ISSN:0003-2654    

DOI:10.1039/AN95479FX033

出版商:RSC    

年代:1954

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95479BX035

出版商:RSC    

年代:1954

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95479FP087

出版商:RSC    

年代:1954

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95479BP095

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

J ULY, 1954 THE ANALYST Vol. 79, No. 940 PROCEEDINGS OF THE SOCIETY FOR ANALYTICAL CHEMISTRY DEATH Alfred Randolph Campbell. WE regret to record the death of MICROCHEMISTRY GROUP AN Ordinary Meeting of the Group was held jointly with the London Section of the Royal Institute of Chemistry at 5.45 p.m. on Friday, May 7th, 1954, at the University, Reading. Dr. A. M. Ward, F.R.I.C., was in the Chair. Demonstrations were given in the Analytical Chemistry Research Laboratory by I. Smith and J. B. Jepson on “New Methods in Paper Chromatography” and by J. Tinsley on “Shandon Continuous Paper-electrophoresis Apparatus.” The following papers were then presented and discussed in the Zoological Lecture Theatre (see summaries below) :-“The Determination of Esterases,” by W. N. Aldridge, B.Sc., Ph.D.; “The Determination of Sugars,” b.y G. Harris, Ph.D., A.R.C.S. ; “The Measurement of Isotopes of Carbon and Hydrogen,” by R. F. Glascock, B.Sc., Ph.D. During the afternoon, visits were made to Messrs. Huntley and Palmers Ltd., and to the National Institute for Research in Dairying. THE DETERMINATION OF ESTERASES DR. W. N. ALDRIDGE said that the two most important methods of determining esterases were those involving manometric techniques and those making use of substrates which, on hydrolysis, produced substances that could easily be determined. In the manometric technique, the acid produced on the hydrolysis of the ester liberated carbon dioxide from a bicarbonate buffer. In the colorimetric techniques, esters containing phenols, phenolphthalein, indoxyl esters, and so on, had been used.It had long been known that esterases had a low substrate specificity. This had led many workers to assume that they had no specificity and that any ester that might be synthesised would be hydrolysed by the esterase. This was quite untrue; nor could it be assumed that the hydrolysis of a particular ester was carried out by one enzyme only. Experiments had been carried out by the author with organo-phosphorus com- pounds. These had shown that, in rabbit serum, there were two types of esterase that hydrolysed 9-nitrophenyl and phenyl esters, and that in rat intestinal mucosa there were two esterases that hydrolysed tributyrin. One of these was commonly called lipase. Organo-phosphorus compounds had proved to be extremely useful, and their use would undoubtedly increase the knowledge of this group of enzymes. It was probably true to say that the lack of knowledge of their function was due mainly to the fact that the activities of the different esterases had not yet been separated, nor their properties examined.THE DETERMINATION OF SUGARS DR. G. HARRIS said that the need for a precise and versatile method of analysing the carbohydrates in extracts of the raw materials for brewing had been largely met by the use of paper chromatography. Modifications of the quantitative procedures elaborated by N. Nelson ( J . Biol. Chem., 1944, 153, 375) and A. E. Flood, E. L. Hirst 393394 PROCEEDINGS [Vol. 79 and J. K. N. Jones (Nature, 1947, 160, 86; J . Chem. SOC., 1948, 1679), who used the solvent mixtures ethyl acetate - pyridine - water and ethyl acetate - acetic acid - water, as recommended by M.A. Jennyn and F. A. Isherwood (Biochem. J., 1949, 44, 402), were described and illustrated by the results of analyses of brewer’s wort, barley extracts and- beers for their constituent sugars. The separation of the sugars in mixtures of di-, tri-, tetra- and pentasaccharides, as in wort, was best effected with ethyl acetate - pyridine - water, or ut-propanol - ethyl acetate - water (Albon, N., and Gross, D., Analyst, 1962,77,410), but the separation of monosaccharides from one another was best achieved with ethyl acetate - acetic acid - water. Each sugar was determined after location on (Harris, G., and MacWilliam, I. C., Chem. Ind., 1954, 249) and elution from the paper chromatogram by means of its reduction of the copper ion in Somogyi’s reagent (Nelson, reference above; Somogyi, M ., J . Biol. Chem., 1952,195, 19), followed by measure- ment of the molybdenum-blue colour formed on addition of Nelson’s arsenomolybdate reagent and comparison of this colour with that given by a known amount of a standard sugar added to the solution before chromatography. The amounts of each sugar in each of the eluates from the paper chromatogram were found by reference to standard curves showing the relationship between weight of sugar (0 to 200 pg) and colour intensity of the molybdenum blue. Standard curves for glucose, fructose, arabinose, xylose, ribose, sucrose, glucodifructose, maltose, maltotriose, maltotetraose and maltopentaose were shown.Of these sugars, xylose and ribose had been found suitable as added internal standards for wort and beer analyses, and ribose for the analysis of extracts of barley. The standard errors of the determinations of the individual sugars were discussed; these were not greater than -+3 per cent. for sugars present in amounts of 100 to 300 pg on the chromatograms, but rose to 10 per cent. for those present in very small amounts, e.g., 10 to 30 pg. The defects in the method described above were discussed, the principal objection to it arising from the fact that many polymeric sugars gave little or no reduction of the cupric ion, and must be hydrolysed to monosaccharides before determination. The use of the anthrone-sulphuric acid reagent of R.Dreywood (Ind. Eng. Chem., Anal. Ed., 1946, 18, 499), as modified by D. L. Morris (Science, 1948, 107, 254), overcame this difficulty, as the hydrolysis and colour development were effected in one step. It suffered from the disadvantage, however, that the colours given by pentoses and pentosans were transient and weak, and the use of these materials as internal standards was in- admissible. A compact apparatus designed by W. D. McFarlane and H. R. Held (Proc. Ear. Brew. Conv., Nice, 1953, p. 110) for chromatography was described, and modifications of procedure recommended by Mr. Hall at the Brewing Industry Research Foundation were discussed. Typical results by the method were shown by reference to the analysis of wort, the standard errors of the estimations of the individual sugars being of the same magnitude as for the method previously described.Some possible future trends in the development of carbohydrate analysis were discussed. THE MEASUREMENT OF ISOTOPES OF CARBON AND HYDROGEN DR. R. F. GLASCOCK first briefly surveyed the methods available for counting the radioactive isotopes carbon-14 and tritium in solid compounds, and then proceeded to describe in greater detail the more efficient gas-counting methods currently used in the Isotope Section at the National Institute for Research in Dairying. The three isotopes carbon-13, carbon-14 and tritium could be determined in a single 5 or 10-mg sample of organic material (Glascock, Biochem. J . , 1952, 52, 699). By the use of high-vacuum techniques, the sample was burnt quantitatively, the products were trapped in liquid- air chilled traps and separated by subliming away the carbon dioxide at -78” C.This was measured manometrically, so providing, if desired, a measure of the carbon content of the sample. Part was set aside for the mass-spectrometric determination of carbon-13, and measured portions of the remainder were introduced into a gas counter for radioactivity determinations. The gas-counting method used was that of S. C. Brown and W. W. Miller (Rev. Sci. Instrum., 1947, 18, 496) : the counter was filled with carbon dioxide at a pressure of 20 cm of mercury and carbon disulphide at 2 cm, the partial pressure of carbon dioxide being adjusted with inactive carbon dioxide after the addition Qf the active sample. The counter was operated in the Geiger region.July, 19541 PROCEEDINGS 395 Although tritium could be measured in a proportional gas counter in the form of hydrogen alone, hydrogen plus methane or as tritiomethane, these gases were subject to several disadvantages.Hydrogen was not condensed at liquid-air temperature and tritium - hydrogen contaminated counters by exchange, either when used alone or in admixture with methane, giving rise to elevated backgrounds-the well-known “memory effect.” Methane, although it did not contaminate counters, was also incondensable and was therefore somewhat difficult to manipulate quantitatively. Butane, however, was a good counting gas in the Geiger region and was fully condensed at liquid-air temperature. In the method described by the author (Nucleonics, 1951, 9, No.5, 28), n-butane was prepared on the 5 to 10-mg scale from water by making it react with dry n-butyl magnesium bromide. With freshly prepared reagent the yield was almost quantitative, and it had been shown that no isotopic fractionation occurred during the reaction. Measured portions of the resulting butane were condensed into a gas counter, the pressure was adjusted to 126cm with inactive butane and the counter was operated in the Geiger region. A slide was shown to illustrate the accuracy and reproducibility of the method. Yields of carbon dioxide from pure compounds were within 1 per cent. of theoretical, and replicate counting and carbon-13 results agreed to within & l a 5 per cent. PHYSICAL METHODS GROUP THE Forty-sixth Ordinary Meeting of the Group was held at 6.30 p.m.on Friday, May 28th, 1954, in the Harris Institute, Preston, Lancs. It was a joint meeting with the Liverpool and District Section of the Royal Institute of Chemistry, and was preceded by a visit to the Works of Siemens Lamps Ltd. The following papers on “Fluorimetry” were presented and discussed :-“A Twin-beam Null-point Fluorimeter,” by J. P. Dowdall, BSc., A.R.C.S., D.I.C., A.R.I.C., and H. Stretch, A.R.I.C.; “The Quenching of the Fluorescence of Traces of Uranium in Fused Sodium Fluoride by Iron and Plutonium,” by G. N. Walton, M.A., B.Sc. Mr. A. A. Smales, B.Sc., F.R.I.C., was in the Chair. BIOLOGICAL METHODS GROUP THE Summer Meeting of the Group was held on Wednesday, June 16th, 1954, at the National Institute for Research in Dairying, Shinfield, nr. Reading, Berks. Dr. S. K. Kon was in the Chair for the morning session, at which the following papers were presented and discussed :-“The Estimation of Depolarising Substances by Use of the Isolated Semispinalis Cervicis Muscle of the Chick,” by K. J. Child, B.Pharm., A.R.I.C., and Eleanor J. Zaimis, M.D., BSc.; “Use of the Frog’s Rectws Abdominis for Assay of Laudexium and Suxamethonium,” by G. B. Chesher, BSc., and H. 0. J. Collier, B.A., Ph.I>. A visit to the laboratory of J. E. Ford and M. E. Gregory followed, to see and discuss the research in progress on the bio-assay of cobalamins. The proceedings terminated at 4.50 p.m. with a vote of thanks moved by Dr. Collier to Professor Kay and his staff for their generous hospitality. After lunch a tour was made of the Institute’s laboratories.

ISSN:0003-2654    

DOI:10.1039/AN9547900393

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

396 OBITUARY [Vol. 79 Obituary SAMUEL ERNEST MELLING THE sudden death of SAMUEL ERNEST MELLING on April 17th has severed almost the last direct link between the contemporary generation of analytical chemists and those legendary pioneers who founded this Society and who first directed it on its so successfui course. Melling was trained in the laboratory of the late A. H. Allen of Sheffield. That must, indeed, have been a remarkable school that could turn out so many men of Melling’s calibre, men who attained and maintained eminence and honour in an exacting profession. Melling was proud of being a “Surrey Street” man. He was quietly proud, too, in a natively whimsical way, of the Lancashire town of Wigan where he was born in 1877, and for which he became Public Analyst-his first of many public appointments-nearly forty years ago.He left Allen’s laboratory in 1903 to become partner of the late James Carter Bell, a contemporary and friend of Allen. The appointment for Wigan followed, and thereafter the laboratory at Higher Broughton, with its quaint air of isolation from the strain and urgency of city life, became the centre of one of the busiest practices of analytical chemists in the North of England. Carter Bell died in 1913, and Melling carried on alone until, in 1921, he was joined by the late Edward Ardern. Some of Ardern’s outstanding pioneer work on sewage purification had been shared and supplemented by Melling, and the partnership was the natural outcome of the association of two contrasting, but essentially complementary, personalities. With justifiable pride, since it was his life’s work and the measure of the success of unremitting effort, Melling watched the practice grow until “Melling and Ardern” became acknowledged as one of the leading firms of consulting chemists in the country.Indeed, he hoped it would be his monument, and it was characteristic that in the last year of his life, when he knew his strength was failing, he undertook an elaborate programme of laboratory reconstruction and re-organisation from which, at the age of 76, a less courageous spirit might well have recoiled. Melling’s professional eminence was recognised by appointment to high offices in many societies, and he regarded as the supreme honour of his career his election to the Presidency of this Society in 194345. But an obituary notice cannot be a biography, and it is not for these things that he will be most remembered by his friends.A vivid personality that never failed in any company to be the centre and focus of a group, a master of anecdote, a shrewd but never malicious critic, a histrionic talent that was unsuppressed because it was largely unconscious, and over everything a kindly nature and a disarming and unexpected humility, these made the good companion that Melling so undoubtedly was. To those who knew him well he was more. He was a prodigious worker who spared neither himself nor his associates, a stern critic, a generous opponent, a loyal friend. Above all, he was a deeply religious man, with an unfashionably sincere conviction of the predominant value of spiritual things.Mrs. Melling, a charming and devoted wife, predeceased him by a few months, after several years of failing health. There were no children of the marriage. It is the modern tragedy that a generation of scientists that is obsolescent at the age of thirty must concentrate all its energy on vocational training and cannot afford the luxury of education. Melling, born in less clamant times, enjoyed a breadth of culture too often denied to his less fortunate successors; his life was the fuller for it. He “ceased upon the midnight with no pain,” and we who are left, acquiescent or active in tearing veil after merciful veil from unimaginable horror, might well pause and sigh, not for the passing of an individual, but for ourselves. J. G.SHERRATT CRESSACRE GEORGE MOOR CRESSACRE GEORGE MOOR, who died on February 8th, 1954, shortly before his 86th birthday, had become a prominent figure among analytical chemists at a remarkably early age. Moor spent two years at Cambridge and two at Oxford studying under Vernon-Harcourt, and became M.A. Cambridge. He was one of the band of workers inspired by A. H. Allen, and some of his first papers were published under their joint names, on the composition of vinegar (Apzallyst, 1893), the detection of exhausted ginger, and changes in butter on storage (AnaZyst, 1894). Before this, in 1882, howevcr, lie had read a paper to the British AssociationJuly, 19541 THE DETERMIXATION OF LEAD IN FOODSTUFFS 397 on sewage disposal. He also practised at various times in association with William Chattaway, T.Peannain, Martin Priest and W. Partridge, at first in London, then both in London and Exeter, and finally again in London. During these early years he was a prolific author, publishing numerous articles, many of which appeared in The Analyst, either alone or in conjunction with others, but his name will live chiefly on account of the books of which he was author or joint author. The two best known of these are “Aids to the Analysis of Foods and Drugs,” now edited by Dr. J. R. Nicholls, and “Aids to Bacteriology,” now edited by Dr. H. W. Scott-Wilson. Other books were “Applied Bacteriology,” “The Chemical and Bacteriological Examination of Water” and “Suggested Standards of Purity for Foods and Drugs.” In the early 1890’s he became Public Analyst for Exeter, but his allegiance to chemistry was always a divided one, and he resigned the appointment and closed his laboratory in 1901 to go to West Africa prospecting for gold. This venture in mining did not last long, as he was soon back in London to rejoin Chattaway’s old laboratory and was appointed Public Analyst for Dorset, Poole and Penzance.During the 1914-18 war he served in the R.A.M.C. as a Captain and, although he remained associated with the practice of Moor and Partridge, he never returned to the laboratory, his natural wanderlust again taking him to Africa to engage in mining, a pursuit in which he never lost interest and on which he had written a book in 1910. He returned to England in the early 1930’s and again joined Priest, who was then Public Analyst for Camberwell, in his laboratory in Lewisham, but abandoned this once again to return to mining in Cornwall. Moor was elected a member of the Society on January 4th, 1893, and served on the Council, 1898-99. He was elected a Fellow of the Institute of Chemistry in 1898. T. MCLACHLAN

