangement was applied to the determination of chlorine in organic vapours diluted with nitrogen. Appreci- able blank values were observed, which were attributed to a chloride flux used to plate the copper with indium. Also, the sensitivity of this arrangement decreased drastically as the indium surface aged. We have previously noted that when a solution of tin(I1) chloride was aspirated into the hydrogen - nitrogen diffusion flame blue Sn-CI and red Sn-H emission were ~bserved.~ The SnCl emission was concentrated in a sharp cone in the centre of the flame, and when an excess of hydrobromic acid was added the greenish blue emission of Sn-Br was observed.The SnCl and SnBr emissions would thus appear to take place in the coolest region of the flame (k, in a small central cone just above the burner port) where the droplets are still evaporating. Because the thermal energy of the flame in this region is insufficient to excite the SnCl and SnBr spectra, obviously a chemiluminescent reduction reaction is taking place. Thus the flame is acting as a reaction medium for the production of chemiluminescence. Strong InC1, InBr and In1 emissions were found to occur in this cool diffusion flame. Those of InCl and InBr lie in the near ultraviolet and are not clearly visible, but the In1 emission is visible as a dark blue colour in the central regions of the flame.The indium halide emission was much more intense than that of the tin halide, and the main emission peaks of the three halide species were well separated. * For details of earlier parts of this series, see reference list, p. 652. t Paper presented a t the Second SAC Conference 1968, Nottingham. 0 SAC and the authors.644 DAGNALL, THOMPSON AND WEST [Analyst, Vol. 94 The indium halide emission is thought to be caused by a chemiluminescent reduction reaction of an indium(II1) species leading to an excited indium(1) halide molecule. The low thermal energy in the central regions of the flame is otherwise insufficient to account for the intense indium halide emission.There are several reasons why the indium halide emissions are so intense in the cool diffusion flame. For example, in the presence of hydrogen, indium, unlike aluminium and boron, does not form a stable oxide species but relatively stable monohalides of indium do exist. In addition, indium trihalides, except for the fluoride, have low melting-points (see Table I). Thus no involatile, unreactive clotlets are formed in the flame. Yet again, a highly reducing oxygen-free flame, with negligible background at the wavelengths of interest, produces indium(1) halides in an excited state, possibly by exothermic reduction from the indium(II1) state. TABLE I MELTINGPOINTS OF THE TRIHALIDES OF GALLIUM AND INDIUM’ Temperature, O C - Fluoride .. .. ..960 (sublimes) 1170 Chloride . . .. .. 78 686 Bromide . . .. .. 122 436 Iodide . . .. .. 212 210 X GaX, InX, EXPERIMENTAL APPAFUTUS- A Unicam SPSOOA flame-emission atomic-absorption spectrophotometer, fitted with the standard air - acetylene (rectangular) burner head and a non-standard burner head described subsequently in the text, was used. Fuel gas-Hydrogen, from a cylinder. Dilzcent gas-Nitrogen, from a cylinder. Indium nitrate, 2-94 X REAGENTS- M-Dissolve 3.378 g of indium metal (99.9 per cent.) in 50 ml of analytical-reagent grade nitric acid and dilute to 1 litre with distilled water. The final nitric acid concentration is about 0.70 M. Indium chloride, 2.30 x 10-2~-Dissolve 2.642g of indium metal in 1 1 O m l of Aristar hydrochloric acid and 1OOml of distilled water, and dilute to 1 litre with distilled water.The final hydrochloric acid concentration is about 1 . 0 ~ . Gallium nitrate, 4-0 x M-Dissolve 0.6972 g of gallium metal in 20 ml of analytical- reagent grade nitric acid (dissolution is very slow) and dilute to 250 ml with distilled water. Hydrochloric acid, 1 M-Prepare by diluting Aristar hydrochloric acid, standardising and subsequently adjusting the solution to exactly 1 M. Nitric acid, 1 M-Prepare by diluting andytical-reagent grade nitric acid, standardising and subsequently adjusting the solution to exactly 1 M. Hydrobromic acid, 1 M-Prepare as for nitric acid. Hydriodic acid, 1 M-Prepare as for nitric acid, and store in a dark-coloured bottle. In this study two burner heads were used: the standard 1.8 x 7.5-cm air - acetylene emission head supplied with the Unicam SP9OOA spectrophotometer and a 7-cm length of 10 mm i.d.Pyrex tubing, which was sealed into the conventional burner stem by a ring of PVC tubing. When the tube burner was used the burner stem was brought nearer to the monochromator slit by removing the brass block behind the burner clamp mounting. This brought the “tube flame” into a similar position to the flame obtained on the normal 1-8 x 7-5-cm air- acetylene burner head. The position of the Pyrex tube with respect to the monochromator slit was critical and was optimised before taking any measurements. The optimum tube diameter for the measurement of InCl emission was found to be about 10mm. Nitrogen was used as the nebulising gas (flow-rate 4 litres minute-l), with the conventional Unicam SP9OOA nebulising system (set at 15 p.s.i.).Hydrogen was intro- duced as usual at the bottom of the burner through a propane jet at a pressure somewhat above that necessary to prevent the flame from lifting off the burner (flow-rate 1 litre minute-’,August, 19691 MOLECULAR-EMISSION SPECTROSCOPY IN COOL FLAMES. PART IV 645 with the “tube burner”). The tube burner permitted the use of lower hydrogen flow-rates than the standard emission burner head for the maintenance of a stable flame. The tube burner gave about %fold greater emission intensity from InCl under optimised conditions than the conventional burner head. (Two to three-fold increases in S, and HPO emission, from sulphur dioxide and phosphoric acid, respectively, were also observed.) These increases were thought to be caused by less air entrainment giving a larger oxygen-free reaction zone in the central region of the flame.This idea was borne out by a comparison of the tempera- tures in the centre of the two flames while nebulising distilled water (Table 11), and also by 1 InCl A4 I n2 L Wavelength, nrn Fig. 1. Emission spectrum of InC1: A, sample with indium 2-94 x 1 0 - 8 ~ , chloride 1 x 1 0 - z ~ and nitrate 7 x 10-ZM; and B, blank with indium 2.94 X 10-SM and nitrate 7 x 10-ZM. Slit width 0.01 mm, gain 3,7 and band width 2 350 360 370 380 390 400 Wavelength, nm Fig. 2. Emission spectrum of InBr: A, sample with indium 2.94 x ~O-*M, bromide 1 x 1 0 - z ~ and nitrate 7 x 1 0 - s ~ ; and B, blank with indium 2.94 x 10-8~ and nitrate 7 x 10-ZM.Slit width 0.01 mm, gain 2,6 and band width 2646 DAGNALL, THOMPSON AND WEST [Artalyst, VOl. 94 the fact that when distilled water is aspirated into the tube flame a fine mist is clearly visible in a 5-cm long cone in the centre of the flame. (In1 emission was visible around the edge of this cone.) SPECTRUM OF INDIUM IN THE NITROGEN - HYDROGEN FLAME- When a solution of indium nitrate was nebulised into the nitrogen - hydrogen flame, the lines and bands shown in Table I11 were observed. The spectra of InC1, InBr and In1 are shown in Figs. 1, 2 and 3. No InF spectrum was observed under any conditions. This is probably due to the much greater stability of indium(II1) fluoride compared with other indium halides. Thus, there is insufficient energy liberated on reduction to give an excited Wavelength, nrn Fig.3. Emission spectrum of In1 : A, sample with indium 2.94 x 10-s~, iodide 1 x 1 0 - 2 ~ and nitrate 7 x 1 0 - 2 ~ ; and B, blank with indium 2.94 x 1 0 - s ~ and nitrate 7 x lo-%. Slit width 0.01 mm, gain 3,O and band width 2 TABLE I1 COMPARISON OF TEMPERATURES AT CENTRE OF THE FLAMES Temperature, “C r 1 Burner head 1 cm above burner 2 cm above burner Acetylene emission head* . . .. 280 330 Tube burner, 1 cm* . . .. .. 170 185 * Hydrogen 1 litre minute-1 and nitrogen 4 litres minute-’. The method of temperature measurement is described in a previous paper.’ TABLE I11 Species Description and wavelengths* Atomic lines 451.1 and 410.2 nm Diffuse broad bands with maxima a t 382, 373 and 368nm A series of bands between 620 and 550 nm (because of the poor resolution of the silica prism There was also a weak Diffuse band system between 450 and 400 nm.There was also an unassigned weak band In In, In-H in this spectral region only three main bands were resolved). band system about 450 nm at 490 nm In-0 * Wavelengths were checked with reference 8. In-F molecule (0,O band at 319.7 nm).8 The “C”8 systems of the InCl and InBr spectra (InCl, 0,O band at 267.2 nm and InBr, 0,O band at 285.2 nm) were not observed. This was probably because the energy liberated in the reduction of indium(II1) species to form InCl and InBr molecules is insufficient to excite the molecules to this high energy state.August, 19691 MOLECULAR-EMISSION SPECTROSCOPY IN COOL FLAMES.PART IV 647 When air or nitrous oxide was added through a third jet in the burner stem,g the InC1, InBr, InI, In, and InH emissions were destroyed, even at very low air or nitrous oxide flow-rates. A nitrogen - hydrogen flame burned on a total-consumption burner gave negligible emission, probably because the turbulent nature of the flame causes considerable entrainment of air and gives rise to a much hotter and more oxidising flame than that obtained with the simple pre-mixed diffusion flame. INDIUM CHLORIDE EMISSION When a solution containing indium and chloride ions was nebulised into the diffusion flame the characteristic spectrum shown in Fig. 1 was obtained. The emission was three times more intense with the tube burner than with the conventional air - acetylene head and consequently the former was used in all further studies.Maximum InCl emission (measured at 360 nm) occurred with the burner top 3 cm below the bottom of the mono- chromator slit and the hydrogen flow-rate such that the flame was near the lift-off point (0.6 litre minute-1). In practice, it was found that the optimum results were obtained with the burner head 1* cm below the monochromator slit and the hydrogen pressure some way above the lift-off point, about 1 litre minute-1). The nitrogen pressure was set at 15 p.s.i. on the instrument gauge. These conditions resulted in a more stable flame than that obtained with a lower hydrogen flow-rate, although this gave slightly less InCl emission. The InCl emission was almost independent of the hydrogen flow-rate between 0.8 and 1.2 litres minut e-1 .EFFECT OF INDIUM AND CHLORIDE CONCENTRATIONS ON InCl EMISSION- A stock solution of indium nitrate in a 25-fold excess of analytical-reagent grade nitric acid was used in this study. It was found that, with a fixed chloride concentration of 5 x l o - 3 ~ and varying amounts of indium (in sample and blank), the InCl emission was almost constant until the chloride-to-indium ratio became 4 : 1. Even at a chloride-to-indium ratio of 8: 1, the InCl emission dropped only about 30 per cent. At a lower fixed chloride concentration of 2 x M, the decrease in the InCl emission began at a chloride-to-indium ratio of 1.5: 1. With a fixed indium concentration of 1-45 x 10-3 M the InCl emission in- creased with increasing chloride concentration, and was still increasing at a chloride-to- indium ratio of 100: 1 (see Table IV).The graph of the InCl emission with increasing chloride concentration is concave, with respect to the concentration axis, up to a chloride-to-indium ratio of about 7 : 1 ; with higher ratios it becomes increasingly convex. This behaviour suggests that some reaction is taking place that is strongly dependent on the chloride concentration (ie., a rnass-action effect). TABLE IV EFFECT OF INCREASING CHLORIDE CONCENTRATION ON InCl EMISSION Indium concentration: 1-47 x M ; slit width 0.015 mm M ; band width 3 Recorder reading at 360 nm Nitrate concentration: 3-5 x Chloride 1 concentration, M Gain InCl Background* 1 x 10-3 2.10 2-7 2.0 2 x 10-3 2.10 7 2-0 5 x 10-3 2.10 28.5 2.0 10 x 10-3 2.10 58.5 2.0 15 x 10-3 2-10 76.5 2.0 1 x 10-2 2.1 17 0.5 2 x 10-2 2.1 28-5 0.6 6 x 2.1 54.5 0.6 10 x 10-8 2.1 76 0.5 15 x 2.1 87 0.6 20 x 10-2 2.1 92 0.6 * The blank solution was 1.47 x 10" M with respect to indium and 3.5 x lo-* M with respect to nitrate. The effect of the indium concentration on the InCl emission was not critical, provided the chloride-to-indium ratio was less than about 3.Part of the increase observed in Table IV is caused by increased background at 360nm in the presence of chloride (see below), but648 DAGNALL, THOMPSON AND WEST [Afialyst, Vol. 94 this increase levels off at a chloride concentration of 5 x 1 0 - 3 ~ , and would only account for about 1 per cent. of the signal at the highest chloride concentration.EFFECT OF NITRATE CONCENTRATION ON InCl EMISSION- M with respect to chloride and 2.94 x 10-3 M with respect to indium was almost independent of the nitrate concentration over the range 6 x lo-, to 2 x 1 0 - 1 ~ nitrate. EFFECT OF CHLORIDE CONCENTRATION ON THE In, EMISSION- The background at 360nm, observed on aspirating an indium nitrate blank solution, was thought mainly to emanate from In, emission, because the flame itself did not exhibit any background at this wavelength. The intensity of the In, bands at 374 and 381 nm from an indium nitrate solution was found to be directly proportional to the indium con- centration. The In, emission showed a similar dependence on the flame parameters as the InCl emission. The In, emission at 374 and 381 nm from a fixed nitrate concentration of 3.5 x 1 0 - 2 ~ and an indium concentration of 1-36 x 1 0 - 3 ~ was dependent on the chloride concentration (Fig. 1) and increased 26-fold when the chloride concentration was altered from zero to 5 x 10” M.Only a further slight increase was observed at higher chloride con- centrations. This increase in the emission at 374 and 381 nm was caused by In, and not by InCl, as the latter gives negligible emission at these wavelengths and at the slit widths and gains used. At low indium concentrations the increase in the In, emission was much less and, at an indium concentration of 2.7 x 10“‘ M and a chloride concentration of M, the emission increased only 1.3 times from its initial value before levelling off. When the chloride-to-indium ratio was less than unity the increase in the In, signal at 374nm was almost proportional to the chloride concentration.Thus it would appear that the background at 360 nm (if, as assumed, it is due to In,) is slightly dependent on the chloride concentration. If there is an excess of indium present, then this increase should be almost proportional to the chloride concentration. Calibration graphs were plotted over the chloride concentration ranges 2 x 104 to 4 x 1 0 - 3 ~ (7-1 to 142 p.p.m.) and 5 x to W 3 M (1.8 to 35.5 p.p.m.), with indium concentrations of 5.42 x lo4 and 1-36 x 1 0 - 3 ~ and nitrate concentrations of 0.14 and 3.5 x 1 0 - 2 ~ , respectively. Both graphs were slightly concave with respect to the concen- tration axis, the former being the more concave. The blank signal at 360nm for the two graphs was 25 and 80 per cent., respectively, of the signal of the most concentrated solution of each graph; the blank solutions were 5.42 x and 1.36 x 1 0 - 3 ~ with respect to indium and 1-4 x 10-1 and 3-5 x ~ O - , M to nitrate, respectively.The 5 x 10-5 to 10-3 M chloride calibration graph in the presence of an indium concen- tration of 2.71 x 10-4 M and a nitrate concentration of 7 x M was convex at high chloride concentrations. The blank signal at 360 nm was only 25 per cent. of the signal of the most concentrated solution. This curious behaviour is difficult to explain, but it must be appre- ciated that the InCl emission does not take place under conditions of thermal equilibrium. Also, the emission is dependent on one or more “unknown” reaction that produces excited InCl molecules.The shape of the calibration graph is not only dependent on the chloride- teindium ratio, but at low concentrations also on the absolute indium concentration. It is difficult to give a limit of detection for chloride, because this would depend on the chloride-to-indium ratio. However, with an indium concentration of 1.36 x M, the limit of detection (signal-to-noise ratio = 1) is about 2 x M (0.7 p.p.m.). At these low chloride levels contamination problems were serious. INDIUM BROMIDE EMISSION The emission of a solution 2 x When a solution containing indium and bromide ions was nebulised into the nitrogen - hydrogen diffusion flame, the characteristic spectrum shown in Fig. 2 was obtained. The peak of the InBr emission lies in the In, band region, and hence the InBr peak at 376nm was used for all measurements.This gave a better signal-to-blank ratio than the peak at 373 nm. As the In, emission, which was thought to constitute largely the background emission at 376 nm, was dependent on the halide concentration, a large excess of hydrochloric acid was added to all solutions. For this purpose Aristar hydrochloric acid, for which a maximum bromide and iodide impurity of O-OOOl per cent. is claimed, was used. Thus it was reasonedAugust, 19691 MOLECULAR-EMISSION SPECTROSCOPY IN COOL FLAMES. PART IV 649 that the addition of bromide ions to a solution containing a large excess of chloride ions should not alter the In, background emission. The InBr emission (in the presence of a 20-fold excess of chloride) was not very dependent on the hydrogen flow-rate or the position of measurement in the flame.Therefore, the same hydrogen flow-rate was used as before for the InCl investigation, and the burner height was set 1 cm below the monochromator slit. Two stock solutions of indium were used, one containing an excess of analytical-reagent grade nitric acid and the other an excess of Aristar hydrochloric acid. EFFECT OF CHLORIDE CONCENTRATION ON InBr EMISSION- The effect of increasing the chloride concentration on the InBr emission at 376nm in nitrate media was found to give an increase in the InBr emission, as shown in Table V. TABLE V EFFECT OF CHLORIDE CONCENTRATION ON InBr EMISSION Indium concentration: 2.94 x 10-3 M; slit width 0.02 mm Nitrate concentration: 7 x lo-, M ; gain 3,l Bromide concentration: 2 x M; band width 3 InBr signal minus blank Emission from blank Chloride 376 nm solution at 376 nm* concentration, M (recorder reading) (recorder reading) 0 23 12 1 x 10-2 39 28 2 x 10-2 41 29 5 x 10-8 41.5 28 10 x 10-2 39.5 26 20 x lo-* 34 23 * The blank solution contained the same concentration of indium, chloride and nitrate as the sample.The decrease observed at a chloride concentration of 20 x M is probably the result of a viscosity effect on the nebulisation efficiency, because the signal and blank readings decrease in the same ratio. Thus it would appear that buffering the solution with a large excess of chloride should not only overcome any interference from chloride, but also increase the intensity of the InBr emission.EFFECT OF NITRATE CONCENTRATION ON InBr EMISSION- media was found to give a large increase in the InBr emission (Table VI). TABLE VI EFFECT OF NITRATE CONCENTRATION ON InBr EMISSION Indium concentration: 2-30 x 10-8 M; slit width 0.02 mm Chloride concentration: 1 x 10-1 M; gain 3,2 The effect of increasing nitrate concentration on the InBr emission at 376 rim in chloride Bromide concentration: 2 x M; band width 3 Nitrate concentration, 0 2 x 10-2 6 x 8 x 10-2 10 x 10-8 20 x 10-0 InBr signal minus blank at 376 nm 28 41 60 78 86-5 95-6 M (recorder reading) Emission from blank solution at 376 nm (recorder reading) 10 22 34 39 40 40 * The blank solution contained the same concentration of indium, chloride The increase in the InBr emission observed in the presence of hydrochloric and nitric acids was thought to be caused by a decrease in the degree of hydrolysis of indium salts species in the flame clotlets at low pH values.and nitrate as the sample.650 DAGNALL, THOMPSON AND WEST [Analyst, Vol. 94 It is possible that a complex salt containing nitrate groupings, e.g., In(NO,),Br or In(NO,)ClBr, is formed in the flame clotlets, and these may more readily be reduced to excited InBr molecules than any indium chloride (bromide) species. It is also possible that the presence of a large excess of anions may lessen the rate of reaction of indium species with hydrogen, which gives rise to indium metal. To minimise nitrate and chloride interference, therefore, the solution should be buffered with a large excess of nitrate and chloride ions.EFFECT OF INDIUM CONCENTRATION ON InBr EMISSION- The effect of increasing indium concentration on the InBr emission at 376 nm in chloride and nitrate media is shown in Table VII. The InBr emission increases with increasing indium concentration, but is less than expected for a linear dependence. A linear calibration graph was obtained over the concentration range 5 x to 10-3 M bromide (4 to 80 p.p.m.), with an indium concentration of 1-15 x M, a hydrochloric acid concentration of 5 x M and a nitric acid concentration of 10-1 M. The limit of detection for bromide (signal-to-noise ratio = 1) at the above indium concentration was 2 x 10-SM (1.6 p.p.m.). This value could be reduced if the indium, chloride and nitrate concentrations were also reduced. TABLE VII EFFECT OF INDIUM CONCENTRATION ON InBr EMISSION Chloride concentration: 1 x 10-l M; slit width 0.02 mm Nitrate concentration: 1 X 10-1 M; gain 3,lO Bromide concentration: 1 x 10" M; band width 3 InBr signal minus blank Emission from blank Indium at 376 nm