ISSN:0003-2654    

DOI:10.1039/AN9547900396

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

July, 19541 THE DETERMIXATION OF LEAD IN FOODSTUFFS 397 Analytical Methods Committee REPORT OF THE LEAD PANEL OF THE METALLIC IMPURITIES IN FOODSTUFFS SUB-COMMITTEE The Determination of Lead in Foodstuffs THE Analytical Methods Committee has received from its Metallic Impurities in Foodstuffs Sub-committee the following report of its Lead Panel. The Report has been approved by the Analytical Methods Committee and its publication has been authorised by the Council, REPORT The Lead Panel was constituted on February 6th, 1950, with the terms of reference: “to formulate standard and routine methods for the Determination of Lead in Foodstuffs,” The purpose of this appointment was in effect to ascertain the precision and accuracy of methods then generally in use, having regard to the possibility of statutory limits of 1 or 2 parts per million, or of even less than 1 p.p.m., being recommended by the Ministry of Food.At its first meeting the Panel was of the general opinion that the procedure to be adopted should be the development, if possible, of a rapid method, and then of a standard or referee method. For this purpose, samples of fruit syrup and cocoa containing, respectively, about 5 parts and 1 part per million of lead were circulated together with drafts of two methods of analysis, one involving the use of dithizone and the other sulphide for the final assessment. The results proved to be disappointing. Accordingly, a further series of tests with fruit juice and cocoa was undertaken, these two methods again being used, together with a third method that incorporated the colloidal sulphide procedure with the final deter- mination with dithizone as the reagent; in this third method certain precautions were taken, for which the need had been demonstrated by the first test.Again the results were disappointing. The position was then reviewed, and the earlier decision to develop a rapid method before developing a standard method was reversed, largely because it was agreed that the necessity for examining and endeavouring to overcome the many difficulties398 THE DETERMINATION OF LEAD IN FOODSTUFFS [Vol. 79 that had arisen in the collaborative work to date clearly indicated a standard method as the objective. The next test was with samples of syrup, salad cream and beer, by a modified dithizone method elaborated by two members of the Committee.In addition, an entrainment method for the beer was outlined. After this test, members of the Panel undertook intensive examina- tion of certain aspects of the method, and then two samples of syrup containing, respectively, 1-5 and 2.5 parts per million of lead were used. The results of this test were still not con- sidered satisfactory. A letter was sent to a number of manufacturers of analytical reagents asking for assistance in getting completely lead-free reagents. Meanwhile collective work on an entrainment method for use with acid extracts from a sample of cocoa was under- taken, but with no better results. A full and detailed report, together with the modified dithizone method finally used, was submitted to the Analytical Methods Committee.The difficulties (and “snags”) that had arisen during the course of this work were as follows- BLANKS- The relatively large blanks that were found in the earlier tests presented a major difficulty. It was only as the result of a very considerable amount of work that the possibility of minute traces of lead being universally present was fully appreciated. In a determination in which 5 g of the foodstuff are used, a blank of 10 pg means 2 parts per million, a substantial and significant amount when a prescribed maximum limit of, say, 5 p.p.m. is in force, but one quite unacceptable when the limit is 1 part per million or less. This contamination was found to be due to several causes- (a) Reagents as supplied for ordinary analytical work may contain up to 1 or 2 parts per million of lead.This difficulty has been largely overcome by the enterprise of one or two manufacturers, who can now supply acids and alkalies containing not more than 0.01 part per million of lead, and by careful attention in the method of analysis to the preparation of reagents free from lead by washing them with dithizone solutions. In addition all reagents are stored in Pyrex-glass bottles. (b) The glass of apparatus and bottles may yield lead in the course of usage. The remedy is to use Pyrex glass wherever possible. (c) Atmospheric dust normally contains substantial amounts of lead, often to the extent of several thousand parts per million. Particular attention must therefore be paid to any possibility of dust contamination during any stage of a determination.The ideal laboratory is a separate room, air-conditioned, devoted entirely to this determination. DESTRUCTION OF ORGANIC MATTER- Two methods for the destruction of organic matter were considered: oxidation with acids (wet combustion) and incineration to a carbon-free ash (dry combustion). No evidence has been produced that there is any loss of lead by volatilisation if the temperature of incineration does not exceed about 500” C. On the other hand, a loss of lead occurs occasionally with foodstuffs such as fruit juices, which fuse on ignition, as the result of the formation of acid-insoluble lead silicate produced by attack on the silica of the basin. (Dry combustion of such foodstuffs in platinum capsules or basins is not advisable, as traces of platinum may be dissolved on solution of the ash in acid.) The use of magnesium nitrate as an “ashing-aid” was found to overcome this difficulty.The panel was divided in its opinion as to which method for destruction of organic matter was the better, once the initial difficulty of high blanks in the larger amounts of reagents used in the wet combustion method had been overcome. THE OPTIMUM pH OF THE ACID SOLUTION PREPARED FOR TREATMENT WITH DIETHYL- AMMONIUM DJETHYLDITHIOCARBAMATE SOLUTION- The optimum conditions for the extraction of the whole of the lead, while leaving inter- fering elements, such as bismuth, unextracted, necessitated a considerable amount of experi- mental work.July, 19541 THE DETERMINATION OF LEAD I N FOODSTUFFS 399 ASSESSMENT OF LEAD DITHIZONATE COLOUR- The fact that the work for this method was based on the mono-colour technique does not mean that it was considered better than the mixed-colour technique with absorption at defined wavelengths, or the Irving reversion technique. The choice was due to an attempt to make the method available in laboratories where such special apparatus as the Spekker absorptiometer is not available. By the mono-colour technique, an assessment can additionally be made with a tintometer or even by nesslerising.The difficulty that had to be overcome in the mono-colour method was due to the fact that, in the process of washing the excess of dithizone out of the chloroform solution of the lead dithizonate, the use of a solution containing more than a certain percentage of potassium cyanide might result in removal of part of the red solution of lead dithizonate. The procedure laid down in the method does, we believe, avoid this difficulty. Member A B C D E F G Weight of sample, f3 7.0 5.9 7.5 9-0 10.0 10.0 10.0 10.0 9.2 5.0 5.0 5.0 5-00 5-06 10.02 10.19 10.09 10.00 5-0 5.0 5.0 5.0 5.0 5.01 1 5-176 5.180 10.0 10.0 10.0 10.0 TABLE I RESULTS ON COCOA Method of Lead in Lead in breakdown Lead found, blank, sample, wet ash 9.0 3.5 5.6 Y Y 9.5 3-6 6.0 dry ash 9.7 3.0 6-7 3) 10.3 1.0 9.3 Pg Pg CLg wet ash 12.7 $9 14.2 dry ash, no 9.6 Y Y 11.4 M g w " Mg(NO3h dry ash with 7.9 wet ash dry ash 3) wet ash dry ash >Y Y$ Y Y 9Y 7.7 8-3 11.6 9.0 13.8 9-4 wet ash IY ¶> Y> $Y wet ash Y9- >Y wet ash 13.3 Y$ 11.9 dry ash 15.3 Y Y 12.3 5.8 8.8 2.2 2.2 2.5 2.6 2.5 0.9 0.9 1.3 1.3 6.9 8.4 7.4 9.2 5.4 7.2 3.75 4.00 5.2 5.8 10.6 8-1 12-5 8-1 5-1 2-8 5.4 4.6 2-3 4-8 5.3 5-7 4.1 9.2 2.7 9-2 1.8 13-5 1.8 10.5 Result, p.p.m.0.79 1-00 0.90 1.03 0.69 0.84 0.74 0.92 0-59 1-44 0.75 0.80 1.04 1.16 1.06 0.80 1-24 0-81 1.02 0.66 1-08 0.92 0.46 0.96 1.02 1.10 0.92 0.92 1.35 1.05 At the suggestion of the Analytical Methods Committee, one final test on cocoa and syrup was undertaken by the Panel. The results of this, in detail, are shown in Tables I and 11. The full details of the method are appended. It is hoped that this publication will allow the method to be tried out extensively, and that in a short time the results of outside work will confirm it as a standard or referee method.One final observation is that the full method is designed to be workable with all kinds of foodstuffs. When the foodstuff does not contain any significant amounts of phosphates or iron (e.g., cereals, sugars or wines), then Section 2A of the method can be omitted.[Vol. 79 400 THE DETERMINATION OF LEAD I N FOODSTUFFS METHOD FOR LEAD NOTE- In the absence of significant amounts of phosphates, bismuth or iron, Section 2A of the method mav be omitted. When the full method is used, however, the exact details given must be lollowed with care. Weight of g A 6.3 5.0 8.2 8.6 16-1 Member sample, B C D E G 8.52 5.42 8.69 5.0 5-0 5.0 5.0 5-73 5.22 10.06 10.29 6-09 7.12 15.5 13.1 5.2 5.2 8.1 9-8 10.0 TABLE I1 RESULTS ON SYRUP Method of Lead in breakdown Lead found, blank, wet ash 7.5 3.5 n 6.0 3.5 dry ash 10.5 3.0 99 4.0 1.0 Y9 10.0 1.0 wet ash 3.6 2.6 ?9 3.2 2.6 dry ash with 4.7 2.5 rg Pg M g w " wet ash 0 99 Y9 wet ash 5.3 99 5.0 dry ash 5.2 Y9 5-6 wet ash 99 99 99 wet ash 8.7 4.1 9Y 8-3 4-1 97 9.6 3.1 n 11.3 3.1 dry ash 9.7 1-8 2.5 2.5 0.9 0-9 Lead in sample, Pg 4.0 2-5 7-6 3.0 9.0 1.0 0.6 2.2 2-76 2-30 2.00 1.20 2-8 2.5 4-3 4-7 1.6 1.3 5-6 4.0 4.6 4.2 6.5 7.9 8.2 Result, p.p.m.0.63 0.50 0-92 0.35 0-56 0.12 0.1 1 0-25 0.55 0-48 0.40 0.24 0.49 0.48 0.43 0.46 0.26 0.18 0.35 0.31 0.89 0.81 0.80 0-84 0.79 REAGENTS- (see Notes). All reagents must be substantially "lead-free," either by purchase or by special preparation Sulphuric acid, concentrated. Perchloric acid, 60 per cent. Nitric acid , concentrated. Hydrochloric acid, 5 N.Ammonium hydroxide, s$.gr. 0.880. Ammonium hydroxide, 5 N. Sodium iodide solution-A 20 per cent. solution in water. Sodium metabisulphite solution-A 1.25 per cent. solution in water, filtered. Diethylammonium diethyldithiocarbamate-A 0.5 per cent. solution in chloroform, prepared freshly each day. Ammonium citrate solution-A 25 per cent. solution in water. Potassium cyanide solution-A 10 per cent. solution in water. This solution shall be at least 2 days old. Potassium cyanide wash liquid, 0.5 per cent.