solution at 376 nm* concentration, M (recorder reading) (recorder reading) 4.6 x 10-4 18 23 9.2 x 10-4 26.6 40 1.38 x 10-8 37.5 63.5 2.3 x 10-8 66 78 * The blank solution contained the same concentration of indium, chloride and nitrate as the sample.INDIUM IODIDE EMISSION When a solution containing indium and iodide ions was nebulised into the diffusion flame the characteristic spectrum shown in Fig. 3 was obtained. The peak of the In1 emission occurs at 410 nm and corresponds to an indium atomic-resonance line at 410.2 nm.There are also some weaker bands at about 399 nm. At an iodide concentration of M, a nitrate concentration of 7 x M and an indium concentration of 2-94 x lo-, M, the ratio of the In1 emission to the atomic-line emission in the cool diffusion flame was about 30 : 1. This ratio decreases with increasing indium concentration. The addition of a 60-fold molar excess of hydrochloric acid (6 x 10-2~) to a solution 1-45 x 10-SM with respect to indium, 3-5 x 1 0 - 2 ~ to nitrate and 8 x 104~ to iodide increased the In1 signal 4-fold and the background 3-&fold. Optimum In1 emission occurred under flame conditions similar to those for InBr emission, and all measurements were taken with the burner 1 crn below the bottom of the monochromator slit.All solutions were buffered with an excess of hydrochloric and nitric acids. A calibration graph was plotted over the range 4 X 10" to 8 x 1 0 4 ~ (5.1 to 102 p.p.m.) iodide in the presence of 1-15 x M nitric acid. The background signal at 410nm was not only due to the indium atomic-resonance line emission at 410.2 nm, but also to In0 (about 50 per cent.) and possibly In, emission. The total blank signal was 90 per cent. of the intensity of the In1 emission from the 8 x 1 0 4 ~ iodide solution. The limit of detection (signal-to-noise ratio = 1) at the above indium con- centration was 1-5 x 10-6 M (2 p.p.m.). This could be lowered by reduction of the indium, nitrate and chloride concentrations. M indium, 5 x M hydrochloric acid and 5 xAugust, 19691 MOLECULAR-EMISSION SPECTROSCOPY IN COOL FLAMES.PART IV 661 INTERFERENCE EFFECTS OF FOREIGN IONS InCl EMISSION- The test solution was 1 0 - 3 ~ with respect to chloride, 1-18 x 1 0 3 ~ to indium and 7 x 10-2 M to nitrate, and measurements were made at 360 nm. Five-fold molar excesses of sulphate and phosphate (5 x 10-3 M final concentration) gave about 60 per cent. decreases in the InCl signal at 360 nm. A 10-fold molar excess of acetic acid M final concentration) gave a 4 per cent. decrease in the InCl signal. A 100-fold molar excess of hydrobromic acid (final concentration 10-1 M and chloride impurity less than lo4 M) increased the blank signal about &fold and decreased the InCl emission at 360nm about 3-fold. The large increase in the blank signal at 360 nm is mainly caused by the presence of a weak InBr band at 360 nm.By measuring the InCl emission at 350nm (no InBr emission occurs at this wavelength), the blank signal was unaffected by the presence of bromide. A linear calibration graph was obtained at 350nm, in the presence of 2 x 1 0 - l ~ bromide, 2.94 x 1 0 - s ~ indium and 7 x 1 0 - 2 ~ nitrate, over the chloride range 4 x lo4 to 2 x 1 0 - 3 ~ (14 to 71 p.p.m.). The blank solution contained the same indium, nitrate and bromide concentrations as the sample. The limit of detection under the above conditions was about 1 0 4 ~ (3.5 p.p.m.) chloride. The presence of a 100-fold excess of iodide did not affect the InCl or the blank signal at 360nm by more than 15 per cent. The sulphate and phosphate interferences can be overcome by adding barium nitrate solution to precipitate barium sulphate and phosphate before using an ion-exchange procedure to remove metal ions.The interference of other halides and nitrate can be overcome by buffering the sample and blank solutions with a large excess of the ions in question. InBr EMISSION- The test solution was 2 x 1 0 - 3 ~ with respect to bromide, 1.18 x 1 0 4 ~ to indium, 5 x 10-2 M to nitrate and 5 x Phosphate and sulphate had an effect similar to that observed with InCl emission. (At high sulphate concentrations interference caused by S, emission1 at 376 nm is observed.) A 15-fold molar excess of hydriodic acid (final concentration 3 x M) added to the sample and the blank slightly decreased the blank signal and doubled the InBr emission at 376 nm.A 5-fold molar excess of hydrofluoric acid (final concentration M) had no effect on the sample or blank signal. This was thought surprising as fluoride ions would be expected to form involatile species with indium in the flame. To overcome iodide interference it would be necessary to buffer the solution and the blank with an excess of bromide-free hydriodic acid. In1 EMISSION- The test solution was 8 x 1 0 4 ~ with respect to iodide, 1-18 x ~O"M to indium, 6 x 10-2 M to nitrate and 5 x M to chloride, and measurements were taken at 410 nm. A 125-fold molar excess of hydrobromic acid (final concentration 10-1 M), when added to the sample and blank, slightly increased the blank signal and almost completely destroyed the In1 emission at 410 nm.A 25-fold molar excess of hydrobromic acid decreased the In1 emission 5-fold. The large decreases in the In1 emission in the presence of bromide are difficult to explain, because addition of hydrochloric acid to a nitric acid solution of indium and iodide ions increases the In1 emission. Thus it would appear that bromide has a depressing and chloride an enhancing action on the In1 emission. A 12-fold molar excess of hydrofluoric acid ( 1 0 - 2 ~ final concentration) had no effect on the In1 or the blank signal. It is difficult to generalise about interferences because the indium halide emission does not take place under conditions of thermal equilibrium, and the effect of a given interference will be dependent to some extent on the indium, halide and nitrate ion concentrations.The size of the droplets produced by the nebuliser would also be an important factor. SEPARATED AIR - HYDROGEN FLAME- A separated air - hydrogen flame was obtained in a 2.4 cm i.d. Pyrex tube placed over the burner.1 This gave InC1, InBr and InH emission from the interconal region around the burner top. The InH emission was clearly visible and formed a pink coloured ring around the top of the burner. The deep blue coloured indium atomic emission was clearly visible just above the burner top. The InCl and InBr emissions were only about 20 per cent. of the M to chloride, and measurements were taken at 376 nm. Chloride ion was without effect under the conditions used. APPLICATION OF OTHER FLAME TYPES652 DAGNALL, THOMPSON AND WEST intensities observed in the diffusion flame.However, on cooling the outer surface of the separator with compressed air, the InCl and InBr emission increased about 4-fold. When cooling was stopped, the emission increased further, thus indicating that condensation on the separator had occurred during the cooling period. This condensation is undesirable because of the “memory” effects. In1 emission was not observable at 410 nm because of the great intensity of the indium atomic emission at 410 nm in this “hot” oxygen-free interconal region. The separated flame was considered to be inferior to the essentially more simple diffusion flame and was, therefore, not studied further. SHEATHED AIR - HYDROGEN FLAME- An air - hydrogen flame was sheathed with nitrogen via an arrangement that allowed a laminar “wall” of nitrogen to flow up the outside of the flame.This effectively produced a separated flame in which the secondary combustion zone was lifted off about 3 cm. However, this gave only very weak InCl and InBr emission, which suggests that the flame is hotter than the mechanically separated flame. In1 was not observed because of the high intensity of the indium atomic emission at 410nm. The increase in the intensity of the indium resonance emission in this flame is, no doubt, caused by the minimal formation of In0 species, as well as the high temperature of the pre-mixed (and separated) flame. BEHAVIOUR OF GALLIUM SPECIES IN THE NITROGEN - HYDROGEN DIFFUSION FLAME- Weak GaH and GaO emission was observed when a solution of gallium and nitrate ions was nebulised into the diffusion flame.When chloride, bromide or iodide ions were present, GaCl, GaBr and GaI emission was also observed. The GaCl emission (0,O bands at 338-4 and 334.7 nm) was very weak and observable only at relatively high chloride concentrations (about M). The GaBr emission (0,O bands at 354.9 and 350.3 nm) was slightly stronger, but still too weak for analytical use. The GaI emission (0,O bands at 391.1 and 386-3 nm) was quite strong, and measurements at 391 nm (the most intense peak) can be used analytically. No GaF emission (0,O bands at 301.8, 299.0 and 298-9 nm8) was observed under any conditions. The GaCl and GaBr emission is weak because of the stable nature of gallium(II1) halide species as compared with indium(II1) halide species. The energy liberated on reduction of the gdium(II1) species is insufficient to give many excited GaCl and GaBr molecules, but it is sufficient in the case of gallium(II1) iodide to give excited GaI molecules. The production of excited gallium or indium(1) halide molecules is unlikely to be caused by the reaction of the free metal with hydrogen halide molecules, e.g., M+Hhal --+ Mhal+H(or&H,) In this type of reaction, that with hydriodic acid would be expected to be the least favourable, while that with hydrofluoric acid reaction would be the most favourable. Also the visible In1 emission occurs in the coolest central region of the flame where the incomplete evaporation of droplets is clearly observable. It is also unlikely that reduction to the metal will occur in these regions. We thank the Science Research Council for the award of a research studentship to one of us (K.C.T.), and Unicam Instruments Ltd. for the loan of the flame spectrophotometer used in this study. REFERENCES 1. 2. 4. 5. 6. 7. 8. 9. NOTE-References 1, 2 and 3 are to Parts I, I1 and 111 of this series, respectively. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1967, 92, 506. -,-,- , Ibid., 1968, 93, 72. Candler, C., “Atomic Spectra,” Hilger & Watts Ltd., London, 1964. Ezell, J. B., jun., Atomic Absorption Newsletter, 1967, 6, 84. Gilbert, P. T., jun., Analyt. Chem., 1966, 38, 1920. Cotton, F. A., and Wilkinson, G., “Advanced Inorganic Chemistry,” Interscience Publishers Inc., Pearse, R. W. B., and Gaydon, A. G., “The Identification of Molecular Spectra,” Third Edition, Mackison, R., Analyst, 1964, 89, 745. 3. -, -, -, Ibid., 1968, 93, 618. New York and London, 1962, p. 337. Chapman and Hall Ltd., London, 1966. Received September 12th, 1968 Accepted January 31st, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400643

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

ArtaZyst, August, 1969, Vol. 94, pp. 653-658 653 Spectrophotometric Determination of Micro Amounts of Nitrogen as Indophenol BY P. J. ROMMERS AND J. VISSER (Philips Research Laboratories, N . V . Philips' Gloeilampenfabrieken, Eindhoven, Netherlands) A critical study has been made of the well known spectrophotometric determination of nitrogen as ammonia, originally based on the reaction of ammonia with hypochlorite and phenol producing a blue coloured indophenol. This has resulted in a modified method with a gain in sensitivity and reproducibility. The absorptivity of indophenol found with this method is 0.60 per p.p.m. of nitrogen per cm at a wavelength of 625 nm, which is an improvement on the methods previously reported. The standard deviation is 0.0013 p.p.m. at the 0.2 p.p.m.level in 50 ml of colour developed solution. The solution obtained after colour development is also suitable for the solvent extraction with isobutyl alcohol. The method described in this paper can be applied to 0.05 to 20pg of nitrogen. PREVIOUSLY reported methods1 to 7 show considerable variation in sensitivity and repro- ducibility caused by the use of various reagents and the performance of the reactions under different conditions. These methods depend on the reaction of ammonia with hypochlorite to give chloramine, which reacts with phenol to form indophenol. This process proceeds slowly at room tem- perature but can be accelerated by heating. These reactions can be performed in either a weakly acidic or an alkaline medium, but we preferred an acidic medium to prevent the possible loss of ammonia, which might occur in the heating of an alkaline solution.Both chlorine water and sodium hypochlorite solution are used as oxidants ; however, these reagents have the disadvantage of being unstable. Namiki, Kakita and Got@ introduced chloramine-T (CHs.C,H,.S02.NClNa.3H20), which resulted in a considerable improvement because of the great stability of its solution and the ease with which it can be dissolved in well defined concentrations, so that a greater reproducibility was achieved. As a phenol-containing reagent, an aqueous solution of phenol is most suitable in an acidic medium, and in an alkaline medium a sodium phenate solution is used. The phenol solution appeared to be more stable than the phenate solution.Consequently, we chose a chloramine-T solution and a phenol solution. Colour development was carried out in a boric acid medium as reported by Bolleter, Bushman and TidweL4 In the present investigation, a study has been made of the variables involved in the reactions. Special attention has been paid to the time in which the reaction of chloramine-T with ammonia takes place. REAGENTS AND APPARATUS- Only analytical-grade chemicals were used. Some batches of phenol, which are slightly pink, give rise to a high reagent blank and therefore cannot be used. The water, obtained by the passage of de-ionised water through a Monobed mixed resin (Amberlite MB-1, analytical grade) column, was used immediately after preparation. Development of the indophenol was carried out in a thermostatically controlled water-bath at 60" C .Absorbance measurements were made with a Unicam SP500 spectrophotometer with 1, 2 and 4-cm cells. All other techniques are given below under Method. EFFECT OF THE AMOUNT OF BORIC ACID- Variations of the absorbance from the variation of the amount of boric acid solution from 5 to 25d, with the pH after colour development kept constant at 12, were within 5 per cent. EXPERIMENTAL 0 SAC and the authors.654 ROMMERS AND VISSER : SPECTROPHOTOMETRIC DETERMINATION [Analyst, Vol. 94 Without boric acid, but otherwise following the procedure as described under Method, there was a decrease in the absorbance of about 30 per cent.; however, 6 ml of boric acid was chosen as a suitable amount. It was also possible to carry out this photometric procedure immediately after an ammonia distillation.Ammonia was distilled into a 50-ml graduated flask containing 5ml of the boric acid solution, and distillation was continued until the distillate was 25 ml, then the colour development was carried out as described under Method. EFFECT OF THE AMOUNT OF PHENOL- Changes in the amount of phenol (again with the final pH maintained at 12) did not affect the absorbance more than 10 per cent. The use of a smaller amount (1 to 3 ml) of phenol, however, appeared to lead to an increase in absorbance of the reagent blank of two to four times the value obtained when 3 to 10ml of phenol were used. EFFECT OF THE AMOUNT OF CHLORAMINE-T AND OF THE TIME INTERVAL BETWEEN THE ADDITION This experiment was directed to finding a relationship between the absorbance and the amount of chloramine-T, the ammoniacal nitrogen concentration and the interval between the addition of chloramine-T and phenol. Only the amount of chloramine-T added was varied, all other variables retaining the same value as in the Procedure.Fig. 1 shows the relationship between the absorbance and the time interval for various amounts of chloramine-T added at room temperature (20pg of nitrogen). Table I shows the behaviour of the reagent blank with various amounts of chloramine-T after an interval of 10 seconds (addition of chloramine-T at room temperature). OF CHLORAMINE-T AND PHENOL- TABLE I EFFECT OF THE AMOUNT OF CHLORAMINE-T ON THE REAGENT BLANK, AFTER AN INTERVAL OF 10 SECONDS. CHLORAMINE-T ADDED AT ROOM TEMPERATURE; MEASURED IN A 2-Cm CELL AT 625 nm Chloramine-T, ml .. .. 1 2 3 4 6 10 Absorbance . . .. . . 0.005 0.005 0.005 0.006 0.007 0.010 It follows from Fig. 1 that the time interval is least critical if 4 or 6 ml of chloramine-T are used. With 5 ml or more, however, the blank begins to increase, as indicated in Table I. The effect of the interval on the absorbance, for a given amount of chloramine-T. ameared to be related to the nitrogen concentration. 01 I I 1 I I I 0 I 0 20 30 40 50 60 Time, seconds Fig. 1. Effect of the interval between the addition of cbloramine-T and phenol on absorbance; 20 pg of nitrogen in 60 ml measured in a 2-cm cell at 625 nm: A, 1 ml of chloramine-T : B, 2 ml of chloramine-T ; C, 3ml of chloramine-T; D, 4ml of chloramine-T ; E, 6 ml of chloramine-T ; F, 10ml of chloramine-T O” t O l I I 1 1 1 0 10 20 30 40 50 I Time, seconds Fig.2. Effect of the interval on the sensitivity for various amounts of nitrogen : 4 ml of chloramine-T, measured in a 4-cm cell at 625 nm. Sensitivity expressed as the absorbance per 10 pg of nitrogen: A, 10 pg of nitrogen in 50 ml (1 x the absorbance) ; B, 20 pg of nitrogen in 60 ml (& x the absorbance) ; C, 40 pg of nitrogen in 50 ml (& x the absorbance) Fig. 2 shows this relationship, with 4 ml of chloramine-T added at room temperature. The figure clearly shows the decrease in sensitivity if greater amounts of nitrogen are used. ItAugust, 19691 OF MICRO AMOUNTS OF NITROGEN AS INDOPHENOL 655 appeared that the time between the addition of chloramine-T and phenol was rendered less critical by cooling, so that the reproducibility of the procedure was increased.Fig. 3 shows this effect with 4 ml of chloramine-T and 20 pg of nitrogen in 50 ml, the chloramine-T added after the solution had been cooled to below 3" C. Thus a linear relation- ship was found between the nitrogen concentration and the absorbance up to 20pg of nitrogen in 50ml. I t is essential to add the chloramine-T before the phenol. Reversing the sequence of addition was found to decrease the absorbance to EFFECT OF THE HEATING TIME- The heating time appeared to be an important factor. Heating in a water-bath at various temperatures resulted not only in different heating times for attaining maximum absorbance but also in different sensitivities.Time, seconds ot I I I I I 0 5 10 IS 20 25 Time, minutes Fig. 3. Effect of the interval on Fig. 4. Effect of heating time on absorbance: 20 pg of nitrogen in 60 ml; absorbance: 20 pg of nitrogen in 50 ml, 4ml of chloramine-T (added in the cold) measured at 625 nm in a 4-cm cell: measured at 625 nm in a 2-cm cell A, heating at 100' C; B, heating at 60' C Fig. 4 shows this effect with 20 ,ug of nitrogen in 50 ml, after heating at 100" and 60" C. Heating in a thermostatically controlled water-bath at 60" C was found to lead to good reproducibility, while heating for 16 minutes at 60" C appeared to be a convenient time, during which a series of at least ten analyses could be performed. EFFECT OF PH- ment, as shown in Fig. 5. The absorbance of the indophenol was markedly affected by the pH after colour develop o t t t , , 0 8 9 10 I 1 I2 13 .I PH Fig. 6. Effect of pH: 20 pg of nitro- gen per 50 ml measured at 625 nm in a 2-cm cell The maximum absorbance was attained at pH 9.9 to 10.0 as previously reported by Bolleter, Bushman and Tidwell.4 However, we preferred a pH of 12 (+0-5), thus enabling656 ROMMERS AND VISSER : SPECTROPHOTOMETRIC DETERMINATION [~ndyst, VOl. 94 a solvent extraction of the indophenols to be carried out immediately after the colour develop- ment as described under Method, especially when small amounts of nitrogen are present. METHOD This corresponds to an addition of 5 ml of 3 M sodium hydroxide. REAGENTS- All reagents used were of analytical-reagent grade. Standard nitrogen solution-A 3.819-g sample of ammonium chloride that had been dried at 110" C was dissolved in water to give 1 litre of a solution containing 1 mg ml-1 of nitrogen. This stock solution was used for preparing dilute solutions.Boric acid solution, 5 9er cent. aqueous solution. Chloramine-T solution, CH,.C,H,.S02.NCINa.3H20-A 5 per cent. aqueous solution (stable for several months) was prepared. Phenol sohtion-An 8 per cent. aqueous solution (stable for several months) was prepared. Sodium hydroxide solution, 3 M. APPARATUS- Colour development was performed in a thermostatically controlled water-bath at 60" C (k 1" C). Measurements of the absorbance were made on a Unicam SP500 spectrophotometer with 1, 2 and 4-cm cells. PROCEDURE- To a solution containing between 0.05 and 20 pg of nitrogen, in the form of ammonia, in a 50-ml graduated flask, add 5.0 ml of boric acid solution and dilute the whole to about 25 ml with water.Place the flask in a bath with melting ice and allow the solution to cool until the tem- perature is below 3" C. Then remove the flask from the ice-bath and add from a fast-running pipette, with swirling, 4.0 ml of chlorarnine-T solution. After 20 seconds (f2 seconds) add rapidly, with thorough mixing, 10 ml of phenol solution and then place the flask in a water- bath at 60" C and heat for 16 minutes (k30 seconds). Remove the flask and add from a pipette, with swirling, 5.0 ml of the sodium hydroxide solution. Cool to room temperature under running tap water, then dilute the solution, make up to the mark and mix.Measure the absorbance in a cell of suitable length at 625 nm. Use a reagent blank as a reference solution. PREPARATION OF THE REAGENT BLANK- Dilute 5-0 ml of boric acid solution in a 50-ml graduated flask to about 25 ml and continue as described under Procedure. PREPARATION OF CALIBRATION GRAPH- To a 50-ml calibrated flask, containing 5.0 ml of boric acid solution, add several aliquots (up to 20 ml) of a standard ammonium chloride solution containing 1 pg ml-l of nitrogen. Dilute the whole to about 25ml with water. Thereafter follow the technique as described under Procedure. Table I1 shows the results obtained. TABLE I1 CALIBRATION RESULTS Nitrogen in 60 ml of Absorbance at Absorptivity per colour-developed solution, pg 626 nm in 2-cm cell p.p.m. per cm 1 0.024 0.600 3 0.072 0.600 6 0.120 0.600 8 0.192 0.600 10 0.238 0.696 13 0.310 0.596 16 0.355 0.692 17 0.402 0.69 1 20 0.476 0.594 26 0.578 0.578 30 0.670 0.658August, 19691 OF MICRO AMOUNTS OF NITROGEN AS INDOPHENOL 657 EXTRACTION- Namiki, Kakita and GotP reported a solvent-extraction method with isobutyl alcohol, which provided complete extraction of the indophenol.This solution has an absorption maximum at 655 nm. With this extraction used in our method a considerable lowering of the limit of determination was attained. The solution obtained after colour development in accordance with the technique described in Procedure can be used for the extraction. For this purpose the solution, after cooling, was poured into a 100-ml separating funnel, 10 ml of water being used for quantitative rinsing of the flask.Add 18 g of sodium chloride and shake the solution until most of the salt has dissolved. Add from a pipette 10ml of isobutyl alcohol and shake the solution for 1 minute. Draw off and discard the aqueous solution. Filter the isobutyl alcohol layer with a fast running filter-paper into an absorptiometric cell of a suitable length. Measure the absorbance at 625 nm with a reagent blank as reference solution. The reagent blank is obtained by shaking the blank from the Procedure with 10 ml of isobutyl alcohol for 1 minute. This extraction method is recommended for amounts of nitrogen from 0.05 to 1Opg in a final volume of 50ml. Calibration results for the extraction method are given in Table 111. TABLE I11 CALIBRATION RESULTS AFTER EXTRACTION WITH 10 ml OF ISOBUTYL ALCOHOL Nitrogen in 50 ml of Absorbance at Absorptivity per colour-developed solution, pg 655 nm p.p.m.per cm 0.25 0.059 (4-cm cell) 0.590 2 0.118 (l-cm cell) 0.590 5 0.296 (1-cm cell) 0.592 10 0.587 (l-cm cell) 0.587 DISCUSSION SENSITIVITY AND REPRODUCIBILITY- The results in Table I1 show that the calibration graph is linear in the range between 0 and 20 pg of nitrogen in 50 ml of colour-developed solution. The absorbance of the reagent blank was 0.005 at 625 nm, measured in a 2-cm cell against water. The sensitivity was compared with that of the methods previously reported. The absorptivity per p.p.m. of nitrogen per cm in the method described was 0.596. Bolleter, Bushman and T i d ~ e l l , ~ Noble,l Riley,' and Tetlow and Wilson6 gave 0*407,0.138, 0-255 and 0.436, respectively, for this value.With the extraction method the absorptivity per p.p.m. of nitrogen per cm in isobutyl alcohol was 0.590. In the method reported by Namiki, Kakita and Got6 this value was 0-106. Table I11 shows some calibration results obtained by extraction with 1Oml of isobutyl alcohol. Measurements were made with a reagent blank as the reference solution. The absorbance of the blank itself, measured against pure isobutyl alcohol, was 0.014 in a l-cm cell. The within-batch standard deviation of the absorbance of this blank was about 0-001 in a l-cm cell. By choosing three times the standard deviation of the blank as a criterion of deter- mination, this yields a lower limit of 0-05p.g of nitrogen. The reproducibility by the direct method was determined from a series of ten analyses of a given ammonia solution by two analysts each at dates 2 years apart.The pooled standard deviation of the results of both analysts was found to be 0-0013 p.p.m. at a level of 0-2 p.p.m. of nitrogen in 50ml of the colour-developed solution. STABILITY- extracted into isobutyl alcohol it is stable for at least 1 week. APPLICATION OF THE METHOD- This method has been especially developed for the determination of small amounts of nitrogen, after distillation as ammonia, to supplement the titrimetric determination of ammonia. The indophenol was found to be stable for at least 2 hours in aqueous solution. When658 ROMMERS AND VISSER The spattering of small amounts of sodium hydroxide during the distillation, which can cause slight errors in the titrimetric method, does not influence the results obtained with this photometric method.We reduced a nitrogen distillation apparatus (Messrs. Normag, Hofheim/Taunus, Germany), to half-scale and obtained good results in distillations with as little as 1 pg of nitrogen a.s ammonia. Well known methods of converting the nitrogen in a sample into the ammoniacal form (Kjeldahl digestion reduction, or by means of Devarda’s alloy, etc.) in conjunction with this method, were applied to the nitrogen determination in materials of widely varying type. We have used this method for several years and obtained good results in the following determinations: 1000 to 2000 p.p.m. of nitrogen in 10-mg samples of zirconium; 30 to 200 p.p.m.of nitrogen in 100-mg samples of iron, vanadium, niobium, magnetic alloys and oxides; 4 to 200 p.p.m. of oxides of nitrogen in 1 litre of exhaust gases; and 4 p.p.m. of nitrogen in a 400-mg sample of silicon. COLOUR REACTIONS- indophenol- Bolleter, Bushman and Tidwel14 proposed the following steps for the formation of NH, 4- HOCl (cbloramine-T) NH,CI + H,O 3 According to Tetlow and Wilson6 reaction (1) proceeds rapidly. In addition, another reaction is possible, as indicated by Mellors- However, from our experiments it appears that, at least in a boric acid medium, the reaction proceeds more slowly. NHJI + 2HOC1 --+ NCI, + 2H2O (5) after which the nitrogen trichloride can decompose into nitrogen and hydrochloric acid. This is a possible explanation of the fact that the formation of the indophenol appeared to be so dependent on the concentrations both of the ammonia and of the chloramine-T and on the time (and therefore on the temperature) in which the reaction can take place. After addition of the phenol the chloramine-T reacts with the phenol, thereby reducing its own concentration and blocking both reactions (1) and (5). At the same time this would be an explanation of the proposed order of addition of the reagents. 1. 2. 3. 4. 5. 6. 7. 8. REFERENCES Noble, E. D., Analyt. Chem., 1955, 27, 1413. Scheurer, P. G., and Smith, F., Ibid., 1955, 27, 1616. Crowther, A. B., and Large, R. S., Analyst, 1956, 81, 64. Bolleter, W. T., Bushman, C. J., and Tidwell, P. W., Analyt. Chem., 1961, 33, 592. Tetlow, J. A., and Wilson, A. L., Analyst, 1964, 89, 453. Namiki, M., Kakita, Y., and Got8, H., Talanta, 1964, 11, 813. Riley, J. P.. Analytica Chim. Ada, 1963, 9, 576. Mellor, J. W., “A Comprehensive Treatise on Inorganic and Theoretical Chemistry,” Volume VIII, Received September 9th, 1968 Accepted January 6th, 1969 Longmans, Green and Co., London, 1928, p. 698.