-Add 25 ml of 10 per cent. potassium cyanide solution to 475 ml of water and shake this solution with about 10ml of chloroform and 1 or 2 drops of the dithizone solution. Reject the extract, wash the solution again with chloroform and reject the lower layer. This lower layer must be efficiently collected by swirling the separator, and all the emulsion must be drawn off.Chloroform-Shake 250 ml of chloroform with 1 ml of 10 per cent. potassium cyanide solution and 25 ml of water containing about 20 drops of 5 N ammonium hydroxide, separate and reject the aqueous layer, wash the chloroform with water and filter it.July, 19541 THE DETERMINATION OF LEAD IN FOODSTUFFS 401 Dithizone, stock solation-Dissolve dithizone (as supplied) in chloroform, B. P., to give a 0.1 per cent. w/v solution. Dithizone, working solution-Shake 10 ml of the stock dithizone solution with 9 ml of water and 1 ml of 5 N ammonium hydroxide. Draw off the remainder into a centrifuge tube, centrifuge and transfer the clear brown upper layer by teat pipette to a 10-ml burette.Magnesium nitrate solution-Dissolve 10 g of magnesium nitrate, Mg(N0,),.6H20, in water and dilute to 100ml. Standard lead sohtion, 1 ml = 0*00001 g (10 pg) of lead-(a) Dissolve 1.60 g of lead nitrate in water, add 10ml of concentrated nitric acid and dilute to 1 litre. (b) Dilute exactly 5 ml of solution (a) to 500 ml. Prepare dilution (b) freshly as required. 1. PRELIMINARY TREATMENT OF THE SAMPLE- A : "Wet-ash" digestion-Digest 5 to 10 g of sample, or 50 ml of a liquid, such as beer, in a Kjeldahl flask with 5ml of concentrated sulphuric acid, adding nitric acid in small portions in the usual way until the residue remains colourless or pale yellow. Finally add 1 ml of 60 per cent. perchloric acid and heat to fuming point..Cool, dilute with 5 ml of water, heat again to fuming point, add 5 ml of water, then add 10 ml of 5 N hydrochloric acid and boil for 5 minutes. If there is any insoluble matter, filter the solution through an acid- washed filter-paper into a conical beaker and wash the filter with hot water. B : "Dry-ash" digestion-Mix in a silica basin 5 to 10 g of sample, or 50 ml of a liquid, such as beer, with 10ml of magnesium nitrate solution, evaporate the mixture to dryness, char the organic matter over a burner and burn the residue to a white ash in a muffle furnace or over an Argand or similar burner at a temperature not exceeding 500" C. Moisten the residue with 5 ml of water, add 10 ml of hydrochloric acid and boil the mixture for a few minutes. Filter through an acid-washed filter-paper into a conical beaker and wash the filter with hot water.2. EXTRACTION AND DETERMINATION OF THE LEAD- A T o the solution resulting from the preliminary treatment described above add 2 drops of methyl red indicator and then make it just alkaline with ammonium hydroxide, sp.gr. 0.880. Then make the solution just acid with 5 N hydrochloric acid and add 10 ml in excess (not more). Warm the solution to 40" to 90" C, add 2 ml of 20 per cent. sodium iodide solution and reduce the liberated iodine with 2 ml of freshly prepared sodium metabisulphite solution. Cool the solution, transfer it to a separating funnel and adjust the volume to 50 to 75 ml. Add 10 ml of diethylammonium diethyldithiocarbamate solution by pipette and shake the funnel vigorously for 30 seconds.Allow the layers to separate and transfer the chloroform layer to a 100-ml flask; then give the aqueous layer two small washes with chloroform without mixing and add these washings to the flask. Repeat the extraction with 10 ml of diethyl- ammonium diethyldithiocarbamate solution and add the second extract to the main extract. Reject the aqueous layer. To the combined extracts add 2.0 ml of diluted sulphuric acid (1 + 1) and evaporate the chloroform. Add 0.5 ml of 60 per cent. perchloric acid to the residual solution and heat it until fuming occurs and the fuming solution is clear and colourless. Cool the solution, add 10 ml of water and 5 ml of 5 N hydrochloric acid, boil it for 1 minute and then cool it again. B-Add 2ml of sodium metabisulphite solution, 2ml of ammonium citrate solution and 2 drops of bromothymol blue indicator.Add 5 N ammonium hydroxide to the full blue colour of the indicator. Cool the solution, add 1 ml of 10 per cent. potassium cyanide, transfer it to a 50-ml or 100-ml separating funnel and dilute it to 40ml. Add by pipette exactly 10 ml of chloroform. To the solution in the separating funnel add 10 ml of 5 N ammonium hydroxide and then dithizone solution, a few drops a t a time, with vigorous shaking, until there is a distinct excess. This will be indicated by a change from bright red to purple in the bottom layer and a yellowish shade in the upper layer. Shake the funnel for 30 to 60 seconds and allow the layers to separate. Run the lower chloroform layer into 50 ml of 0.5 per cent.potassium cyanide wash liquid contained in a second separating funnel, shake this for 30 seconds and allow the layers to separate. The chloroform extract should now be pure red and free from any trace Filter and store this stock solution in a refrigerator. Separate and reject the lower layer.402 THE DETERMINATION OF LEAD IN FOODSTUFFS [Vol. 79 of blue or green. Dry the outlet of the separator with a small roll of filter-paper, reject a few drops of the chloroform solution to clear the tap of the separator, and filter the rest of the chloroform solution through a dry Whatman No. 41 filter-paper into a dry test tube. Measure the optical density, using a 1-cm cell and Ilford green filter No. 604, or using a spectrophotometer at a wavelength of 520 mp.If the amount of lead contained in the sample is such that the optical density of the lead dithizonate is beyond the scale of the instrument used, it is advisable to extract the lead from the chloroform solution by shaking it with 1 per cent. nitric acid solution, diluting this aqueous extract to a known volume and taking an aliquot part for the determination in the way described above. BLANK OR CONTROL Prepare a reagent blank solution under the same conditions as the test, omitting only the sample, and determine the optical density. STANDARD GRAPH- Measure 0, 1.0, 2-0, 3.0 and 4.0 ml of standard lead solution into flasks containing 2 ml of diluted sulphuric acid (1 + 1) and dilute each solution to about 15 ml. Add 2 ml of sodium metabisulphite solution, 2 ml of ammonium citrate and 2 drops of bromothymol blue indicator.Add enough 5 N ammonium hydroxide to give the full blue colour of the indicator and, after cooling, 1 ml of potassium cyanide solution. Transfer each solution to a 50-ml or 100-ml separating funnel, dilute it to 40 ml and add exactly 10 ml of chloroform. Add 10 ml of 5 N ammonium hydroxide and then dithizone solution, a few drops at a time, with vigorous shaking, until the presence of an excess is indicated. Complete the extraction, with washing, exactly as described for the test solution. NOTES ‘ ‘LEAD-FREE ” REAGENTS- Hydrochloric , sulphuric , perchloric and citric acids , ammonium hydroxide and potassium cyanide are now obtainable “lead-free” to an extent that traces of lead contamination do not exceed 0.01 part per million.In the absence of “lead-free” reagents it is desirable that the reagents be rendered lead-free as follows- Acids-Sulphuric, nitric and hydrochloric acids can be rendered substantially lead-free by distillation , Pyrex all-glass apparatus being used. Ammonium hydroxide-Distil 400 ml of ammonium hydroxide, sp.gr. 04380, from a Pyrex-glass flask fitted with a safety trap into 250 ml of redistilled water kept cold by means of a bath of ice, controlling the pressure of the liberated ammonia gas by adjusting the rate of heating. Subsequently determine the strength of the distillate by titration with N hydrochloric acid. Ammonium citrate solution-Dissolve 125 g of AnalaR ammonium citrate in distilled water up to a volume of 400 to 450 ml, make it faintly alkaline to litmus paper with 5 N ammonium hydroxide , and extract it with chloroform and appropriate additions of the stock dithizone solution. Continue extraction until all metals have been removed and the extract is faintly green. Then make the solution acid by adding lead-free 5 N hydrochloric acid and extract with further portions of chloroform until the final extract is colourless. Potassium cyanide, 10 per cent. sahtion-Dissolve 50 g of AnalaR potassium cyanide in water and dilute to 100 ml. Extract this strong solution with chloroform and 1 or 2 drops. of dithizone solution until the extract is no longer coloured red, but has a greenish shade, Use as small an excess of dithizone as possible, because the excess is not readily removed. The excess of dithizone is extracted by shaking the solution with successive portions of chloroform. Dilute the extracted cyanide to 500 ml with water, warm it to remove chloroform and then cool it. STORAGE OF REAGENTS- may be suitable for some reagents, e.g., sodium citrate and potassium cyanide. MATERIAL FOR APPARATUS- It is advisable to store all “lead-free” reagents in Pyrex-glass bottles. Polythene bottles The use of Pyrex-glass apparatus throughout is advisable.

ISSN:0003-2654    

DOI:10.1039/AN9547900397

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

July, 19541 NEAL 403 The Determination of Titanium High-precision Absorptiometry BY W. T. L. NEAL The titanium content of titanium-base alloys and pure titanium metal can be determined absorptiometrically with a precision (coefficient of variation) of 0.03 per cent., with a Unicam SP600 spectrophotometer at a wavelength of 4100 A, use being made of the colour of the titanium - hydrogen peroxide compound in solutions with an optical density of 2-6 to 3.0 in 1-cm cells. In this paper an analysis is made of the effect of factors liable to influence the precision and accuracy of the determination, and the techniques required to secure high precision are described in detail. NORMALLY, absorptiometry is used for the determination of minor or trace elements in a sample, and the accuracy with which this can be achieved is usually better than 1.0 per cent. and can be as good as 0.5 per cent.The procedure is to make a solution of the sample in a suitable solvent, add reagents to produce a colour characteristic for the element to be determined, dilute to a known volume and measure the optical density of the solution at a suitable wavelength relative to a reagent blank or distilled water. This value for the optical density of the solution is compared with a calibration graph produced by measuring the optical density of similarly prepared solutions containing known amounts of the element. The only essential features making the technique of high-precision absorptiometry more precise than that of conventional absorptiometry are that the optical density of the sample solution is measured, not against a blank solution of zero, or nearly zero, absorption, but against a reference solution containing a known concentration of the element to be determined and coloured in the same way as the sample or test solution; and that the optical densities of these solutions are much higher (in the range 2 to 5) than those used ineminor element or trace analysis, which are usually less than 1.l It can be shown theoretically (see below) for solutions that obey the Beer - Lambert law, that the greater the concentration .of the solutions, the higher will be the precision.How far this statement holds for solutions that do not obey the Beer - Lambert law can only be discovered by experiment. It will be shown that the titanium - hydrogen peroxide solutions do not behave in the way predicted by the theory when their optical density rises above about 1.2 and that there is a certain concentra- tion above which the precision decreases.The precision of the method is also a function of the spectrophotometer used and of characteristics in the solutions, such as acid concentration and temperature. Extensive preliminary experiments were carried out to assess how closely these variables must be con- trolled in order to attain the greatest accuracy. As a result of this work, a method has been developed in which the titanium content of a sample can be determined with a precision (coefficient of variation) of 0.03 per cent. The results can be found for a single sample in 14 to 2 hours after receipt together with the time required for its solution, which may add + to 3 or more hours.Six to eight individual determinations can be carried out per day. OUTLINE OF METHOD- The method is based upon the formation of the familiar colour of the titanium - hydrogen peroxide compound in 20 per cent. w/v sulphuric acid and measurement of the relative transmission of the test solution (that is, a solution from the sample being examined) , against a reference solution of accurately known concentration by means of a Unicam SP500 spectro- photometer at 4100~. The concentration of the test solution can then be read from a calibration graph, or by a procedure to be described that avoids the use of a calibration graph.404 NEAL: THE DETERMINATIOK OF TITANIUM [Vol.79 PRINCIPLE OF METHOD- For solutions obeying the Beer - Lambert law, the following relationships can be deduced. Let D,, D,, etc. = optical density of solutions 1, 2, etc. C,, C,, etc. = concentration of solutions 1, 2, etc. I0 = intensity of incident light. I,, I,, etc. = intensity of light transmitted through solutions 1, 2, etc. K = a constant that is a function of the solution and of the incident I = path length of light through cell (Le., cell length). By definition- Dl = 10~1OIO/Il .. .. .. .. .. . . (la) D2 = 1~~1oIo/I2 .. .. .. .. .. . . (lb) and, according to the Beer - Lambert law- D, = KIC, .. .. .. .. .. .. . . (2a) D, = KEC, .. .. .. .. .. .. . . (2b) whence Io/Ii = 10KICt . .. .. .. .. .. . . (3a) Dividing (3a) by (3b) we have- .. .. .. .. .. . .(4a) .. .. .. .. . . (4b) light. and Io/Iz = 10KzCa . .. .. .. .. .. . . (34 .. * - (4.4 I,/& = 1OK'(C1-G) - - 10KG(1-Ca/C1) . - - 1O-Dlca-1, .. .. .. .. (where a = C,/C,). AI Theref ore 1 - 1,/11 =I, = 1 - 10-Dx(a-1) . . .. .. . . (5a) Now, if C , is greater than C,, solution 1 will be used as the reference solution, and the instrument will be adjusted to read 100 per cent. transmission with solution 1 in the light beam; that is I , = 100. (where AI = I , - I,). Thus we may rewrite (5a) as- = - 10-D1(a-l)] .. .. .. .. . . (5b) where AI is expressed as percent age transmission. In Fig. 1, curve A is the plot of AI against corresponding values of D, for a = 1-25. It will be seen that, as D, increases, AI increases, becoming a maximum at D, = infinity. Fig. 1.Theoretical (A) and experimental (B) variation of change in transmission ratio with optical densityJuly, 19541 BY HIGH-PRECISION ABSORPTIOMETRY 405 Experiments were carried out to compare the behaviour of titanium - hydrogen peroxide solutions with that predicted theoretically. Pairs of solutions were prepared such that while their optical densities varied over the range 0 to 5 (in a 1-cm cell) the ratio of the concentra- tions of the individual solutions of each pair, a, was constant at 1.25. The transmission ratios of these pairs of solutions were measured, and the results are plotted as curve B, Fig. 1. It will be seen that there is a marked deviation from ideality above an optical density of about 1.2. Theoretical and experimental curves have been plotted for values of a ranging from 1.1 to 1.5, the general form of the curves being the same as that shown in Fig.1, and in particular the position of the peak of the experimental curve is independent of a. The theoretical relationships can be analysed mathematically and the position of minimum error can be determined, as has been shown, for example, by Hiskey., Such an approach cannot be made when the solutions show the marked deviation from the Beer - Lambert law indicated in Fig. 1. However, the following semi-empirical approach can be made if it is remembered that the primary requirement is the best possible calibration curve. The accuracy of the determination of a concentration at any point on a calibration curve is defined exactly as the product of the concentration corresponding to that point and the slope of the curve at that point, where the slope is d(AI)/dC, and the error in deter- mining AI is constant for all values of AI.Now calibration curves are nearly linear over the range a = 1 to a = 1.1 or even 1-2 in the practicable optical density range 0 to 5 or 6, so that the slope, S, of the calibration curve may be considered as equal to AI/(C, - CJ, where A I is the transmission ratio of the pair of solutions, C, and C,. The over-all accuracy, A , attainable can then be expressed as the product of this slope and the mid-point of the concentration range C, to C,, that is- = AI(1 + a)/2(1 - a). Hence the accuracy is proportional to AI at constant a, and the criterion for the best calibra- tion curve is that it has the greatest A1 for a given value of a within the range 1.0 to 1.2.The following limitation applies to this semi-empirical method of assessing the accuracy of a calibration curve. It has been assumed that the most accurate part of a calibration curve on which to work is the steepest part, where a is equal to or not much greater than unity. This is not true for solutions obeying the Beer - Lambert law when D, is less than 0.4343 (see Hiskey,, Fig. 6), but applies for all higher values of D,. Hiskey also showed that the minimum error decreases as D, increases. When these conclusions are applied to the experimental curve (Fig. l), it can be seen that the Beer - Lambert law is obeyed for all optical densities up to about 1-2, so that the minimum possible error on the calibration curves below D, = 0-4343 will still be greater than that for calibration curves at values of D, greater than 0.4343.This will in general hold for all systems to which the high-precision absorptiometric method is applied, so that the criterion for maximum accuracy stated above can be safely applied. It will be noted that, as long as solutions obey the Beer - Lambert law, the best calibration curve results from the use of solutions of as high an optical density as possible, with values of a in the range a = 1.0 to a = 1.2 (where the slope of the curve is greatest). However, for the titanium - hydrogen peroxide solutions used in the present method, curve B in Fig. 1 shows that the best calibration curves are to be found in the optical density range 2.1 to 2.5, i.e., a concentration range of 15 to 18mg of titanium per 100ml of solution.OTHER PRELIMINARY EXPERIMENTS In order to assess what influence changes in the many possible variables had on the value determined for the transmission ratio between a pair of solutions, extensive experiments were performed and the results were assessed by normal statistical analysis. These experi- ments will not be described in detail but the conclusions will be stated and, where it is thought necessary, these will be supported by representative results. PREPARATION OF CELLS- the glass absorption cells. It must be emphasised at the outset that great care is necessary in the preparation of In order to attain reproducible values for the transmission ratio406 NEAL THE DETERMINATION OF TITANIUM [Vol.79 between a given pair of solutions, it has been found necessary to rinse the emptied cells first with dilute sulphuric acid, twice with either industrial spirit or acetone and then twice with ether. The cells are then inverted on a filter-paper pad to drain and dry, after which the outer optical surfaces should be polished with a clean dry chamois leather. The cells should be filled in such a way that the outer surfaces do not become contaminated with the solution. If such a procedure is adopted, the transmission ratio of a given pair of solutions in the same pair of cells should lie within a range of k0.l per cent. transmission after emptying, cleaning and refilling. VARIABLES THAT ARE A FUNCTION OF THE SOLUTION- These variables are the acid concentration and temperature of the solution, the amount of reagent added, the stability of the colour of the titanium - hydrogen peroxide compound and other elements present.All but the last of these are inherent in the method, whereas other elements are incidental to the particular specimen being examined. The effect of some elements that are common contaminants of titanium will be described later; here attention will be confined to the inherent variables listed above. - - Slit width, mm Fig. 2. Variation of instrumental response with slit width of Unicam SP500 spectrophotometer for solutions of various concentrations. The solutions contained the following amounts of titanium per 100 ml: curve A, 12 mg; curve B, 16 mg; curve C, 20 mg; curve D, 24mg; and curve E, 32mg Amount of reagent-About 3 ml of 20-volume hydrogen peroxide are required to develop the full colour of the titanium - hydrogen peroxide compound in 100 ml of solution containing 15 mg of titanium and 20 g of sulphuric acid, and any excess of hydrogen peroxide does not cause a change in the intensity of the colour. Stability of colour-The transmission ratio between several pairs of solutions was found to remain constant for at least 14 days.Variation in acid concentration-This can occur in two different ways. Either the con- centration of acid may differ in the individual members of a pair of solutions between which the transmission ratio is to be measured, or it may be the same for the members of a pair while differing from one pair to another.Variations of the first type alter the observed transmission ratio between a pair of solutions. A difference of 2 4 g of sulphuric acid perJuly, 19541 BY HIGH-PRECISION ABSORPTIOMETRY 407 litre at a total acid concentration of 200g of sulphuric acid per litre causes a change in the transmission ratio of 0.1 per cent. (equivalent to 0.02 per cent. of titanium, see below). Variations of the second type do not cause any change in the transmission ratio, even over the range 100 to 300g of sulphuric acid per litre. Variations of temperature-These can also occur in two ways. Either the temperature may differ between the individual solutions of a pair, or, while being constant for the members TABLE I INFLUENCE OF ACID CONCENTRATION AND TEMPERATURE ON THE TRANSMISSION RATIO BETWEEN A PAIR OF SOLUTIONS WHEN THE ACID CONCENTRATION IS SIMILAR IN BOTH Reference solution contains 150-00 mg of titanium per litre Titanium concentration Temperature of reference of reference and solution, test solutions, mg per litre " C 150.04 {BE 24.2 180.05 Acid concentration of reference and test solutions, g per litre 95 190 290 f A \ 99.2 99.3 99.2 99.1 99.3 99.3 99.1 99.3 99.3 41.6 41.7 41.6 41.6 42.0 41.7 41-7 41.5 41.8 of any pair, may differ from pair to pair.Variations of the first kind are unimportant, as the solutions attain the same temperature within 2 to 3 minutes of the absorption cell being put into the cell holder, i.e., after 2 to 3 minutes a constant value for the transmission ratio is attained. Variations of the second kind correspond to the day-to-day variations of the laboratory temperature and are therefore important.It has been found that temperature variations of this kind do not cause any variation in the transmission ratio of a pair of solutions Slit width 20 rnn I 3.000 4.000 5.000 Wavelength. A Fig. 3. Absorption curve for titanium - hydrogen peroxide colour, and bandwidths of light corresponding to slit widths of 2.0, 1.0 and 0.5 mm, with a Unicam SP500 spectrophotometer when the acid concentration of the individual solutions is the same. Table I contains some representative results showing the constancy of the transmission ratios determined under these conditions. VARIABLES THAT ARE A FUNCTION OF THE SPECTROPHOTOMETER- When balancing the Unicam SP500 spectrophotometer with a given reference solution in the light beam and the transmission control set to 100 per cent., it is possible to achieve balance at different settings of the slit width and sensitivity control.Only one of these settings may be varied at will, however, the other being a dependent variable. Of these4-08 NEAL: THE DETERMISATIOK OF TITAXIGM [Vol. 79 two variables, only the slit width is calibrated, so that it is necessary to use the slit width as the independent variable. The response of the instrument, which is the deflection of the meter needle for a given change in the setting of the transmission scale, alters as the slit width is changed. The relation between slit width and instrumental response is shown in Fig. 2. (The units on the instrumental response axis are equal to the reciprocal of the change in percentage transmission required to deflect the meter needle one division.) The different curves are produced by placing solutions of different concentration in the light beam, and the horizontal broken line represents the minimum response considered permissible, if the instrument is to be read with certainty to 0.1 per cent.transmission. It corresponds to a deflection of the meter needle of one-fifth of a division for a change of 0.1 per cent. trans- mission. It has already been shown that the optimum strength of the reference solution is 16 mg of titanium per 100 ml, so that a slit width greater than 0-7 mm should be used (see Fig. 2). Variation in the slit width also alters the bandwidth of the light passing through the solution and so can alter the value of the transmission ratio between a pair of solutions, by virtue of the fact that the bandwidth extends for different distances on either side of the peak of the absorption curve, as shown in Fig.3. Such a variation in bandwidth would certainly alter the absolute value of the optical density measured against, say, distilled water, although when, as in this method, two similar solutions are to be compared, the effect may be expected to cancel out, at least in parf. Experiments carried out to test this possibility suggest that the effect is not significant. Nevertheless, when dealing with sample solutions that contain other absorbing species, it may be possible to avoid interference by having a small bandwidth for the light from the monochrometer.If this is impossible, then it becomes necessary to apply a correction, as for iron, to be described below. This can only be done if the slit width is kept constant, so that the overlap of the bandwidth on the absorption band of the interfering ions is constant. Hence, it must be arranged that a constant slit width is used. This width should be such that an adequate instrumental response is secured, but no wider. It must be noted that variations in the intensity of the light source, response of the photo-tubes, and so on, such as can occur from instrument to instrument, can alter considerably the positions of the curves shown in Fig. 2, and hence require the use of a different slit width. In this work a slit width of 0.8 mm was used.CONCLUSIONS- The necessary conditions for the highest accuracy in the determination of titanium by this method are that (a) the concentration of the reference solution should be 150mg of titanium per litre in 20 per cent. w/v sulphuric acid solution; (b) 5 ml of 20-volume hydrogen peroxide solution should be used to colour 100 ml of solution; (c) the acid concentrations of the reference and test solutions should not differ by, more than 2.5 g of sulphuric acid per litre; (d) the slit width of the spectrophotometer should be kept constant and should be as narrow as possible consistent with attainbg an adequate response from the instrument ; (e) the absorption cells should be scrupulously cleaned as described above; (f) the transmission ratio of the solutions can be measured at any convenient time after colouring; and (g) the transmission ratios of the solutions can be measured at any prevailing laboratory temperature, METHOD REAGENTS- Standard titanium solzction-Prepared and st andardised as described below.Sul$hwic acid, concentrated-Anal ytical-reagent grade. Sztlphwic acid, 20 per cent. w/v-Dilute the concentrated acid to a sp.gr. of 1-123 at Hydrogen $eroxide, 20-volume. Ammonium sulphate-Analytical-reagent grade. Balance-A semi-micro balance, weighing to 0-00001 g. Flasks-Several 100-ml lipped conical flasks, with suitable covers, and 100-ml calibrated Spectrophotometer-Unicam SP500 or Beckmann DU spectrophotometer. 