ISSN:0003-2654    

DOI:10.1039/AN9699400653

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, August, 1969, Vol. 94, $9. 659-663 659 The Determination of Aliphatic Aldehydes via Polarography of their Girard T Derivatives BY B. FLEET AND P. N. KELIHER (Chemisfry Department, Imfierial College, London, S . W.7) A comparison of several nucleophilic reagents for the determination of aliphatic aldehydes via polarography of their azomethine derivatives has been made. The reduction of Girard T derivatives was found to give the best defined waves under the experimental conditions used. The El/, values for the reduction of the azomethine grouping were found to fit the modified Taft relationship. THE increasing importance of volatile carbonyl compounds in studies of flavouring con- stituents, atmospheric pollution and cigarette smoke has emphasised the need for a suitable method of analysis for this group of compounds.The most generally useful reaction for the determination of carbonyl compounds is the reaction with one of a group of nucleophilic bases, the so-called carbonyl reagents. Of these, the oximation reaction1s2 has been most widely used because of the greater nucleophilic power of hydroxylamine (pKA for NH,OH+ is 6.0)., The main disadvantage of indirect titrimetric methods based on the titration of the excess of hydroxylamine is that a large excess of reagent cannot be tolerated. Intolerably long reaction times are often required to ensure complete conversion into the oxime. Gas - liquid chromatography has been used for the determination of carbonyl compounds after concentrating them from their medium as the oxime or 2,4-dinitrophenylhydrazone deriva- tives.The isolated derivatives can then be chromatographed directly on a silicone oil c01umn,4,~ or the original compounds regenerated by a "flash exchange" technique6,7,* in which a-ketoglutaric acid6 s 7 replaces the carbonyl compound from its 2,4-&nitrophenylhydrazone. The exchange reaction is carried out in a capillary tube attached to the injection port of the chromatograph, and is heated to 250" C, at which temperature the liberated carbonyl com- pound is flushed directly on to the column. The disadvantage of this technique is that isolation of the carbonyl derivative is often tedious and time consuming. Derivatives of carbonyl compounds that contain the azomethine group, -CH =N-, can also be determined polarographically.The advantage of this method is that it obviates the need for isolation of the derivatives as the reagents themselves are electro-inactive, and hence can be present in the solution in large excess. Thus the derivatives of carbonyl compounds with ammoniaJ9 510 hydroxylamine,llJ2 semicarbazide13J4J5 or a Girard reagent16 9 1 7 are more easily determined than the free carbonyl compounds themselves, where complicating side reactions such as hydration , acetal formation or enolisation tend to lower the concentration of the free carbonyl group. A detailed survey of the reduction of compounds containing the azomethine group has been carried out by Lund.l* In general, reduction of compounds RR"C=N-Y-R depends upon the nature of Y . If Y is nitrogen, then the most common course of the reduction is a 4-electron process resulting in fission of the N-Y bond. When, however, Y is a carbon atom, then a 2-electron reduction of the azomethine bond occurs.Fleet and Zuman" have confirmed this for the reduction of a wide range of aldehyde and ketone semicarbazones. With one or two isolated exceptions, reduction occurred with the uptake of four electrons to form the primary amine and urea. R'R"C=N-NHCONH, + H+ -+ R'R"C=N-NHCONH, H+ R'R"C=N-NHCONH, + 2H+ + 2e- --+ R'R"C=NH,+ + NH,CONH2 - H+ R'R"C=NH,+ + 2H+ + 2e- --+ R'R"CHNH,+. 0 SAC and the authors.660 [Analyst, Vol. 94 In almost all cases the energy levels of the two steps are so close together that a single 4-electron wave is observed. It is possible to determine a wide range of aldehydes and ketones by adding the carbonyl sample to a 0.1 M solution of semicarbazide hydrochloride containing 50 per cent.of ethanol in an acetate buffer at pH 4 _+ 0.2. The waves for the lower aliphatic aldehydes, however, are not well defined and the aim of the present work was to examine other nucleophilic reagents in order to improve the method for this class of compounds. EXPERIMENTAL REAGENTS- Carbonyl compounds were obtained commercially, and a 1 0 - 2 ~ stock solution of each compound was prepared in absolute ethanol. Sulphuric acid, acetate and phosphate buffer solutions were prepared from analytical-grade reagents and adjusted to a constant ionic strength of p = 0.5 with sodium chloride. The apparent pH values of the acetate buffers in 50 per cent.ethanol were about one pH unit higher than the measured aqueous values. Girard T reagent-Dissolve 16.7 g of Girard T reagent and 8.2 g of anhydrous sodium acetate in 50 ml of water. Add 6 ml of glacial acetic acid and dilute the solution to 100 ml in a calibrated flask. APPARATUS- A Radelkis polarograph, type OH-102 (Metrimpex, Hungary), was used to record polaro- grams by the conventional 2-electrode system. The polarographic vessel was a Kalousek cell with a separated saturated calomel reference electrode (S.C.E.) . Capillary constants measured at the potential of the S.C.E. in 0.1 M potassium chloride solution were t = 3.1 s and m = 2-30 ms-l at h = 43 cm. Half-wave potential measurements, and accurate measurements of the potential of the dropping-mercury electrode required for logarithmic analysis of the wave shape, were made by using a 3-electrode system.The potential of the dropping-mercury electrode was measured potentiometrically against a reference S.C.E. in an auxiliary circuit with a digital voltmeter (Hewlett-Packard). A modified Kalousek vessel was used, in which the junction of an auxiliary S.C.E. was immersed in the solution studied (Fig. 1). FLEET AND KELIHER: DETERMINATION OF ALIPHATIC ALDEHYDES I o r potentiometer Dropping mercury Polarographic Second reference reference electrode e I ect rode Fig. 1. Three-electrode assembly for accurate measurement of E+ The relative pH values of buffer solutions containing 50 per cent. of ethanol were measured with a Vibron pH meter, model 39A (E.I.L. Ltd., Surrey).RESULTS AND DISCUSSION Four nucleophilic reagents were examined : hydroxylamine, phenylhydrazine, trimet hyl- ammonium acetylhydrazide chloride (Girard T reagent) and pyridinium acetylhydrazide chloride (Girard P reagent). The derivatives of these compounds with propanal and heptanal were chosen for preliminary study. A series of 0.1 M solutions of each reagent, buffered at pH values as noted and containing 50 per cent. of ethanol, was prepared. An aliquot of the carbonyl compound was then added to give a concentration of the azomethine derivative in situ of 5 x 1 0 " ~ , and a polarogram recorded. This technique has the advantage that under the conditions of a large excess of reagent, the formation of the derivative is very rapid. As all of the reagents studied are electro-inactive, no interference with the wave of the carbonyl derivative occurs.The preliminary investigation showed that the Girard T derivatives gave well defined polarographic waves but the other three derivatives gave poorly defined waves similar toAugust, 19691 VIA POLAROGRAPHY OF THEIR GIRARD T DERIVATIVES 661 the semicarbazones. A more detailed investigation of the optimum conditions for measuring the Girard T derivatives was carried out. No marked improvement in the shape of the wave was obtained when the concentration of the Girard T reagent was varied from 0.01 to 1 M. The addition of potassium, lithium, magnesium and tetramethylammonium ions also had little effect. pH DEPENDENCE OF HEXANAL GIRARD HYDRAZONE- A series of constant ionic strength buffers containing 0.1 M Girard T was prepared to cover the pH range 1-68 to 11.0, and polarograms were measured after the addition of an aliquot of the standard hexanal solution to give a concentration of 5 x lo4 M.The resulting dependence is shown in Fig. 2. Potent i at Fig. 2. pH dependence of waves of 5 x ~ O - ' M hexanal Girard hydrazone in 50 per cent. ethanol: (l), pH 1-68; (2), pH 2-58; (3), pH 3.27; (4), pH 4.02; (5), pH 4.77; (6), pH 5.35; (7), pH 5-81; (8), pH 6-83; (9), pH 7-87; (lo), pH 8.7; (ll), pH 11.0. Starting potentials: (l), -0.7 V; (2, 3 and 4). -0-8V; (5 and 6). -0.9 V; (7 and 8), -1-0 V; (9 and lo), -1.1 V; (ll), -1-3 V; sensitivity 6-pA f.s.d. ; 60 mV/absc The splitting of the wave into two at low pH values is analogous to the reduction of the semicarbazone derivatives, but in this case the heights of the two parts of the wave are not equal.At higher pH values the two waves emerge but the total wave height, unlike the case with the corresponding semicarbazones,14 shows a decrease up to pH 4. The height of the wave remains constant over the pH range 4 to 7, when the conventional decrease in the form of a dissociation curve is observed. An indication of the number of electrons involved in the reduction (fin) was obtained by a comparison of the wave height with the wave for the corresponding semicarbazone derivative, for which n equals 4. This clearly showed that the wave for the Girard hydrazone over the pH range corresponded to a 4-electron reduction. Thus the increased wave height at lower pH values must correspond to a non-integral number of electrons. Cohen, Bates and Lieber- man1119 have shown that hydrazine can be liberated by solvolysis or alcoholysis of Girard P reagent.Although this reaction is much slower than with the Girard T reagent, it can be significant when large excesses of reagent are used, and at low pH values it would, therefore, contribute to an increase in wave height. The optimum pH region for the measurement of the Girard T hydrazones is, therefore, between 4 and 7. RECOMMENDED PROCEDURE- An aliquot of the sample, containing from 1 to 10 pmoles of carbonyl group, is added to 2 ml of the buffered Girard T reagent in a 10-ml calibrated flask; 0.1 ml of 0.01 per cent. Triton X-100 solution is added and the solution made up to volume so that the final solution contains 50 per cent.v/v of ethanol. The solution is then transferred to the polarographic cell and de-aerated by passage of oxygen-free nitrogen through it for 3 minutes. The polaro- gram is then recorded from zero volts versus S.C.E.662 FLEET AND KELIHER : DETERMINATION OF ALIPHATIC ALDEHYDES [A flabst, VOl. 94 Potential Fig. 3. Polarographic waves of 6 x 1 0 - 4 ~ aliphatic aldehyde Girard hydrazone in 60 per cent. ethanolic acetate buffer, pH 6-81: (l), propanal; (2), butanal; (3), pentanal; (4), hexanal; (6), heptanal; (6), octanal; (7), nonanal. Starting potential: - 1.0 V; sensitivity 6-pA f.s.d.; 60 mV/absc The polarograms obtained with the aliphatic aldehydes C, to C, are shown in Fig. 3. A rectilinear dependence of limiting diffusion current on concentration was established for the propanal and heptanal derivatives over the range 1 x lo4 to 5 x 1 0 - S ~ .The values of the diffusion current constant, I, and the half-wave potential are shown in Table I. TABLE I POLAROGRAPHIC RESULTS FOR ALKANAL GIRARD T HYDRAZONES Carbonyl compound Butanal . . .. .. Pentanal . . .. .. Hexanal .. .. .. Heptanal . . .. .. Octanal . . .. .. Nonanal .. .. .. Propanal . . .. .. Ell t - 1.209 - 1.199 - 1.178 - 1.175 - 1.166 - 1.157 - 1.160 ax* -0.1 -0.116 -0.13 -0.16 -0.16 -0.17 -0.18 Diffusion current constant I = i/cm"a t l l a 3.78 f 0.10 3-78 f 0-10 3.78 f 0.10 3.78 & 0.10 3.76 f 0.10 3.76 f 0.10 3.76 f 0.10 MECHANISM OF THE ELECTRODE PROCESS- Prelog and HafligeP studied the polarographic reduction of the Girard T derivatives of aliphatic and alicyclic ketones and proposed a 2-electron reduction to the substituted t rimethylammonium ace tyl h ydr azide .R 2H+ R\ \C=N-NH-CO-CH,N+(CH,) , + 2e- - CHNHNHCOCH,N+(CH3)s. R f / R / This was supported by Young.17 Brezina, Volkova and VolkeJ20 however, showed that the reduction of the structurally similar dimethylglycylhydrazide derivatives (Girard D reagent) occurred by a 4-electron process to form a primary amine. 4H+ \C=N-NHR+ + 4e- - CHNH,+ + H,NR. / The latter observation supports the conclusions of the detailed study by Lundl* that reduction R, of \C=N-Y-R occurs by a Li-electron process, when Y is oxygen or nitrogen. The work of Masni and Ohmori21 also clearly confirmed that the reduction occurred by a &electron process and that the earlier conclusions of Prelog and Hafliger were incorrect. The results.of the present investigation also confirmed that the reduction occurred with the uptake of four electrons rather than two. R'/August, 19691 VIA POLAROGRAPHY OF THEIR GIRARD T DERIVATIVES 663 STRUCTURAL RELATIONSHIPS- The effect of structure on the half-wave potential for the reduction of organic compounds has been studied in detail by Zuman.22s23 For aliphatic substances the effects of structure can be treated by a linear free-energy relationship, i.e., the modified Taft equation- where AE,,, is the difference between the half-wave potentials of the parent RH (where R is the reducible grouping) and the substituted RX (where X is an alkyl group), pxR is a reaction constant that relates the susceptibility of the group R to substituent effects and reaction conditions, andax* is the substituent constant that depends on the over-all elec- tronic effect of the group.A plot of El,, against substituent constants (available from standard tables) should give a straight line from which the reaction constant p can be calculated. AEW2 = pxRuX* -1.22 c ~~ -1.20 -1.18 -1.16 -1.14 -0 ~~ .09 -0.10 -0.1 I -0.12 -0.13 -0.14 - 0 1 5 -016 -0.n Taft total polar substituent constant 0,’ Fig. 4. Graph of E+ versus Taft total polar substituent constant ox* : 1, propanal; 2, butanal; 3, pentanal; 4, hexanal; 6, heptanal; 6, octanal; 7, nonanal The importance of this relationship from an analytical viewpoint is that it enables El,, values to be predicted for members of a homologous series, once the value of pxR is known.Also any deviations from linearity in a p a plot can indicate a change in the mechanism of the electrode process, which can also have important analytical consequences. The El, values measured for representative members of the n-alkanal series were found to fit the modified Taft equation (Fig. 4). The p value for the reaction is 0-376. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. REFERENCES Jencks, W. P., in Cohen, S. G., Streitwieser, A., jun., and Taft, R. W., Editors, “Progress in -, J . Amer. Chem. SOC., 1959, 81, 475. Conant, J. B., and Bartlett, P. D., Ibid., 1932, 54, 2881. Soukup, R. J., Scarpellino, R. J., and Danielczik, E., Analyt.Chem., 1964, 36, 2266. Leonard, R. E., and Kiefer, J. E., J. Gas Chromat., 1966, 142. Ralls, J. W., Anulyt. Chem., 1960, 32, 332. Keeney, M., Ibid., 1957, 29, 1489. Stephens, R. L., and Teszler, A. P., Ibid., 1960, 32, 1047. Zuman, P., Nature, 1950, 165, 486. -, Colln Czech. Chem. Commun., 1960, 15, 839. Gardner, H. J., and Georgans, W. P., J. Chem. SOC., 1956, 4180. Manousek, O., and Zuman, P., J. Electroanalyt. Chem., 1960, 1, 324. Souchay, P., and Graizon, M., Chim. Analyt., 1954, 36, 86. Fleet, B., and Zuman, P., Colln Czech. Chem. Commun., 1967, 32, 2066. Fleet, B., Analytica Chim. Acta, 1966, 36, 304. Prelog, V., and Hafliger, O., Helv. Chim. Acta, 1949, 32, 2088. Young, J. R., J . Chem. Soc., 1955, 1516. Lund, H., Acta Chem. Scand., 1969, 13, 249. Cohen, H., Bates, R. W., and Liebermann, S. J., J. Amer. Chem. SOC., 1962,74, 3938. Brezina, M., Volkova, V., and Volke. J., Colln. Czech. Chem. Commun., 1964, 19, 894. Masni, M., and Ohmori, H., Chem. Pharm. Bull., Tokyo, 1964, 12, 877. Zuman, P., in “Streitwieser, A., jun., and Taft, R. W., Editors, “Progress in Physical Organic -, “Substituent Effects in Organic Polarography,” Plenum Press, New York, 1967. Physical Organic Chemistry,” Volume 2, Interscience Publishers, New York, 1964, p. 63. Chemistry,” Volume 6, Interscience Publishers, New York, 1967, p. 81. Received December 19th. 1968 Accepted February 20th, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400659

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

664 Artalyst, August, 1969, Vol. 94, $9. 664-669 The Determination of Arsenic in Organic and Inorganic Arsenic Compounds: A Radioisotope-dilution Substoicheiometric Application BY ALPHONSO ARNOLD, SIDNEY DAVIS AND ALMETIA L. JORDAN (Division of Pharmacology and Toxicology, Food and Drug Administration, Department of Health, Education, and Welfare, Washington, D.C. 20204) A substoicheiometric isotopic-dilution method is applied to the deter- mination of microgram amounts of arsenic. The arsenic compound is sub- jected to wet acid digestion, radio-labelled with arsenic-74 standard solution, and complexed and extracted with a limiting amount of zinc diethyldithio- carbamate in chloroform. Results are given for fifty-seven determinations of nine organic compounds and sixty-eight determinations of seven inorganic compounds.Average recovery values of 99.7 f 2.7 per cent. and 98.5 f 2-6 per cent. , respectively, are included. Exclusive of digestion time, twenty replicates of a single sample can be completed in 1 hour. CLASSICAL methods for the determination of arsenic invariably involve the use of either the oxidative property of arsenite in titrimetry, or some oxidised or reduced form of the element as the basis for colorimetric determination. Although titrimetric methods are widely used to determine both macro and semi-micro levels the standard solutions may be unstable and, with iodine, accuracy may be poor at 5-mg levels of arsenite.l In one of the two official methods for the determination of arsenic2 the silver diethyl- dithiocarbamate (AgDDC) coupling of arsenic reduced to arsine is used, and in the other the molybdenum-blue colorimetric procedure.The authors, by using radio-labelled arsenic-74 as a tracer in the evaluation of both procedures, have found that the loss of arsenic is as high as 15 per cent. These losses are frequent except under the most rigidly controlled con- ditions. The slightest variations in the AgDDC procedure, e.g., in the acidity of the absorbing medium, the volume of the absorbing solution and the quality of metallic zinc used as the reductant, can cause serious errors. The molybdenum-blue method requires wet acid digestion of the sample together with a time consuming distillation step. Methods in which neutron activation3s4s6 is used have, in general, good sensitivity and obviate many of the difficulties found in the wet chemical procedures.However, the problems of access to nuclear reactors and the high cost of equipment may be disadvantages in some instances. In 1964, a substoicheiometric isotope-dilution technique3 was reported for the deter- mination of traces of arsenic after activation analysis of semi-conductor materials. In the search for a procedure that may not have the disadvantages of the methods currently in use, the simplicity of this technique, omitting the activation analysis, appealed to the authors as the basis for a rapid and satisfactory method for the deterrnination of arsenic in organic and inorganic arsenic compounds after the sample is wet oxidised. The theoretical principles of the substoicheiometric application are discussed el~ewhere.~ The principle of the method is illustrated by the following example.Radio-labelled solution containing a mass Y of metal is added to the sample solution containing mass X of the same metal. A complexing agent is added to a separate portion of the radio-labelled solution in an amount less than that which corresponds stoicheiometrically with the mass Y that it contains. The same amount of complexing agent is added to the mixture of radio-labelled and sample solutions containing mass Y + X of metal. The complexed metal from both fractions is generally isolated by 0 SAC and the authors.ARNOLD, DAVIS AND JORDAN 665 liquid - liquid extraction into a suitable solvent. The radioactivity obtained from the two solvent extracts permits the calculation of the amount of metal analysed from the relation4 x = y [& - l)] where X is the mass of metal contained in the sample, Y is the mass of metal added as tracer, A , is the activity in counts minute-l of the organic extract of the Y fraction and A , is the activity in counts minute-l of the organic extract of the Y + X fraction.The procedure outlined is an application of the technique of substoicheiometry to isotope dilution without irradiation of the samples. The main steps are as follows: preparation of the sample by wet ashing or dissolution in weakly acidic or alkaline medium; reduction of arsenic to the tervalent form with potassium iodide and ascorbic acid; dilution of the metal with a known amount of radio-labelled arsenite standard solution ; simultaneous formation and extraction of arsenic diethyldithiocarbamate by use of a limiting amount of zinc diethyl- dithiocarbamate in chloroform; determination of the radioactivity of the organic extracts by y-spectrometry; and calculation of the amount of metal determined from the relation previously described.EXPERIMENTAL APPARATUS- poration. used. y-S@ectrometer-NaI (Tl) crystal, 3 x 3 inches, from Technical Measurements Cor- Burrell mechanical shaker. Nalgene containers with covers-No. 6250, from Fisher Scientific Instruments Co., were REAGENTS- De-ionised water, previously distilled, was used in the preparation of all reagents. Arsenic-74 labelled standard sol.ution-Arsenic-74 (5 mCi) as sodium arsenate, greater than 60 Ci g-1 and containing not more than 5 pg ml-l of added arsenic carrier, was obtained from the Radiochemical Centre, Amersham, Bucks. The amount of tracer, as received, was diluted to contain about lo7 counts minute-l ml-1.To prepare the standard solution, mix 1 ml of the tracer with 0-02000 g of arsenic as the trioxide in a 100-ml graduated flask. Make the solution 6 N in hydrochloric acid, add 1 ml of 15 per cent. w/v potassium iodide, and warm to about 70" C on a hot-plate. Let the solution stand for about 30 minutes, add 1 ml of 15 per cent. w/v ascorbic acid, and dilute to the mark with water. This solution contains 200pg of arsenite ml-l and about 106 counts minute-l. Zinc diethyldithiocarbamate-Mix 0.8000 g of analytical-reagent grade zinc sulphate, ZnS0,.7H20, in 100 ml of water with 0.4000 g of analytical-reagent grade sodium diethyl- dithiocarbamate in 100 ml of water.Extract the resulting precipitate of zinc diethyldithio- carbamate into 200 ml of analytical-reagent grade chloroform (5 minutes of equilibration is optimal for complete extraction). This solution is stable for 5 days when kept stored in a dark bottle out of direct sunlight. Five millilitres of a 1 + 24 chloroform dilution is used as the limiting amount for the substoicheiometric results. N-Bromosuccinimide, 0.1 per cent. w/v. METHOD PREPARATION OF SAMPLE- From a solution of the organic arsenic compound in dilute sodium hydroxide or dilute acid, transfer an aliquot containing from 2.5 to 1Omg of arsenic to a 125-ml flat-bottomed flask. Add 10 ml of dilute nitric acid (1 + 2) and not more than 3 ml of sulphuric acid, sp.gr.1.84. Place the flask on a hot-plate and digest until fumes of sulphur trioxide appear; the digestion is complete within 30 minutes, Dilute the acid digest of the sample to 100 ml. (Dissolve inorganic arsenic compounds and make appropriate dilutions ; use aliquots of these dilutions in the determination.) The amount of arsenic digested was chosen to facilitate replicate analyses from a single acid digest.666 [Analyst, Vol. 94 SUBSTOICHEIOMETRIC ANALYSIS- Add an aliquot of the diluted sample digest, not more than 5ml and containing from 25 to 100 pg of arsenate, to a 125 or 250-ml separating funnel, labelled sample solution. Add 1 ml of 15 per cent. potassium iodide w/v, exactly 1 ml (200 pg) of the labelled arsenite standard solution and 20 ml of 6 N hydrochloric acid.Place the funnel carefully on a hot- plate and warm to about 70°C, then remove and let it stand at room temperature for 30 minutes with frequent mixing of the contents. When reduction is complete, add 1 ml of 15 per cent. ascorbic acid w/v and dilute to a total volume of 50 ml with water. Mix well. The hydrochloric acid concentration may vary from 2 to 3 N. With a second separating funnel, labelled standard solution, follow the above sequence exactly, except that the aliquot of sample digest is omitted and water is added instead. Add exactly 5 ml of a 1 + 24 chloroform dilution of the stock zinc diethyldithiocarbamate to both separating funnels. Extract the solutions simultaneously for at least 8 minutes.Allow sufficient time for good, clear phase separation. Decant the upper aqueous phase, pipette 3-ml aliquots of the two organic extracts into counting vessels and measure the y-ac t ivit y . ARNOLD et al. : DETERMINATION OF ARSENIC RESULTS Results obtained for the analysis of inorganic compounds are shown in Table I. Arsenate compounds were reduced before analysis. At 50-pg levels, the accuracy is excellent with respect to the N.B.S. certified pure arsenic trioxide. The apparent discrepancy in relative per cent. error, series 6 to 9, at the same level of determination (50pg) is indicative of the variation of individual results. The explanation for this variance is discussed below. Many determinations were made of these inorganic compounds because of the oxidising influence of the initial sample treatment.At levels of 22 to 100 pg, the over-all results are acceptable by ordinary radiochemical standards. TABLE I SUBSTOICHEIOMETRIC DETERMINATIONS OF INORGANIC ARSENIC No. of Series deter- No. Arsenic compound minations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Sodium arsenate . . .. .. .. 6 Sodium arsenate . . .. .. .. 4 Sodium arsenate . . .. .. .. 6 Sodium arsenate . . .. . . .. 3 Sodium arsenate . . .. .. .. 5 N.B.S. No. 83c arsenic trioxide . . .. 5 N.B.S. No. 83c arsenic trioxide . . .. 4 N.B.S. No. 83c arsenic trioxide . . .. 4 N.B.S. No. 83c arsenic trioxide . . .. 6 Ammonium magnesium arsenate . . 8 Nickel arsenate . . .. .. .. 5 Aluminium arsenide . . .. .. 5 Aluminium arsenate . . .. .. 5 Average standard deviation (relative per cent.) Mercury(I1) arsenate .. .. .. 