20" c. APPARATUS- flasks, calibrated to 0.01 ml.July, 19541 BY HIGH-PRECISION ABSORPTIOMETRY 409 DEFINITIONS- Standard titanium solution-A stock solution of 150 mg of titanium contained in a known weight of solution the volume of which is 100 ml.Sample solution-A similar solution prepared from a known weight of sample. Reference solution-A solution , coloured with hydrogen peroxide, prepared from a 10-ml aliquot of the standard titanium solution. Test solution-A similar solution prepared from a suitable aliquot of the sample solution. A set of solutions-This comprises the reference and test solutions, which are coloured and prepared at the same time. PRELIMINARY REMARKS- The method is based upon the comparison of a solution of unknown titanium concentrk- tion (coloured by the complex formed with hydrogen peroxide) with a similar solution of known titanium concentration, when the transmission ratio of the two solutions is a measure of the difference in concentration.The conditions necessary for an accurate determination of the transmission ratio between these two solutions have already been described. In addition, it is necessary that all aliquots should be taken by weight, as it has been found that aliquots taken volumetrically are not sufficiently reproducible. With volumetric techniques only, the analytical results have a coefficient of variation of 0.07 per cent., but with a weight method for taking the aliquots, the coefficient of variation is reduced to 0.03 per cent. Even if a volumetric method is to be used, the pipette must be recalibrated with solutions containing 5 g of ammonium sulphate and 20 g of sulphuric acid per 100 ml of solution, otherwise the results will be 0.1 to 0.2 per cent.low. PROCEDURES FOR COMPARING CONCENTRATION- Two methods are available for comparing the concentrations of the reference and test solutions. One method is to prepare a reference solution of an accurately known concentra- tion, say 15.000 mg of titanium per 100 ml, and to prepare a calibration curve against such a solution. Aliquots of the sample solution are taken such that the concentration of the test solution will lie in the range covered by the calibration curve. The other method is to take aliquots of the standard and test solutions such that both are of approximately the same strength, say to within A1.0 per cent., and so that the concentration of the reference solution is 15.000 & 0.05 mg of titanium per 100 ml.Then the difference in the concentration of the reference and test solutions is proportional to the difference in transmission ratio, and the concentration of the test solution can be determined by applying a small correction to the known concentration of the reference solution. In what follows, the former method will be referred to as the calibration-curve method, the latter as the correction method. Both methods have their sphere of application, the calibration-curve method when a large number of routine determinations have to be undertaken of samples of unknown titanium content, the correction method when determinations of a few titaniferous alloys of approxi- mately known composition have to be made. In the calibration-curve method, it is necessary to check the calibration curve frequently, as any slight change in the transmission control potentiometer will alter the curve to some extent.Again, the points on the calibration curve must be determined in triplicate at least and at small intervals of titanium concentra- tion, if the curve is to be drawn with the required accuracy. Alternatively, fewer points can be determined in triplicate, and an empirical equation to the curve can be deduced, from which other points can be interpolated or a table of transmission ratio against concentration can be made. Such a procedure is lengthy and tedious if only a few determinations are required, the correction method then being preferable, as the slope of the linear correction graph can be determined rapidly; single determinations of a few points and the drawing of the best straight line give an adequate accuracy.NOTE-If aliquots are to be taken volumetrically, the calibration-curve method must be used, as the aliquots in the correction method must frequently be taken with a burette, and this cannot be done with sufficient accuracy volumetrically. MEASUREMENT OF TRANSMISSION RATIO- Two cells are selected and cleaned as already described; one is marked reference and the other is marked test. These cells should be filled with the reference and test solutions, respectively, and the transmission ratio is measured as follows.410 NEAL THE DETERMINATIOK OF TITANIUM [Vol. 79 The instrument and light source should be turned on 15 minutes beforehand.The wavelength is set to 4100 A and the slit width to the predetermined value, the switch is set to check and the instrument is balanced with the dark-current control. The reference solution is then placed in the light beam, the shutter is opened and the instrument is re- balanced by means of the sensitivity control. The test solution is then placed in the light beam, the switch is turned to “1” and the instrument is balanced by means of the transmission control. The reference solution is then replaced in the light beam and the switch is turned to check. If this is not so, the measurement should be repeated. The measurements are repeated after 1-minute intervals until a constant value is attained. This occurs when the temperature of the cells and solutions is the same (the values are constant after 1 to 2 minutes). Except during measurements of the trans- mission ratio, neither solution should be left in the light beam, or the results may be erratic.The procedure to be followed when measuring the transmission ratios of a set of solutions is as follows. Both cells are filled with the same reference solution, and the transmission ratio (cell check value) is measured. Initially this must be repeated several times, emptying, cleaning and refilling the cells between each determination, so that a reliable estimate of the value can be made. Subsequently one determination will be sufficient provided that the value agrees with the cell check value found previously to &O-1 per cent. This cell check should be made before the measurement of the transmission ratios of test solutions is attempted, so as to ensure that the cells are in good order.The reference solution used for the cell check should not be emptied from the reference cell, but should be used as the reference for the determination of the transmission ratios of all the test solutions of the set. Finally, the cell check should again be made, to ensure that the reference solution - reference cell combina- tion has undergone no change. It is advisable to keep the reference cell covered to avoid possible loss by evaporation. PREPARATION OF THE STANDARD AND SAMPLE SOLUTIONS- Standard soldion-The titanium selected for the preparation of the standard solution should be completely analysed so that the titanium content is accurately known.The accuracy of the method is limited by the accuracy of this value. The standard titanium used in this work was a Van Arkel titanium with the following composition: oxygen, 0-019; hydrogen, 0.022 ; nitrogen, 0.003 ; iron, 0.016 ; manganese, 0-012 ; aluminium, 0.005 ; silicon, 0405; vanadium, 0-002 to 0.003; and lead, 0.002 per cent. Total impurities, 0-087 per cent. ; titanium, by difference, 99.91 per cent. Suitable materials for use as standard samples are Van Arkel titanium or arc-melted pure titanium sponge. The sponge itself is not suitable for a standard material, as its composition varies from lump to lump and even within each lump. The selected material should be milled as finely as possible, and an amount weighed out to contain 15040mg of titanium.This should be placed in a lipped 100-ml conical flask containing 5.0g of dry ammonium sulphate. The flask, contents and cover are then weighed to the nearest 0.1 g. About 6 rnl* 1 ml of concentrated sulphuric acid are added and the flask is heated on a hot-plate to just below the boiling point of the solvent mixture. If the reaction becomes sluggish, or if a white crystalline precipitate appears, more acid (1 ml at a time) should be added. When solution is complete, the flask and contents are cooled and weighed, the difference between the weight of the flask before the addition of the acid and after solution giving the weight of sulphuric acid present. The volume of 20 per cent. w/v sulphuric acid necessary to make the weight of sulphuric acid up to 20 g is calculated, and the contents of the conical flask are transferred to a tared 100-ml calibrated flask with this volume of acid.The transfer is completed with distilled water and the volume made up to the mark at 20” C . The calibrated flask and contents are then weighed and the weight of the solution is calculated ; three or four such solutions should be prepared simultaneously to enable the strengths of the individual solutions to be checked (see later). Sample solutions-These are prepared in the same way as the standard solutions, from 0.1 to 0.2 g of the finely divided sample. PREPARATION OF THE REFERENCE AND TEST SOLUTIONS- The required aliquot is taken from the standaxd or sample solution by means of a pipette or burette and run into a tared 100-mi calibrated flask, whose volume is known to 0.01 ml.This is followed by about 50 mi of 20 per cent. w/v sulphuric acid solution to wash the titanium solution out of the neck of the flask and then by 5.0ml of 20-volume hydrogen peroxide The instrument should still be in balance.July, 19541 BY HIGH-PRECISION ABSORPTIOMETRY 41 1 solution. The solutions are then made up to the mark with 20 per cent. w/v sulphuric acid solution. It is important when these solutions are made up that they should not differ in temperature by more than 0.2" C, which is equivalent to a change in volume of 0.01 per cent. This is conveniently effected by allowing them to stand in a water-bath or a thermostatically- controlled enclosure at 20" C for 10 minutes, Use of a thermostat is in general more con- venient, for, should it be necessary to repeat the preparation of any solution in a given set, this can be done without preparing a fresh reference solution, which would otherwise be necessary.CHECKING THE STANDARD TITANIUM SOLUTIONS- Before the standard solutions are used for the preparation of a calibration curve, or for the preparation of reference solutions for making titanium determinations, it is advisable to inter-check them. First a correction curve should be prepared by taking 9.9, 10.0, 10.1 and 10.2-ml aliquots from any one of the reference solutions, weighing them and preparing coloured solutions with hydrogen peroxide as already described. The transmission ratio of each of these solutions against the 10-ml aliquot is measured and the relationship between difference in transmission ratio and difference in concentration is determined by calculation or graphically, assuming a linear relationship.When calculating the concentration of the solutions, the volumes of the calibrated flasks must, of course, be taken into account. Ten- millilitre weighed aliquots are now taken from each of the standard solutions and a set of coloured solutions are prepared, the transmission ratios of which are determined with one of them as the reference solution. The amount of titanium in the standards should then be calculated and must lie within ,t0.03 per cent. of the nominal value to be accepted. PREPARATION OF CALIBRATION CURVE- Weighed aliquots of 10, 11 , 12,13 and 14 ml are taken from one of the standard solutions and diluted as described. The 10-ml aliquot must be such that its weight is exactly O*looOO -+ O.oooO1 of that of the standard solution from which it was taken (this can readily be done with a capillary pipette calibrated in 0.01-ml units and readable to 0.001 ml).With this solution as the reference solution, the transmission ratios of all the solutions, including the 10-ml aliquot, are measured. These measurements should be repeated twice more with a new set of solutions each time. Each individual point is plotted on a graph of adequate size. As it stands, this curve will be relatively inaccurate in the regions midway between the various sets of points. This could be remedied by determining more points, but a simpler way is to deduce from the determined results the values of the transmission ratio for solutions con- taining exactly 15.0, 16.5, 18.0, etc., mg of titanium per 100 ml and to calculate an equation to the curve.The following is an example of such a calculation- Original results r Concentration, C, Transmission A -l mg of Ti per 100 ml ratio, I , 15.0 100.03 % % 16.5 62.88 18.0 41.54 19.5 28.76 21.0 20.92 First Second log,,^ differences differences 2.0000 0.2014 1.7986 > 00212 0.0206 0.0213 1.6184 1-4588 1-3205 In column 4 the differences of the pairs of values of log,, I are given, while in column 6 are given the differences of the figures in column 4. The mean of the figures in column 6 is 0-0210. From the figures in columns 4 and 5, and deducing the appropriate general formulae for the series, it can be shown that the values of log,, I are given by the expression- where n is given by C = 150 + 14%~.412 NEAL: THE DETERMIXATION OF TITANIUM [Vol.79 From such an equation, intermediate points for the calibration curve can be computed, or a table of values of C calculated for corresponding values of I . The general form of the calibration curve is shown in Fig. 4. EFFECT OF OTHER IONS- The alloys examined to date in this laboratory have been virtually pure binary or tertiary alloys, the impurities, i.e., silicon, manganese, vanadium, aluminium, magnesium and calcium, all being present at levels lower than 0.1 per cent. The presence of these elements at this level did not have any effect on the transmission ratios, nor did potassium and ammonium sulphates, TABLE I1 EFFECT OF IRON ON THE TRANSMISSION’ RATIO OF TITANIUM - HYDROGEN PEROXIDE SOLUTIONS Transmission ratio of pure titanium solution minus determined transmission Concentration ratio in presence of am of titanium, mg per 100 ml % % 15.0 + 0.6 + 0.9 16.5 + 0.3 + 0.7 18.0 + 0.3 + 0.5 when present in amounts up to 1 g per 100 ml of test solution.The effect of iron up to 25 per cent. of the titanium content has been investigated and the results are shown in Table 11, in which the amount of iron present is expressed as mg of iron per 100 ml of test solution. It will be seen that the effect of iron varies with the titanium concentration of the test solution. For example, when the titanium concentrations of the test solution is 16; mg per 100 ml, 1 per cent.of iron in the sample is equivalent to 0-007 per cent. of titanium, but when the titanium concentration is 18mg per lOOml, 1 per cent. of iron is equivalent to 0.014 per cent. of titanium. Fig. 4. Calibration curve TYPICAL RESULTS OF DETERMINATIONS The following results are typical of many obtained during the analysis of titanium- oxygen and titanium - iron - oxygen alloys. In Table I11 are shown results determined by taking the aliquots volumetrically, making use of a calibration curve. The following ratios were determined gravimetrically by taking 10-ml diquots from nine different 100-ml sample solutions-July, 19541 BY HIGH-PRECISION ABSORPTIOMETRY 413 0*10018, 0-10009, 0-10022, 0-10025, 0.10019, 0.10022, 0*10021, 0.10017, OmO26 ; mean In Table IV are shown results found by the correction method with the weight method of taking aliquots.The following are the results of thirteen replicate determinations on a standard Van Arkel titanium- 0.10020: coefficient of variation = 0.05 per cent. 99-93, 99.95, 99.85, 99.91, 99.91, 99-91, 99.85, 99-91, 99-91, 99.95, 99-91, 99.86, 99.87; Each figure quoted above and in Tables I11 and IV is the result of a complete deter- mination, i.e., the titanium content is determined on different portions of the sample ; they are not merely replicates in a single sample solution. TABLE I11 TYPICAL RESULTS BY THE CALIBRATION-CURVE METHOD WITH mean 99.90: coefficient of variation = 0.03 per cent. VOLUMETRIC ALIQUOTS Sample 1 2 3 4 5 6 7 8 9 10 Amount of titanium found, per cent. r A \ 93.82 93.75 92.88 92.77 92.74 96-54 96.60 95.60 95.54 95-60 94.51 94.49 94.33 92-94 92-95 92.94 92.23 92.3 1 92.38 94.59 94-40 92.55 92-53 90.70 90.75 Over-all coefficient of variation: 0-07 per cent. Mean value for titanium, 93.79 92-80 96.57 95.58 94.44 92.94 92-31 94.50 92.54 90.73 % TABLE IV TYPICAL RESULTS BY THE CORRECTION METHOD WITH WEIGHT ALIQUOTS Sample 1 2 3 4 5 6 7 8 9 LO Amount of titanium found, per cent. r A % 98.1 1 95.86 99.45 98-30 99-70 95-71 97-41 98.41 94.66 96.43 98.14 95.82 99.45 98.31 99.64 95-68 97-45 98-46 94-68 96.43 Mean value of titanium, 98.13 95-84 99.45 98.31 99-67 95.70 97-43 98-44 94.67 96.43 % Over-all coefficient of variation: 0.03 per cent. DISCUSSION AND CONCLUSIONS The results by the procedure described have been obtained by the use of normal com- mercially available apparatus, which has not been modified in any way. The coefficient of variation of these values is 0-03 per cent.; that is, the range within which 95 per cent. of all the values will fall is &O-06 per cent. This is considerably better than can be achieved by the usual gravimetric procedures, even by the most careful workers. The time taken to complete a determination after the dissolution of the sample is about 90 minutes. This compares favourably with that taken with gravimetric procedures. The work described above has been carried out as part of the research programme of the National Physical Laboratory, and this paper is published by permission of the Director of the Laboratory. REFERENCE 1. Hiskey, C. F., Anal. Chew., 1949, 21, 1440. NATIONAL PHYSICAL LABOR~~TORY TEDDINGTON, MIDDLESEX January 8th, 1954

ISSN:0003-2654    

DOI:10.1039/AN9547900403

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

414 MILNER AND PHENNAH: THE DETERMINATION OF TITANIUM I N [Vol. 79 The Determination of Titanium in Uranium - Titanium Alloys by Differential Absorptiometry BY G. W. C. MILNER AND P. J. PHENNAH The differential absorptiometric technique has proved suitable for the determination of from 6 to 76 per cent. of titanium in titanhm -uranium binary alloys. The yellow colour produced by titanium with hydrogen peroxide in 2.5 N sulphuric acid solutions was used in this work and optical density measurements were made with both the Beckman DU spectrophoto- meter and the Spekker absorptiometer model H 760. Good agreement was found between the results determined with these instruments, showing that the Spekker absorptiometer has a potential use in the determination of major alloying constituents.Although macro amounts of uranium can be deter- mined differentially, this determination proved impossible in the presence of titanium. The theoretical principles of differential absorptiometry are summarised. HISKEY and co-workers1 ,293 have shown from theoretical considerations that certain absorptio- metric determinations can be made with greater precision if the usual water or blank solution is replaced by a standard solution of the same substance at approximately the same con- centration. This differential technique has been successfully applied in analysis by various workers using the Beckman DU spectrophotometea, or an equivalent instrument, for optical density measurements. For example, Bastian determined large concentrations of coppeI-4 and nickels by means of the natural colours of the ions produced by these elements in perchloric acid solutions.Crouthamel and Hubbarda determined uranium by a thiocyanate procedure in acetone solution, whilst Neal and Short' determined the titanium in pure titanium metal and titanium - titanium dioxide mixtures by means of the titanium - hydrogen peroxide colour. Many analytical laboratories in this country arcs equipped with Hilger Spekker absorptio- meters and not with the more expensive spectrophotometers. When the differential technique was used for the determination of titanium in titanium - uranium binary alloys, therefore, experiments were made to test the applicability of the new Spekker absorptiometer to differen- tial absorptiometry. Various attempts have been made in the past by different workers to use this type of instrument for the determination of the major constituents of alloy systems.StrossS used a system involving the use of neutral filters, these filters being inserted in front of the compensating photocell after the initial balancing of the instrument. Pollak and Nicholass pointed out, however, that this type of technique contravened one of the design principles of this instrument, namely that the light intensity on the indicating photocell should be the same when the instrument is set as when the readings are made. For accurate work, these workers suggested that the neutral filters should be used in front of the indicating photocell only. At a recent meeting of the Physical Methods Group of the Society,' however, some doubts were expressed as to the suitability of the Spekker absorptiometer for differential work.In testing this instrument, our method consisted in assaying the alloys for titanium by an established gravimetric methodlo and then by the differential technique, determining the titanium by means of the titanium - hydrogen peroxide colour, making use of both the Beckman spectrophotometer and the Spekker absorptiometer for optical density measure- ments. THEORETICAL PRINCIPLES OF DIFFERENTIAL ABSORPTIOMETRY Consider first the general case in which the absorptiometer or spectrophotometer is used in the usual way. A solvent blank is used when the instrument is balanced at zero absorption or 100 per cent. transmission. The optical density of the coloured solution is then measured with reference to this setting and the concentration of the solution is determined by reference to a calibration graph, which is prepared by an identical procedure.Suppose that the error in determining a concentration, cl, by this method is & x per cent. of the total concentration. This error has no theoretical interpretation, but is defined to include all errors made in theJuly, 1954 J 416 measurement of the optical density, whether instrumental or operational errors. Moreover, this error will be inherent in all measurements of this kind. The quoted result would be- URANIUM - TITANIUM ALLOYS BY DIFFERENTIAL ABSORPTIOMETRY c1 L- x per cent. or c, 5 a .. . . 100 Now consider the differential technique in which the only variation in the method is that a solution of accurately known concentration, cg, of the substance being determined is used in place of the solvent blank in the balancing of the instrument.Let the value of c2 be such that Occ,<cl. The concentration of the substance as measured is the difference, c1-c2, to which the same overall error x per cent. still applies. The total concentration determined is now the sum of the following two parts: (a) a concentration c2, which is accurately known and therefore has a negligible error, and (b) a concentration, cl&c,, which has an error of x per cent. The quoted result would therefore be- c, + (c,-c2) & x per cent. of (c,-c2) The error introduced by the differential technique in equation (2) is clearly smaller than that introduced by the usual absorptiometric technique in equation (1).As the value of c, is increased, the error is correspondingly reduced to a theoretical limiting value of zero when c2 equals c,. Suppose the value of c1 is so great that it is impossible to measure the optical density of this solution with reference to the reagent blank. For the Spekker absorptiometer this optical density will be greater than 1.3, as this type of instrument is only calibrated to measure optical density values up to a maximum of 1-3. In these circumstances the usual practice consists in diluting the solu- tion to give a final solution with an optical density that can be measured on the Spekker absorptiometer. This diluting and fractionating method unfortunately introduces the possibility of further errors in the determination of the concentration c,.With the differential technique for optical density measurements, however, the need for further dilution of the sample solution is either obviated or greatly reduced. In this technique it is only necessary to increase the concentration of the reference standard, c2, to such a value that the optical density corresponding to the concentration c1-c2 is measurable, i.e., the value is less than 1.3 if the Spekker absorptiometer is to be used. In practice it is advisable to choose a value for c2 so that the optical density reading for the concentration cl-c2 is considerably'less than the maximum value that can be read on the instrument. Further considerations on the choice of a value for c2 are given below (see p.419). The Hilger Spekker absorptiometer model H 760 is a null-reading type of instrument with a compensating photocell for the determination of the balance point. The method of operation consists in placing the most concentrated solution in the light beam whilst adjust- ing the instrument to give a zero galvanometer reading at the zero optical density setting by partially closing an iris diaphragm in front of the compensating photocell. When a less concentrated solution is placed in the light beam, the galvanometer is deflected, and it is restored to its zero value by operating the calibrated drum to close the shutter that controls the intensity of the incident light. The difference in optical density between the two solutions is found from the difference in the drum reading.Although the Beckman spectrophotometer is also a null-reading type of instrument, it only uses a single photocell and then the solution of lowest concentration is used in the balancing of the instrument at zero absorption. In spite of this major difference in operation, equations (1) and (2) are applicable to the results recorded with either instrument, as it is simply the optical density difference between the two solutions that is required in the differential technique. Hiskev based his theoretical treatment of differential absorptiometry on the Beer - A further deduction can be made from equations (1) and (2). Lambert k w , where I , = a = b = I , = which states that- - log,, abc,, intensity of the incident light, intensity of the emergent light from a solution of concentration c,, absorptivity, and optical path length.416 MILNER AND PHEKSAH: THE DETERMINATIOK OF TITANIUM IN Hiskey developed mathematically the expression-- [Vol.79 0.4343 A(?) ACl - I;- -- c1 log,, & I, where Ac, is the small error in the concentration, c,, due to A is the uncertainty of reading the exact position on the transmission scale of the absorptiometer and, in addition, that the value of A - is constant for any instrument with a linear trans- mission scale. The error in the concentration is therefore dependent on the magnitude of the error function 0’4343 and reaches a minimum value when the function attains a minimum value. As the error function is in turn dependent on the value of Il/Io, it is 0*4343 as ordinate against values of IJI, as possible to plot graphically values of .Hiskey states that A (2) (3 I1 11’ To lo€ho & I 11 I, I”, log10 - abscissa, and then the abscissa of the minimum point of this graph corresponds to the value of IJI, at which the minimum error occurs in the concentration being measured. Measure- ments of the optical density of solutions containing t.he titanium - hydrogen peroxide complex with both a Beckman and a Uvispec spectrophotonieter, and also with a Spekker absorptio- meter, gave results that when plotted produced error-function graphs that were identical for each instrument and corresponded to the curve for A = 0 in Fig. 1. The value of Il/Io for the minimum error in the concentration proved to be about 37 per cent. Hiskey extended the above treatment to the differential technique and he developed the expression- 0.4343 A(2) Ac2 - I ’ -- c2 %[1.p, 2 -t- log,, <] where I, is the intensity of the emergent light from the unknown concentration, c2, and- A being the optical density of the standard reference solution as measured against a reagent blank.Similarly, the conditions for the minimum error in the concentration are found by plotting 12/11 against the error function- Il is the intensity of the emergent light from the reference standard concentration, c,, 12/11 = 1OA - 04343 t - 0-4343 ’9 $[lOSlO I2 Tl + log10 ;F- As I z / I l is a function of A , Hiskey also studied the effect of varying the concentration of the reference standard (and therefore its optical density, A ) , the resultant graphs being reproduced in curves labelled- A = 0.1085 to A = 1.736 in Fig. 1.It is clear from Fig. 1 that the minimum value of the error function in a differential technique is always less than that for a normal absorptiometric method and it becomes increasingly less as the concentration of the reference standard is increased. A point is reached, however, when an increase in the concentration of the reference standard produces only a very small decrease in the value of the error function. The application of Hiskey’s theory to the Speltker type of absorptiometer is not im- mediately evident for several reasons. First, the design of the Spekker absorptiometer is different from that of the Beckman spectrophotonieter and the intensity of the emergent light is always adjusted to be constant.Secondly, the Spekker absorptiometer modelJuly, 19541 URANIUM - TITANIUM ALLOYS BY DIFFERENTIAL ABSORPTIOMETRY 417 H 760 utilises a logarithmic cam arrangement for altering the intensity of the incident light and this results in a linear absorption scale, while the percentage transmission scale is logarith- mic and therefore A 2 can no longer be considered constant. Hiskey’s theory must there- fore be modified to make it applicable to the Spekker absorptiometer model H 760. No attempt has been made in this paper to analyse mathematically the particular characteristics of the Spekker absorptiometer or to modify existing theory. Instead, the suitability of the G O ) I 0 J \ \ A = 0.4343 \= 1.736 1 I I 1 0.2 0.4 0.6 0.8 I 12/11 Fig.1. Curves of the error function for different concentrations of the refer- ence standard 20 40 3 al I n g 60 E + 0 .- H .- $ 8 0 390 410 430 450 470 490 Wavelength, mp Fig. 2. Transmission curve for titanium - hydrogen peroxide colour Spekker absorptiometer for differential work has been investigated experimentally by comparing the results found for the titanium content of uranium - titanium binary alloys with those found by the differential method with the Beckman spectrophotometer. Full details of the experimental work and the results found are included in the experimental section. EXPERIMENTAL DEVELOPMENT OF A SUITABLE TECHNIQUE FOR DIFFERENTIAL WORK WITH THE SPEKKER Checking the drum calibration and operation of the drum-The drum calibration of the Spekker absorptiometer used in this work was first checked in accordance with the procedure described by Taylor and Williams.ll This test was also applied to the Beckman spectro- photometer, and the results are shown in Tables I and 11.Repeated tests with the Spekker absorptiometer showed a small error in the scale calibration in the region of zero absorption, and for this reason the part of the drum scale from 0.000 to 0.050 was never used in differential work. The operation for adjusting the position of the drum was standardised in this work. The drum was always rotated in the same direction, the chosen direction being from low absorption readings towards readings of higher value. Technique to overcome cell-length variations-The selection of cells and their use in differential absorptiometry is of considerable importance.A small difference in optical path length may not be apparent when solutions of low optical densities are measured. However, when solutions of relatively high optical densities are being measured, there can be considerable variations in the optical density of a solution if different cells are used. As an example of this effect, the optical density of a titanium solution (18mg of titanium per 100ml) containing hydrogen peroxide was measured by two different methods with the Beckman spectrophotometer: (i) with .the reference standard solution in cell A and the unknown solution under analysis in cell B, when the optical density was found to be 0.167, ABSORPTIOMETER- The instrument was always initially balanced a t an arbitrary zero of 0.050.18 MILNER AND PHENNAH: THE DETERMINATION OF TITANIUM IN and (ii) with the reference standard in cell €3 and the unknown solution in cell A, when the optical density was found to be 0.166.The large error introduced by the use of badly matched cells can be readily overcome by several methods. The method that was preferred in this work consisted of reversing the order of the cells, In all measurements of the optical density values of solutions, the two chosen cells were arbitrarily marked A and B. Cell A was then filled with the standard reference solution and cell B was filled with the solution of unknown concentration and the optical density difference was measured. Optical density readings were again taken after TABLE I [Vol.79 RESULTS FOR DRUM CALIBRATION OF THE SPEKKER ABSORPTIOMETER Solution setting 0.000 0.100 0.200 0.300 0.400 0.600 0-600 0.700 0.800 0.900 1.000 1.100 1.200 Reading 0.099 0.202 0-302 0.403 0.603 0.601 0-702 0.801 0.903 1.007 1-105 1.229 1.302 OpticaJ density Individual error 0.099 - 0.004 0.102 - 0.001 0.102 - 0.001 0.103 Nil 0.103 Nil 0.101 - 0.002 0.102 - 0.002 0.101 - 0.002 0.103 Nil 0.107 + 0.004 0.105 + 0.002 0.129 + 0-026 0.102 - 0.00 1 TABLE I1 RESULTS FOR THE DRUM CALIBRATION OF THE BECKMAN SPECTROPHOTOMETER Setting 0.000 0.100 0.200 0.300 0-400 0.500 0.600 0.700 0.800 0.900 1.000 1.100 1-200 1.300 1.400 1-500 Reading 0.103 0,203 0-303 0-403 0-603 0.603 0.702 0.799 0.905 1.01 1.104 1.206 1.304 1.406 1.606 1.606 Optical density 0-103 0.103 0.103 0.103 0.103 0.103 0.102 0.099 0-105 0.11 0-104 0-105 0.104 0.106 0-106 0-106 Individual error Nil Nil Nil Nil Nil Nil -0*001 - 0.004 + 0.002 + 0-007 + 0.00 1 + 0.002 + 0.00 1 + 0.003 + 0.002 + 0.003 Remarks To nearest 0.001 19 99 99 To nearest 0.002 99 19 19 To nearest 0.005 To nearest 0.01 Range changed here To nearest 0.001 33 93 99 To nearest 0.002 n filling cell B with the reference solution and filling cell A with the unknown solution.The mean of these two readings was taken as the true osptical density difference. This measuring technique was used in the preparation of all the necessary calibration graphs. In addition, the cells used in the calibration experiments were reserved for all the subsequent work on the determination of unknown concentrations by the differential technique.Positioning the cells-The positioning of the cells in the Spekker absorptiometer cell carriage can lead to errors unless care is taken to ensure that the cells always occupy the same positions. The technique adopted with this instrument therefore consisted in always placing the cells flush with the side of the carriage nearer the light source. Also, the operation of placing the cell in the light beam was standardised in this work, the cell always being positioned by pushing the carriage towards the rear of the instrument. THE DETERMINATION OF TITANIUM- Although titanium forms several coloured complexes suitable for absorptiometric work, investigations were confined to the colour produced by titanium with hydrogen peroxide in 2.5N sulphuric acid solutions.This complex was chosen because of its stability overJuly, 19541 URANIUM - TITANIUM ALLOYS BY DIFFERENTIAL ABSORPTIOMETRY 419 reasonable intervals of time, and also because Neal and Short successfully used it for the determination of the titanium content of titanium metal with a Unicam spectrophotometer. The best solvent for uranium - titanium alloys consisted of a mixture of nitric and hydro- fluoric acids, but, after the initial dissolution, sulphuric acid was added and the alloy solution was evaporated to fumes of sulphuric acid to remove the nitric and hydrofluoric acids. The alloy solution was then diluted to give the optimum sulphuric acid normality for the production of the coloured titanium complex with hydrogen peroxide. Uranyl ions fortunately do not produce a colour with hydrogen peroxide in acid solutions and give only a small optical density reading with the conditions used for measuring the optical density of the titanium colour.The optimum wavelength for the titanium optical density measurements is in the region of 410 mp (see Fig. 2), and light of approximately this wavelength can be obtained with the Spekker absorptiometer by means of Ilford No. 601 filters used in conjunction with the mercury-vapour lamp. When the Spekker absorptiometer is