5 Average recovery, per cent. . . .. Arsenic added, mg 0.0550 0.0300 0.0500 0~1000 0.0055 0.0500 0.0500 0-0500 0.0500 0.0827 0.0660 0.0220 0.0522 0.0518 Arsenic found, mg 0.0543 0.0278 0.0607 0.0989 0.0053 0*0500 0.0499 0.0494 0.0496 0.0843 0.0648 0.02 10 0.0609 0.0520 Recovery, per cent. 98-7 92.6 101.4 98.9 96.4 100.0 99.8 98.8 99.2 101.9 98.1 96.5 97.6 100.5 98.5 2.6 Relative error, per cent. 1.8 3.8 1.9 2.0 4.6 2.4 4.2 0.9 0.7 1.0 2.3 6.1 1.8 2.5 The titration of arsenite with N-bromosuccinimide was used as a standardisation method for some of the compounds by the following modifications of the Barakat and Abdalla pr0cedure.l An aliquot of the acid digest was made 6 N in hydrochloric acid and was reduced with potassium iodide solution.The reduced arsenic was transferred to a separating funnel and the excess of iodide extracted with three or four 50-ml volumes of chloroform until the aqueous phase was colourless. The use of ascorbic acid as a reducing agent caused interference in the titration. This interference was similar to that caused by sodium thiosulphate, but to a lesser degree. After the organic extracts were discarded the aqueous phase was drained into a suitable flask, neutralised with 5 N sodium hydroxide, made slightly acidic with dilute hydrochloric acid, and the analysis comp1eted.lAugust, 19691 IN ORGANIC AND INORGANIC ARSENIC COMPOUNDS TABLE I1 N-BROMOSUCCINIMIDE TITRATION RESULTS Series No. 1 2 3 4 6 6 7 8 9 No. of deter- Compound minations 4-Chlorobenzenearsonic acid 4-Chlorobenzenearsonic acid 4-Chlorobenzenearsonic acid Phenylarsine oxide .. .. Phenylarsine oxide . . .. N.B.S. No. 83c arsenic trioxide N.B.S. No. 83c arsenic trioxide N.B.S. No. 83c arsenic trioxide N.B.S. No. 83c arsenic trioxide Arsenic added, mg 0.750 1.200 1.500 4.370 12.510 2.880 5.76 8.64 0.577 0.1 per cent. reagent , * ml 1-84 2.83 3-68 0.900 29.900 6.86 13.70 20.04 1.34 Arsenic found, mg 0.770 1-180 1.540 4.160 12.630 2.88 6-73 8.40 0.560 Recovery, per cent. 102-6 98.7 102-7 99.0 100.2 100.0 99-5 97.2 97.4 667 Standard deviation, per cent. 0.6 0.29 0.33 0.27 0.12 0-17 0.10 0.42 1-80 * One millilitre of 0.1 per cent. reagent is equivalent to 0.419 mg of As3+. Table I1 shows the values obtained from the titration with N-bromosuccinimide. The compounds analysed are representative of those listed in Table 111, in terms of oxidation state.The amounts of arsenic titrated are in relation to the sensitivity of this procedure, which is at the milligram level. TABLE I11 SUBSTOICHEIOMETRIC DETERMINATIONS OF ORGANIC ARSENICALS Series No. 1 2 3 4 5 6 7 8 9 10 11 12 No. of deter- Arsenic compound mina tions 4-Nitrobenzene arsenoxide, semihydrate 4 p-Arsanilic acid . . .. .. .. 5 p-Arsanilic acid . . .. .. .. 6 4-Chlorobenzenearsonic acid . . .. 5 Triphenylarsine . . .. .. .. 5 4-(p-Aminophenylazo)phenylarsonic acid 3 4- (p-Aminopheny1azo)phenylarsonic acid 5 Triphenylmethylarsonium iodide . . 5 Tripropylammonium hexafluoroarsenate 6 4-Hydroxy-3-nitrobenzenearsonic acid . . 5 p-Nitrobenzenearsonic acid .. . . 5 p-Nitrobenzenearsonic acid . . .. 4 Average recovery, per cent. . . .. -4verage standard deviation (relative per cent.) Arsenic added , mg 0.060 0-050 0.080 0.0474 0.0489 0.050 0.100 0.045 0.0419 0.0400 0.0600 0.0500 Arsenic found, mg 0.0577 0-0536 0.0820 0.0459 0.0507 0-0499 0-0934 0.0460 0-0450 0.0390 0.0560 0-0480 Average recovery, per cent. 96.2 107.2 102.5 96.8 103.7 99.8 93.4 102.2 107-4 97.4 94.0 96.0 99.7 2.7 Relative error, per cent. 2.4 3.3 1.5 1.3 4.5 4-1 2.3 2.4 4.3 0-8 4.5 0.5 Included in Table I11 are results on the substoicheiometric analysis of nine organic arsenicals. A comparison of Table I11 with Table I shows the close agreement in accuracy and relative error between the organic compounds, 99.7 2.7 per cent., and the inorganic compounds, 98.5 2.6 per cent.The organic compounds were analysed without regard to their ultimate purity, as a critical evaluation of the discussed technique was not our objective. The slight variations from the theoretical arsenic content are, therefore, not considered indicative of the true accuracy with respect to the technique of substoicheiometry. Radio- tracer (arsenic-74) monitoring of possible digestion losses indicated consistent recoveries of more than 98 per cent. DISCUSSION The reproducibility of the extraction of arsenic with a limited amount of a complexing agent is the most critical aspect of the described technique; the experimental verification of reproducible amounts of extracted metal with ideal systems gives an indication of the precision and the accuracy that can be attained. As in theory, the optimum conditions presume an essentially complete extraction of the metal complex,4 in which the complexing agent is entirely consumed, the results of Fig.1 tend to support the conclusions of Landgrebe and McClendon6 that the mass (metal) extracted in the fraction A , and A , cannot be equal, when all the equilibrium factors are considered. Fig. 2 shows the experimental equivalence between the complexing agent and the amount of arsenic contained in the radio-labelled standard solution. The greatest accuracy is668 ARNOLD et al. : DETERMINATION OF ARSENIC [Alzalyst, VOl. 94 I 1 I 1 I I 1 I I 0 2 4 6 8 Hydrochloric acid, N Fig. 1. The extraction of tervalent arsenic as a func- tion of hydrochloric acid normality with excess of Zn(DDC), (solid line) and with a substoicbeiometric amount of reagent (broken line).A, 6 ml of stock reagent; B, 5 ml of 1 + 24 CHCI, dilution of stock reagent obtained when the substoicheiometric extraction is between 50 and 75 per cent. of the metal contained in the volume of standard solution taken. With this type of curve, a limiting volume of the complexing agent can easily be selected and the instability of the reagent on standing can be measured. G s 100- - * L aJ 0 z I I I I I s _ & I 1 I I 2 6 10 14 18 Y Stock Zn(DDC),- CHCI, solution1 ( I +24), ml Fig. 2. The extractive titration of arsenite with increasing concentration of Zn(DDC), The results from Fig. 2 form the basis for the reproducibility curves shown in Fig.3. The volumes of complexing agent chosen were 5-0 and 4-2 ml, respectively, for the indicated curves. Reproducible extraction improves at the higher concentration. Reproducibility is fair when the amount of extracted metal falls below 60 per cent. Arsenic iodide complexes would be expected to cause interference in the substoicheio- metric extraction. A study was therefore designed to determine the optimum amount of potassium iodide that can be used under the experimental conditions of this study. The resulting curve seen in Fig. 4 indicates the need for milligram amounts of iodide to reduce microgram amounts of arsenate. To minimise possible interference, an experimental equiva- lent of ascorbic acid was incorporated. In the initial studies, it became evident that the precision of the analysis was also dependent on obtaining a good statistical average of the count-rates of the fractions A , and A,. In general, the determinations were performed on five replicates of sample and standard solutions.The best precision invariably resulted from the simultaneous equilibration of the two fractions. The associated error of the mean count-rates was obtained by the conven- tional standard deviation computation. As the amount of added arsenic carrier contained in the tracer as received was low and decreased further on dilution, the error associated with the standard solution was considered negligible. The over-all error of the analysis, expressedAugust, 19691 IN ORGANIC AND INORGANIC ARSENIC COMPOUNDS 669 1 I I 1 I I I 50 I50 250 3 50 Labelled trivalent arsenic carrier, mg.Fig. 3. Reproducibility of the substoi- cheiometric extraction of arsenite from 2-9 N hydrochloric acid with a substoicheiometric amount of reagent: A, sulphuric acid; B, 5 ml and C, 4.2 ml of Zn(DDC), - CHCl, (1 + 24) Potassium iodide, mg Fig. 4. The reducing effect of potassium iodide on the quantitative extraction of 100 pg of arsenite with a slight excess of Zn(DDC), - CHC1, (10 ml of 1 + 24 stock dilution) as relative standard deviation, per cent., was obtained from the expression for the standard deviation of ratios’: Standard deviation of - A = dB2a2 + A2b2 B B2 where A is the total activity of organic extract of Y (A2); B is the total activity of organic extract of Y + X (A,) ; a is the counting error associated with A ; and b is the counting error associated with B.The described application has been restricted to arsenic in compounds of fairly simple composition. Attempts to increase the sensitivity of the analysis by decreasing the concen- tration of the complexing agent and the amount of labelled standard solution were un- successful ; the reproducibility results were erratic and unpredictable. Several causes may be cited, e.g., the absence of a blank correction, which is a severe limitation on the sensitivity at low levels, and the instability of the reagent at low dilution. It has been shown* that high sensitivity can be obtained when an ideal system is studied, with the ratio of metal determined ( X ) to the amount contained in the standard solution (Y) as 1 : 1.In this connection the technique may receive use as a standardisation procedure. To extend the method to biological materials or other higher cationic media would certainly require separations of the complexes (cationic and anionic species), which would react with the dithiocarbamate radical. Such isolation procedures would probably diminish the speed and simplicity of the method without a proportionate gain in sensitivity, especially at the low levels of trace metals usually present in biological materials. The authors acknowledge the assistance of Mr. J. Van Dyke of the Division of Statistical Services, Food and Drug Administration, and Mr. George E. Story, Intermediary Metabolism Branch, Food and Drug Administration, for the use of some of the compounds analysed. REFERENCES 1. 2. 3. 4. 6. 6. 7. 8. Barakat, M. Z., and Abdalla, A., Analyst, 1960, 85, 288. “Official Methods of Analysis,” Tenth Edition, Association of Official Agricultural Chemists, Zeman, A., RBZiCka, J., Starf, J., and KleCkovd, E., Talanta, 1964, 11, 1143. RbZiCka, J., and Starf, J., Institute of Nuclear Research, Report 992/63, Rez, 1963. Grimanis, A. P., and Souliotis. A. G., Analyst, 1967, 92, 549. Landgrebe, A. R., and McClendon, L. T., N.B.S. Technical Note No. 404, July, 196!5-June, 1966, Mulligan, F., and Wormall, A., “Isotopic Tracers,” University of London, The Athlone Press, Landgrebe, A. R., McClendon, L. T., and Devoe, J. R., “Radiochemical Methods of Analysis,” Received July 18th. 1968 Accepted January 15th. 1969 Washington, D.C., 1965, Section 24.011, p. 357. p. 170. 1954, p. 295. International Atomic Energy Agency, Vienna, 1965, Volume 11, p. 321.