used for differential work, the galvanometer deflec- tion is a major factor in determining the maximum concentration of titanium that can be used in the standard reference solution. In setting the instrument for optical density measurements, the iris diaphragm in front of the compensating photocell is closed, the calibrated drum is next adjusted to zero absorption, or another suitable value, and the light is allowed to pass through the titanium solution on to the indicating photocell. With these conditions the current in the galvanometer circuit is dependent on the intensity of the light falling on the photocell, and the recommended procedure consists in adjusting a resistor in the galvanometer circuit, known as the sensitivity control, until the galvanometer just gives a full-scale deflection.As the concentration of the titanium in the solution is increased, the intensity of the light falling on the indicating photocell is decreased, and eventually a stage is reached when it is impossible to obtain a full-scale deflection of the galvanometer even when the sensitivity control is turned fully clockwise.This state of reduced sensitivity is clearly undesirable in differential absorptiometry, and several methods are available to overcome it. For example, the intensity of the light source could be increased, or alterna- tively, a more sensitive galvanometer could be used. However, this investigation has been confined entirely to the assessment of the Spekker absorptiometer as supplied by the manu- facturers, and the titanium concentrations were chosen to produce a full-scale deflection of the galvanometer. In addition to the above conditions, it is essential that the Beer - Lambert law is obeyed. A straight-line calibration graph should be produced by plotting the optical density differences between the reference standard and the titanium solution of higher concentration against the increases in concentration over that of the reference standard.Deviations from Beer’s law usually occur when the light passing through the solution is not monochromatic, but the use of the mercury-vapour lamp in the Spekker absorptiometer should largely obviate this difficulty. The intensity of this light source is a factor determining the maximum titanium concentration of the reference standard, as it is essential to have sufficient emergent light reaching the photocell to produce a full-scale galvanometer deflection. This kind of limitation does not arise with the Beckman spectrophotometer, as in this instrument a tungsten lamp is used for work in the visible region, and the nature and intensity of the incident beam are controlled by the width of a slit.The higher the concentration of titanium in the reference standard, the greater must be the intensity of the incident beam so as to give a sufficient intensity of the emergent beam. The incident-beam intensity can only be increased with this instrument by further opening of the slit. This operation makes the light less mono- chromatic, however, and a stage will obviously be reached when Beer’s law is not obeyed. An example of the results when Beer’s law fails to apply is shown in Fig. 3. The differential method ensures that the errors from the optical density measurements are minimised as much as possible. Consequently the errors introduced into the determination of the titanium content of alloys will arise mainly from the dilutions needed to adjust the titanium concentration to a value in the region of the optimum reference standard. For the greatest precision it is essential to keep these dilutions to a minimum. When a small volume of the final coloured titanium solution was used, it was considered that suitable dilutions of the sample solution might be troublesome. Moreover, very large final volumes would be wasteful of the reagents used.For these reasons a final volume between these extremes was selected, a volume of 100 ml being considered most suitable for this determination. A good quality grade-A 100-ml calibrated flask was chosen and used for the preparation of all420 the titanium solutions.The use of the same flask throughout the experiments avoided the accurate calibration that would be necessary with more than one 100-ml flask. MILNER AND PHENNAH: THE DETERMINATION OF TITANIUM IN [Vol. 79 DETERMINATION OF THE MAXIMUM CONCENTRATION OF THE TITANIUM REFERENCE STANDARD This maximum concentration was determined with a standard titanium solution contain- ing 1 mg of titanium per ml in 2-5 N sulphuric acid. The solution was prepared from Specpure titanium metal sponge. Suitable aliquots of the solution were treated with 4ml of 100- volume hydrogen peroxide and then each was diluted to 100 mi with 2.5 N sulphuric acid. The optical density of each solution was measured in 0-5-cm cells, first at a wavelength of THAT CAN BE USED WITH THE SPEKKER ABSORPTIOMETER- Fig.3. The effect of. increasing the concentration of the reference standard as measured with the Beckmaan spectrophotometer at a wave- length of 410 m p with l-cm cells. Reference standards containing 0, 2, 4, 6, 8 and 10 mg of titanium per 60 ml were used 410 mp with the Beckman spectrophotometer, and secondly with Ilford No. 601 filters together with the mercury-vapour lamp and the Spekker absorptiometer. The optical density values for the low titanium concentrations were first measured against a 2.5 N sulphuric acid blank. A solution containing 4 mg of titanium per 100 nil was next used in place of the acid blank, and the optical density measurements were repeated. The concentration of the reference standard was then progressively increased, and so were the concentrations of the titanium solutions used for the optical density measurements.This procedure was continued until the Spekker absorptiometer failed to give a full-scale galvanometer deflection, and the experimental results were then plotted on a graph, as shown in Fig. 4. It will be seen from this figure that a linear relationship holds throughout the entire concentration range that was studied with both instruments. From the results of this work, the range from about 12 to 20 mg of titanium per 100 ml of solution wi3s chosen for the determination of titanium in alloys. Before alloys could be analysed, however, it proved necessary to prepare an accurate calibration graph over this range of concentrations. Details of the method used are as follows.A quantity of Specpure titanium metal sponge was accurately weighed (0.30137 g) and transferred to a platinum dish. A 15-ml portion of nitric acid, sp.gr. 1.42, and then about 7 ml of distilled water were added. After the addition of a few drops of 40 per cent. hydro- fluoric acid, the dish and contents were carefully :heated on a hot-plate. Further small addi- tions of water and hydrofluoric acid were made periodically until the metal had dissolved completely. The solution was then cooled, 17-5 ml of sulphuric acid, sp.gr. 1-84, were carefully added, and the solution was evaporated to strong fumes of sulphuric acid to remove all nitric and hydrofluoric acids. The solution was cooled, carefully diluted with water and then accurately made up to a volume of 250 ml.This solution contained 1.205 mg of titanium per ml. By means of a 25-ml grade-A burette, graduated in 0.05-ml steps, 12-ml to 16-ml portions of this solution were transferred in turn to the 100-ml calibrated flask.July, 19541 URANIUM - TITANIUM ALLOYS BY DIFFERENTIAL ABSORPTIOMETRY 421 After the addition of 4ml of 100-volume hydrogen peroxide, each solution was diluted to the mark with 2.5 N sulphuric acid. For the optical density measurements, the solution prepared from the 12-ml aliquot, which contained 14-46mg of titanium per 100m1, was used as the reference standard. Separate linear calibration graphs were made for the Spekker absorptiometer and for the Beckman spectrophotometer with optical density differences plotted against titanium concentrations, which were expressed as the number of milligrams of titanium per 100ml by which the solution differed from the reference standard.These graphs were then used in the following method for determining the titanium content of uranium - titanium alloys containing from 5 to 75 per cent. of titanium with an accuracy of better than &O-5 per cent. 0 4 8 I 2 16 20 24 Concentration of titanium, mg per 100 ml Fig. 4. The effect of increasing the concentration of the Full line: results with the Spekker absorptiometer, Ilford No. Broken line : results with the Beckmann spectrophotometer at reference standard. 601 filters, mercury-vapour lamp and 0-5-cm cells a wavelength of 410 mp, with 0.5-cm cells PROCEDURE FOR THE ANALYSIS OF URANIUM - TITANIUM ALLOYS SOLUTION OF ALLOY- Transfer 2 g of sample to a platinum dish and add about 7 ml of water, 15 ml of nitric acid, sp.gr.1-42, and a few drops of 40 per cent. hydrofluoric acid. Warm the mixture carefully, making further additions of reagent as required until the sample has dissolved (any green residue that remains will dissolve at a later stage during the treatment with sulphuric acid). Cool the solution, add 17.5 ml of sulphuric acid, sp.gr. 1-84, evaporate t o fumes of this acid and continue fuming for about 10 minutes. Cool the solution and dilute it accurately with water in a 250-ml calibrated flask. PREPARATION OF STANDARD TITANIUM SOLUTION- Prepare the following standard solutions of titanium by dissolving the requisite weight of Specpure metal sponge (better than 99.8 per cent.pure) according to the procedure just described: (a) 250 ml of solution containing 5.0 mg of titanium per 20 ml and (bj a range of standard solutions containing from 12 to 19mg of titanium per 20 ml. PRELIMINARY DETERMINATION OF TITANIUM CONTENT OF THE SAMPLE- Transfer a 2-ml aliquot of the alloy solution to a 100-ml calibrated flask then add 2 ml of 100-volume hydrogen peroxide and dilute the mixture accurately to the mark with 2-5 N sulphuric acid. Also, transfer a 20-ml aliquot of the standard titanium solution containing 5 mg of titanium per 20 ml to a 100-ml calibrated flask, add 2 ml of 100-volume hydrogen peroxide and dilute it to the mark with 2.5 N sulphuric acid. Measure the optical density of each solution with the Spekker absorptiometer and the mercury-vapour lamp, Ilford No.601 filters and 0-5-cm cells. Use 2.5 N sulphuric acid with 2 ml of hydrogen peroxide422 MILNER AND PHENNAH: THE DETERMINATION OF TITANIUM IN per 100ml as the blank solution. Calculate the weight of titanium in the 2-ml aliquot of sample solution from the ratio of the two optical ldensities and the weight of titanium in the aliquot of the standard solution. [Vol. 79 ACCURATE DIFFERENTIAL DETERMINATION OF THE TITANIUM CONTENT- Depending on the results of the preliminary determination, take an aliquot of the alloy solution containing from +12 to 19 mg of titanium by means of calibrated grade-A pipettes and transfer it to the 100-ml calibrated flask used in the preparation of the calibration graph. Add 4ml of 100-volume hydrogen peroxide, dilute the mixture accurately to the mark with 2.5 N sulphuric acid and, after mixing, transfer the solution to a clean, dry beaker (solution A).Take a further aliquot of the alloy solution, transfer it to the same 100-ml calibrated flask, dilute it accurately with 2.5 N sulphuric acid. and, after mixing, transfer the solution to a clean, dry beaker (solution B). From the range of standard titanium solutions previously prepared, select an aliquot containing as nearly as possible the same weight of titanium as is present in the aliquot taken from the alloy solution. Transfer this solution to the same 100-ml calibrated flask, add 4ml of 100-volume hydrogen peroxide and accurately dilute the mixture to the mark with 2-5 N sulphuric acid. After mixing, transfer the solution to a clean, dry beaker (solution C).Adjust the temperature of all solutions to the value used in the preparation of the calibration graph and make the following optical density measurements with either the Spekker absorptiometer and the mercury-vapour lamp, Ilford No. 601 filters and 0.5-cm cells, or the Beckman spectrophotometer a t a wavelength of 410mp. Use the same cells that were used in the preparation of the calibration graph- (a) Set the scale reading to a value of 0.430 and balance the instrument a t this setting with solution B. Measure the optical density due to the uranium in this solution against a 2.5 N sulphuric acid blank solution containing; 4 ml of 100-volume hydrogen peroxide per 100ml in the other cell. Interchange the solutions between the cells and repeat the measurements.Calculate the optical density for each measurement and take the arithmetic mean. (b) Set the scale reading to a value of zero or that value found most suitable in the preliminary check (we used 0.05 for our Spekker absorptiometer) with either solution A or solution C in the cell, whichever has the higher optical density. Measure the optical density difference between these two solutions. Interchange the solutions between the cells and repeat the measurements. Calculate the optical density difference for each experiment and take the arithmetic mean. Calculate the weight of titanium in the alloy sample by the methods shown in the specimen calculations below. SPECIMEN CALCULATIONS Two conditions arise in the general application of this method.(i) THE OPTICAL DENSITY OF SOLUTION A IS HIGHER THAN THE OPTICAL DENSITY OF SOLUTION C- Here the presence of uranium in the alloy solution, A, effectively increases the optical The uranium contribution must therefore density difference between the two solutions. be subtracted. Example- Optical density of solution B against sulphuric acid blank Optical density difference of solution A against solution C True optical density difference = 0.300 - 0.038 = 0.038 (mean of 4 readings). = 0.300 (mean of 4 readings). = 0.262. On referring this optical density difference to the calibration graph it was found to correspond to 3.98 mg of titanium per 100 ml. The reference standard contained 12.05 mg ofJuly, 19541 URANIUM - TITANIUM ALLOYS BY DIFFERENTIAL ABSORPTIOMETRY 423 titanium per 100ml. Therefore the aliquot of the alloy solution (in solution A) contained (12.05 + 3.98) mg of titanium.As a 25-ml aliquot of the sample solution (250ml) was taken for the preparation of solution A, the total weight of titanium in the 2-g sample was therefore- 10 x (12.05 + 3.98) mg = 0.1603 g of titanium. (ii) THE OPTICAL DENSITY OF SOLUTION c IS HIGHER THAN THE OPTICAL, DENSITY OF Here solution C must be used for the initial adjustment of the Spekker absorptiometer. The presence of uranium effectively reduces the optical density difference between the two solutions. The contribution of the uranium must therefore be added to the observed optical density difference. ExampLe- SOLUTION A- Optical density of solution B against sulphuric acid blank Optical density difference of solution C against solution A True optical density difference = 0-105 + 0.002 = 0.002 (mean of 4 readings).= 0.105 (mean of 4 readings). = 0.107. By reference to the calibration graph, this optical density difference was found to corre- spond to 1-63 mg of titanium per 100 ml. This reference standard contained 13.37 mg of titanium per 100 ml and, as this concentration proved to be higher than that in solution A, the titanium content of the sample aliquot in solution A was therefore (13-37 - 1.63) mg of titanium. With this sample it proved necessary to dilute 10 ml to 100 ml with 2.5 N sulphuric acid before a 25-ml aliquot was taken for the preparation of solution A. The total weight of titanium in the 2-g sample was therefore- 100 x (13.37 - 1.63) mg = 1.174 g of titanium.DETERMINATIONS OF TITANIUM IN ALLOYS BY THE DIFFERENTIAL METHOD SYNTHETIC SOLUTIONS- Synthetic solutions, corresponding to solutions prepared from 2-g portions of alloys covering the range 5 to 75 per cent. of titanium, were prepared from supplies of Specpure titanium metal sponge and Specpure U,O,. Both materials were dissolved separately, the titanium being first dissolved in nitric and hydrofluoric acids and then these acids being removed by evaporation to fumes with sulphuric acid. The U,O, was dissolved in nitric . acid and the resulting solution was similarly evaporated with sulphuric acid. These two solutions were then mixed and adjusted to a volume of 250ml and a normality of 2.5 in sulphuric acid.The titanium content of each mixture was then determined according to the above procedure, both the Spekker absorptiometer and the Beckman spectrophotometer being used for optical density measurements. The results of this investigation are shown in Table 111. TABLE I11 DETERMINATIONS OF TITANIUM IN SYNTHETIC TITANIUM - URANIUM SOLUTIONS Nominal composition Titanium, Uranium, 5 95 10 90 25 75 50 50 75 25 - % % f Actual composition Titanium, Uranium, f3 g 0.1039 1.8980 0.2017 1.7940 0.4994 1-4990 0.9990 0.9981 1.5004 0.501 1 L \ Titanium found with r Beckman Spekker Titanium spectrophoto- absorptio- present, meter, meter, 5.19 5.19 5.19 10.10 10.08 10.07 25.00 24.95 24-96 49-97 49.92 49-89 74-99 75.18 74-90 Yo % %424 MILNER AND PHENNAH [Vol. 79 ALLOY SAMPLES- In this work the samples were prepared in the form of thin strips instead of in the form of fine turnings. This technique was adopted to reduce the oxidation of the sample before analysis. The results for titanium by the differential technique described above are shown in columns 2 and 3 of Table IV. There was good agreement between these results and those in column 4, which were determined by a gravimetric procedure described in an earlier report .lo TABLE Iv’ DETERMINATIONS OF TITANIUM IN ALLOY SAMPLES BY THE ABSORPTIOMETRIC AND GRAVIMETRIC PROCEDURE Titanium content by analysis I L \ Absorptiometnc method -- content of the alloy, spectrophotometer, absorptiometer, Gravimetric method, % % % % Nominal titanium BeCkmtUl Spekker 10 20 60 60 8.29 19.30 46.82 58-80 8-24 19.33 46-88 68-78 8-45 19.37 46.5 69.0 Thanks are due to Mr. A. A. Smales for suggesting the application of the differential absorptiometric technique to this problem. REFERENCES 1. Hiskey, C. F., Anal. Cltcm., 1949, 21, 1440. 2. Hiskey, C. F., Rabinowitz, J., and Young, I. G., Ibid., 1950, 22, 1464. 3. Hiskey, C. F., and Firestone, D., Ibid., 1962, 24, 342. 4. Bastian, R., Ibid., 1949, 21, 972. 5. - , Ibid., 1951, 23, 680. 6. Crouthamel, C. E., and Hubbard, H. M., Report ANL-4940. 7. Neal, W. T. L., and Short, H. G., Meeting of the Physical Methods Group of the Society, London, 8. Stross, W., Metallurgia, 1949, 39, 231. 9. Pollak, F. F., and Nicholas, J. W., Ibid., 1951, 44, 319. 10. Milner, G. W. C., and Phennah, P., Atomic Ener-9 Research Establishment Report C/R 1236 11. Taylor, R., and Williams, A. F., Chem. & Ind., 1952, 43, 1061. March 3rd, 1953. See also Neal, W. T. L., Artaiyst, 1954, 79, 403. H.M. Stationery Office, 1963. ANALYTICAL CHEMISTRY GROUP ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL, NR. DIDCOT, BERKS. December lst, 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900414