ISSN:0003-2654    

DOI:10.1039/AN9699400664

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

670 *Analyst, August, 1969, Vol. 94, ++. 670-673 The Gas-chromatographic Determination of 2-Chloro-4-nitrobenzamide, 3,s-Dinitrobenzamide and 3,S-Dinitro-o -toluamide in Animal Feeding Stuffs BY R. A. HOODLESS AND R. E. WESTON (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) A method is described for the determination of 2-chloro-knitrobenzamide (aklomide), 3,5-dinitrobenzamide (DNBA) and 3,6-dinitro-o-toluamide (zoa- lene) in animal feeding stuffs. These additives are extracted from the feed with methanol and the extract is treated with methanol - hydrogen chloride solution to form the methyl esters of the corresponding substituted benzoic acids. These are determined by gas chromatography with an electron-capture detector.DNBA and zoalene have similar retention times in the prescribed procedure, so that some method by which they can be distinguished is desirable. A modification of the method for this purpose is proposed. 2-CHLORO-4-NITROBENZAMIDE (aklomide), 3,5-&nitrobenzarnide (DNBA) and 3,5-dinitr0- o-toluamide (zoalene) are incorporated in poultry feeds to provide a measure of prophylaxis against certain poultry infections. The methods that have been described for DNBA1s2 and zoalene3s4 are principally colorimetric and are not necessarily specific for the additive in question. Interference can arise from chromogenic material of the feed or from other additives that may be present, and it is useful to exploit an alternative method when the additive content of a feed requires confirmation.Gas chromatography was selected for this purpose and a high degree of sensitivity is ensured by the use of an electron-capture detector. Direct gas chromatography of the pure drugs proved unsatisfactory, probably on account of their low volatility. As these three drugs are acid amides, hydrolysis of the amide linkage is easily accomplished. Subsequent esterification of the free acid thus formed yields deriva- tives that can be readily chromatographed. These steps can be carried out in one operation by treating the drugs with a solution of hydrogen chloride in methanol, by which the methyl esters of the corresponding benzoic acids are obtained. It has been reported that ortho- substituted benzoic acids are esterified only with difficulty,s but we have found that,with the modification of the Fischer - Speier technique adopted, a satisfactory yield of the methyl ester is obtained.In the procedure described, esterification does not proceed to completion and it is, therefore, important to treat both sample and standard solutions in the same manner. AS the retention times for the methyl esters of 3,5-dinitrobenzamide and 3,5-dinitro- o-toluamide are similar, it may be necessary to distinguish between them. This can be achieved by a preliminary clean-up of the additive extract from the feed by the use of an alumina column, followed by treatment with alkaline permanganate before proceeding with esterification and gas chromatography. When treated in this way the methylated derivative of 3,5-dinitrobenzamide gives a response, whereas the oxidised product of 3,5-dinitro-o-tolu- amide does not. The alumina-column clean-up procedure is necessary to remove interference arising from other feed extractives that react with alkaline permanganate.As an alternative, the characterisation of the methyl esters of these additives was attempted by the use of the “p”-value technique of Beroza and Bowman,6 but without success. 0 SAC; Crown Copyright Reserved.HOODLESS AND WESTON 671 EXPERIMENTAL APPARATUS- An isothermal gas chromatograph, fitted with an electron-capture detector, was used, and comprised a 60-cm glass column, 0.d. 6 mm and i.d. 3 mm; the column packing was silanised Chromosorb G, 100 to 120 mesh, coated with 1 per cent. Apiezon L and 0.2 per cent. Epikote resin 1001.Condition overnight at 200” C with agas flow-rate of 100 ml minute-l. The column characteristics were temperature, 160” C; carrier gas, nitrogen; flow-rate, 100 ml minute-’; and applied detector potential, 20 V. bottom to i.d. of 3 mm. REAGENTS- Injection syringe-Hamilton, 0 to 10p1, fitted with guides. Chromatographic columns-About 25 cm in length and 9 mm i.d., constricted at the All reagents should be of analytical-reagent grade quality. Alumina-Woelm neutral Brockman No. 1. Stir 100 g with 200 ml of water for 5 minutes. Allow the mixture to settle and decant off the supernatant liquid. Repeat twice more with 100-ml portions of water. Collect the slurry on filter-paper in a Buchner funnel and wash the alumina on the filter with two 50-ml portions of methanol.Air dry under reduced pressure. Diethyl ether-Peroxide-free, re-distilled. Methanol - hydrogen chloride solution-Pass hydrogen chloride into methanol until there is a 10 per cent. increase in weight. Potassium permanganate solzction, 0.1 N. Sodium carbonate solution, 0.1 N. Sodium hydrogen carbonate solution-Prepare a saturated solution in water. Sodium sulphate, anhydrous, granular. Sodium sulphite sohtion-Dissolve 20 g of sodium sulphite, Na2S0,.7H,O, in water and dilute to 100 ml. This reagent should be freshly prepared. PROCEDURE- Extraction and esteriJication-Transfer 10 g of the feeding stuff to a 250-ml conical flask and add 100 ml of methanol with a pipette. Stir the mixture for 1 hour, taking precautions against loss of methanol by evaporation.Decant the mixture into a dry centrifuge tube and spin it for about 5 minutes at 2000 r.p.m. Transfer by pipette a 2-ml aliquot of the super- natant liquid to a 50-ml round-bottomed flask. Add 10 ml of methanol - hydrogen chloride solution and boil it for at least 1 hour under a reflux condenser fitted with a calcium chloride guard tube. (If zoalene is known to be present, it is advantageous to extend the time of refluxing to 2 hours to ensure an improved yield of the methyl esters.) Allow the mixture to cool and transfer the contents of the flask to a 100-ml separating funnel. Wash the flask with 10ml of sodium hydrogen carbonate solution and add this to the separating funnel. Repeat the washing with another 10-ml portion of saturated sodium hydrogen carbonate solution.Extract the mixture with three 10-ml portions of diethyl ether and filter each ether extract through 5 g of granular anhydrous sodium sulphate into a 25-ml calibrated flask. Wash the sodium sulphate with a small portion of ether, add the washings to the flask and dilute to volume with ether. Dilute this solution as necessary with ether. Prepare standard solutions of the feed additives in methanol in concentrations com- parable with those expected to occur in the rnethanolic extract of the feeding stuff. Take 2-ml aliquots of these solutions and treat them in exactly the same manner as the sample extract as described above, commencing “Transfer by pipette a 2-ml aliquot of the supernatant liquid to a 50-ml round-bottomed flask. . . .” Gas chronzatograflhy-Dilute the ether solutions as necessary so that the content of the given additive in 5 pl is within the range specified in Table I.Inject 5 pl of the standard solution and record the peak height. Then inject 5 pl of the sample solution and compare the peak heights. If the peak heights of the standard and sample solutions are not directly comparable, the concentration of the sample solution should be adjusted to make them so. The injection must be either into the top of the column or directly “on column.” The ranges of linear response for the prophylactics and their retention time relative to lindane are shown in Table I.672 HOODLESS AND WESTON : GAS-CHROMATOGRAPHIC DETERMINATION [Analyst, Vol. 94 TABLE I RESPONSE AND RELATIVE RETENTION TIME RESULTS Substance Relative retention time of derivative 100 Lindane .... Aklomide .. .. 0-2 to 1.6 36 DNBA .. .. 0.8 to 4.8 95 Zoalene .. .. 1 to 10 104 Range of linear response, ng - The repeatability of the esterification procedure was tested by refluxing five separate 20-pg amounts of each prophylactic with methanol - hydrogen chloride solution and extracting the ester produced as described above in the procedure. The results are shown in Table 11. TABLE I1 REPEATABILITY OF ESTERIFICATION Aklomide .. . . 50,61,49,4a, 50 Prophylactic Peak height, mm DNBA .. .. . . 56, 56, 53, 53, 55 Zoalene . . .. . . 50, 47, 49, 49, 50 DISCRIMINATION BETWEEN DNBA AND ZOALENE- Prepare a chromatographic column by plugging one end of a clean dry tube with cotton- wool. Transfer 5 g of the prepared alumina to the tube and pack by gently tapping against the side of the tube.Transfer 5ml of the methanolic feed extract solution, which should contain at least 50 pg of the prophylactic drug, to the column and allow the liquid to pass through by gravity, collecting the eluate in a 50-ml round-bottomed flask. Wash the column with two 5-ml portions of methanol, allow the column to drain after each portion and collect the portions in the 50-ml flask. Evaporate the combined eluates to dryness on a water-bath. Add 5 ml of 0.1 N potassium permanganate followed by 5 ml of 0.1 N sodium carbonate and boil under a reflux condenser for 30 minutes. Allow the mixture to cool and transfer the con- tents of the flask to a 100-ml separating funnel. Wash the flask with 2 ml of N hydrochloric acid and add this to the separating funnel.Repeat with 2 ml of sodium sulphite solution. Extract the mixture with three 10-ml portions of diethyl ether and collect the extracts in a 50-ml round-bottomed flask. Evaporate the combined ether extracts to dryness, add 10 ml of methanol - hydrogen chloride solution to the residue and boil for 1 hour under a reflux condenser fitted with a calcium chloride guard tube. Allow it to cool and transfer the contents of the flask to a separating funnel. Wash the flask with 10 ml of saturated sodium hydrogen carbonate solution and add this to the separating funnel. Repeat with another 10-ml portion of sodium hydrogen carbonate solution. Extract the mixture with three 10-ml portions of diethyl ether and filter each ether extract through 5 g of granular sodium sulphate into a 25-ml calibrated flask.Wash the sodium sulphate with a small volume of ether, add the washings to the flask and dilute to volume with ether. Inject 5 pl of this solution into a gas chromatograph adjusted to the same sensitivity setting as used for zoalene. Feed extracts containing zoalene give no response when this procedure is followed, but a peak of reduced intensity is given when dinitrobenzamide is present. RESULTS The recovery of these prophylactic drugs in a number of poultry feeds was checked by adding known volumes of the drug solutions containing 0.5 mg ml-l to weighed amounts of feed. The feed was then extracted with methanol. The results are shown in Table 111. The recoveries compare favourably with those obtained by alternative methods.INTERFERENCE BY OTHER ADDITIVES- Amprolium, acinitrazole, aminonitrothiazole, acetyl-(#-nitropheny1)-sulphanilamide, furazolidone, meticlorpindol, nitrofurazone and sulphaquinoxaline were added to feeds at a level of 500 p.p.m. and taken through the method described, and no interference in the determination of aklomide, DNBA or zoalene was observed. Dimetridazole, however, interferes in the determination of aklomide and to a slight extent in the determination of DNBA and zoalene.August, 19691 OF Z-CHLORO-~-NITROBENZAMIDE IN ANIMAL FEEDING STUFFS TABLE I11 RECOVERY OF ADDITIVES FROM POULTRY FEEDS 673 Prophylactic Feed Aklomide . . .. Layers mash Intensive layers pellets Hen battery mash DNBA .. .... Layers mash Intensive layers pellets Hen battery mash Zoalene . . .. . . Prepared feed Layers mash Hen battery mash Turkey fattening mash Weight added, per cent. 0~0100 0~0100 0~0100 0~0100 0.0200 0*0200 0~0100 0~0100 0*0100 0~0100 0.0200 0.0200 0.0200 0~0200 0*0100 0~0100 0.0200 0*0200 0*0100 0~0100 0*0100 0~0100 0.0200 0*0200 0~0100 0~0100 0*0200 0.0200 0~0100 0~0100 0~0100 0~0100 0.0060 0.0060 0.0050 0.0060 0~0100 0~0100 0.0 100 0~0100 0.0050 0.0050 0~0100 0~0100 0~0100 0*0100 Weight found, per cent. 0.009 1 0.0091 0.01 13 0.0104 0.0208 0.0242 0.0096 0.0082 0.0104 0~0100 0.0208 0.0148 0.0188 0*0180 0.0097 0.0091 0.0188 0.0188 0-01 LO 0.0108 0.0098 0-0089 0.0178 0.0186 0.01 18 0.0118 0.0191 0.0182 0.0096 0.0126 0.0112 0*0100 0.0049 0.0056 0.0052 0.0063 0.0097 0~0090 0.0106 0.0110 0.0066 0.0056 0*0100 0.0096 0.0109 0-0104 This paper is published with the permission of the Government Chemist. REFERENCES 1. 2. 3. 4. 6. 6. Cavett, J. W., and Hoetis, J. P., J , Ass. Off. Agric. Chem., 1969, 42, 239. Daftios, A. C., and Schall, E. D., Ibzd., 1962, 45, 291. Getzendaner, M. E., and Gamer, W. L., Ibid., 1961, 44, 18. Smith, G. N., Down to Eavth, 1962, 18, 13. Ingold, C. K., “Structure and Mechanism in Organic Chemistry,” Bell & Sons Ltd., London, Beroza, M., and Bowman, M. C., Analyt. Chem.. 1966, 37, 291. Received December 2nd, 1968 Accepted February 7th, 1969 1963, p. 777.

ISSN:0003-2654    

DOI:10.1039/AN9699400670

出版商:RSC    

年代:1969

数据来源: RSC