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

July, 19541 MILNER AND SMALES 425 The Determination of Niobium in Stainless Steel BY G. W. C. MILNER AND A. A. SMALES A procedure is described for the determination of the niobium content of stainless steels. The niobium is first separated from many of the other constituents of the steel by precipitation with tannic acid and cinchonine from a slightly acid solution. Silica is used as a carrier for the niobium precipitate, and, after filtration and ignition, the silica is removed as its volatile fluoride. The oxide residue is next fused with potassium bisulphate, and the melt is dissolved by extraction with an aqueous solution of tartaric acid. The niobium concentration is finally determined absorptiometrically, use being made of the yellow colouration produced with potassium thiocyanate.For accurate work it is necessary to incorporate niobium-95 tracer to correct for small losses of niobium occurring in the chemical separation. However, the results for niobium in the absence of the radioactive tracer are generally satisfactory for quality control. The determination of tantalum, by radio- activation, in the separated mixed earth-acid oxides is also briefly discussed. THE methods for the determination of niobium in steels in current use in this country are gravimetric with separation procedures involving either the use of cupferron,l or hydrolysis together with the precipitation of niobium as its magnesium salt.2 The final oxide residues from such procedures contain titanium dioxide and tantalum pentoxide in addition to the niobium pentoxide. The titanium dioxide content is determined absorptiometrically, use being made of the yellow colour produced by this element with hydrogen peroxide, and the necessary correction is applied to the weight of total oxides.The tantalum pentoxide content of the steels is generally assumed to be very small, and the niobium content is then found by applying the factor 0-6990 to the weight of oxides after correcting for titanium dioxide. In these gravimetric methods it is necessary to start the analysis with 5 g of sample, and the chemical separations needed to isolate the niobium from the other steel constituents are rather tedious. One modern approach to the determination of the niobium content of materials consists in replacing the gravimetric by an absorptiometric technique so that the complete separation of the niobium from other constituents then becomes unnecessary.A method for the deter- mination of niobium in steels by this approach has just recently appeared in the American literature3; use is made of the colour produced by niobium with hydroquinone after the preliminary separation of niobium by hydrolysis. We4 have recently developed an absorptio- metric procedure for the determination of niobium in certain types of South African ores. In this method the niobium is first separated from many ore constituents by precipitation with tannic acid and cinchonine from weakly acid solutions, with silica as a carrier. The niobium in the precipitate is finally determined absorptiometricaUy by means of the niobium - thio- cyanate colour, prepared according to Freund and LevittJss recommended conditions ; small amounts of niobium lost in the chemical treatment are corrected for by the use of radioactive niobium.In the presentation of this paper at a meeting of the Society, several members asked about the possibility of applying the procedure to the determination of niobium in steels. Subsequent investigations described in this paper have shown that the ore method can be successfully applied to this problem after slight modification. EXPERIMENTAL The development of a suitable technique for the complete solution of the sample is the usual initial problem in the analysis of ores and minerals. The recommended method for South African ores and their mineral-dressing fractions, for example, involved attack on the material with hydrofluoric and sulphuric acids in addition to fusions with sodium car- bonate and potassium bisulphate.Fortunately, niobium-bearing steels dissolve on treat- ment with acids, and fusions are therefore unnecessary, so that there is a considerable saving in operational time. After the separation of niobium from many other ore constituents with tannic acid and cinchonine, tungsten and titanium proved to be the only two elements likely to cause serious interference with the absorption measurements in certain circumstances.426 MILNER AND SMALES: THE DETERMINATION OF [Vol, 79 For example, when the ratio of niobium pentoxide to tungsten trioxide is 1 to 10 in the final thiocyanate solution, an error of about + 20 per cent.results when 125 pg of niobium pentoxide is determined.5 This error becomes about + 1.6 per cent. when the niobium pentoxide to tungsten trioxide ratio is 1 to 1. Although small amounts of tungsten are generally present in stainless steels, the ratio of niobium pentoxide to tungsten trioxide in the steels available to us was much greater than N and interference from this element was therefore quite small. The colour of the titanium - thiocyanate complex is less sensitive than that of the niobium - thiocyanate complex and whereas a ratio of niobium pentoxide to titanium dioxide of 1 to 10 results in an error of + 5.6 per cent. in the determination of 125 pg of niobium pentoxide, the error is negligible when the 'ratio has a value of 1 to 1. No difficulties were expected from titanium in steel analysis, therefore, as steels that have been stabilised with niobium generally have a low titanium content.As in ore analysis, a sample weight of 0.2 g :proved suitable for the determination of niobium in steels. In preliminary experiments this weight of steel was dissolved in a platinum dish in 10 ml of diluted sulphuric acid (1 + 1) together with 5 to 10 ml of water. As radioactive niobium was incorporated with each sample so that the efficiency of the niobium separation could be determined, hydrofluoric acid also was used in the initial solution procedure to ensure the exchange between the active and inactive niobium. It proved necessary, however, to remove the hydrofluoric acid before the precipitation with tannic acid, and this was effected by evaporating the sample solution to fumes of sulphuric acid and continuing the heating for about 10 minutes.After the salts were dissolved in water, the tannic acid - cinchonine precipitation procedure was applied and the determination of niobium was completed exactly as for ore samples. Preliminary results for the B.C.S. Steel No. 246 by this procedure were in excellent agreement with the reported figure of 0.82 per cent. for niobium and, moreover, the activity measurements showed that in favourable circumstances as much as 98 per cent. of the niobium in the sample passed into the final tartrate solution. On extending the experiments to different samples, however, it often proved impossible to completely dissolve the salts after heating to fumes with sulphuric acid.This behaviour was attributed to the high chromium content of stainless steels, as chromium sulphate is only soluble in water with difficulty in certain conditions. This difficulty did not occur, however, when the sample was evaporated to fumes of sulphuric acid followed by gentle heating for a short time. Such a technique: did not ensure the complete removal of fluoride ions, and so the niobium recoveries were about 90 per cent. This trouble from chromium was not encountered when the sulphuric acid was replaced by 10 ml of perchloric acid, sp.gr. 1.70, the solution then being heated to remove the hydrofluoric acid completely. Details of the procedure incorporating radioactive niobium that was finally adopted are as follows. METHOD REAGENTS- Cinchonine sohtion-Dissolve 5 g of reagent in 100 ml of diluted hydrochloric acid Potassium thiocyanate, 3 M-Dissolve 29.0 g of this reagent in water and dilute to a volume of 100ml.Stannous chZoride, 2 M-Dissolve 4 6 g of stannous chloride and a small pellet of tin in 5 ml of hydrochloric acid, sp.gr. 1.16, and then dilute to 10 ml with water. Acetone-Use AnalaR quality reagent, which should produce no colouration when 10 m1 are mixed with a solution consisting of 10 ml of hydrochloric acid, sp.gr. 1.16, 1 ml of 2 M stannous chloride and 5ml of water. Niobium tracer-Obtain 1 millicurie of carrier-free radiochemically pure niobium-96 dissolved in a few millilitres of 8 M hydrochloric acid from the Radiochemical Centre, Amersham. Add approximately 2 ml of hydrofluoric acid to the solution, transfer it to a polythene bottle with water and dilute it to a volume of about 200ml.Dilute 25ml of this stock solution to 250 ml in a second polythene bottle for the working solution and store both solutions behind lead or brick shielding. Perchloric acid, sp.gr. 1-70. XydroJ-Euoric acid, 40 per cent. Sul$hur dioxide, saturated solutio%. Ammonium chloride. (1 + 1). .427 July, 19541 NIOBIUM IN STAINLESS STEEL Sodium hydroxide, 50 9er cent. wlv solution. Tannic acid. S U ~ ~ ~ U Y ~ C acid, ~ 9 . g ~ . 1-84. Potassium bisulphate. Tartaric acid. Nitric acid, s 9 . g ~ . 1-42. PROCEDURE- Dissolve 0.20 g of steel in 10 ml of perchloric acid, sp.gr. 1.70, together with about 6 ml of water in a suitable platinum dish (Note 1).Then evaporate to fumes of perchloric acid, cover the dish and continue the heating to oxidise the chromium. Cool the solution, add to it 5 ml of 40 per cent. hydrofluoric acid and a small volume of tracer niobium solution (Note 2 ) , and re-evaporate it to fumes of perchloric acid. Continue heating the solution for several minutes, then cool it, wash the sides of the dish with water and repeat the fuming procedure to remove the hydrofluoric acid. After cooling the residue, dissolve the salts in about 50 ml of water and transfer the solution to a 250-ml squat beaker. To this solution add 10 ml of water saturated with sulphur dioxide and then boil the solution to reduce the chromium and to remove any excess of sulphur dioxide. Add 100 mg of silica in the form of a sodium silicate solution and then 5 g of ammonium chloride, and dilute the resultant solution with water to a volume of about 100ml.Make the sample solution alkaline to litmus by the addition of a 50 per cent. w/v sodium hydroxide solution and then make it just acid with diluted hydrochloric acid (1 + l), adding 5 ml of the latter reagent in excess. Dissolve 1 g of tannic acid in the sample solution and then add 5 ml of a cinchonine solution prepared by dissolving 5 g of cinchonine in 100 ml of diluted hydrochloric acid (1 + 1). Add two macerated Whatman accelerators to the solution, heat it to boiling and continue the boiling for 10 minutes. Then allow the mixture to digest while warm for about one hour. Filter through a paper-pulp pad, washing the precipitate with a warm 1 per cent.hydro- chloric acid solution containing 1 per cent. of cinchonine. Dry the precipitate and ignite it at about 700" C in the same platinum dish. Add a few drops of diluted sulphuric acid (1 + 1) to the residue in the platinum dish, followed by about 2 ml of 40 per cent. hydrofluoric acid. Evaporate the mixture to fumes of sulphuric acid to remove silica. Cool the residue and repeat the treatment with a further 2-ml portion of hydrofluoric acid. Then heat the dish strongly to remove the sulphuric acid completely and finally ignite the dish and its contents in a Meker burner flame for a few minutes. Carefully fuse the residue with 2 g of potassium bisulphate to give a clear melt, then let it cool and dissolve the melt in a solution of 1 g of tartaric acid in 20 ml of water.Dilute this solution to a volume of 100 ml with water in a calibrated flask (solution A) and, after thorough mixing, take 10 ml of this solution and dilute it to 50 ml with a 1 per cent. w/v tartaric acid solution (solution B). Into a 50-ml calibrated flask place 10 ml of hydrochloric acid, sp.gr. 1-16, 1 ml of 2 M stannous chloride solution and 5 ml of water. Add 10 ml of acetone and mix the solution thoroughly; then cool it to room temperature by immersing the flask in a water-bath for 16 minutes. With a pipette, place 10 ml of solution B in the 50-ml flask and, after mixing, add 10 ml of 3 M potassium thiocyanate solution. Dilute the solution to the mark with water and make optical density measurements after 5 minutes with a 4-cm cell and either a spectrophotometer at a wavelength of 385 mp or a Spekker absorptiometer with Wratten No.2 and Chance OV1 filters and a mercury-vapour lamp. Correct the reading for the value found for a blank solution prepared by the above method from a 0.20-g sample of iron, Determine the amount of niobium by referring the corrected sample optical density readmg to a calibration graph prepared with a standard niobium solution containing 6 pg of niobium pentoxide per ml in 1 per cent. tartaric acid. Place 50 ml of the counting-standard solution (Note 2) in a beaker and measure its activity with a y-scintillation counter. Replace the solution in the beaker by 50 ml of the sample solution A and repeat the count. Correct all counts for the background count of the instru- ment.Correct the amount of niobium found for experimental losses by multiplying by the factor- Counts per minute of the counting standard Counts per minute of the sample solution Finally calculate the percentage of niobium in the sample.428 MILNER AND SMALES: THE DETERMINATION OF [Vol. 79 NOTE 1-If radioactive niobium is not to be added to the sample, the initial solution procedure can be considerably simplified as follows- Place 0*20g of the steel in a 250-ml squat beaker and dissolve the steel in 10ml of diluted hydrochloric acid (1 + 1) by the dropwise addition of nitric acid, sp.gr. 1-42. Then add 10ml of perchloric acid, sp.gr. 1-70, and evaporate the solution to fumes of this acid, continuing the evaporation to fully oxidise the chromium.After cooling, dissolve the salts in about 50 ml of water. To this solution add 10 ml of water saturated with sulphur dioxide, and boil it to reduce the chromium and to remove the excess of sulphur dioxide. Complete the procedure as above. NOTE 2 T h e volume of tracer solution taken should be such as to result in a 50-ml portion of the final tartrate solution having an activity of a few thousand counts per minute. An equal volume of the tracer solution should also be taken for the counting standard, placed in a 100-ml calibrated flask and diluted to the mark with 1 per cent. tartaric acid solution. RESULTS The niobium found when the procedure with radioactive niobium was applied to several typical stainless-steel samples containing from 17 to 22 per cent.of chromium and from 7 to 13 per cent. of nickel are shown in columns 2,3 and 4 of Table I. The values in column 3 show that by this type of procedure approximately 95 per cent. or more of the niobium in the sample reaches the final tartrate solution. In these circumstances it was considered that the results for niobium in the absence of tracer would be acceptable for many requirements. The initial solution procedure for the steel is considtxably simpliiied when radioactive niobium is not needed, and application of the modikd procedure described in Note 1 to the above steels gives the results in column 5 of Table I. With the exception of the British Chemical Standard sample, the steels were also analysed by Bagshawe and Elwell's gravimetric method; these results are reported in column 6 of Table I.TABLE I DETERMINATIONS OF NIOBIUM IN STAINLESS STEELS BY ABSORPTIOMETRIC AND GRAVIMETRIC PROCEDURES Results with tracer niobium Other elements r A Results in absence Gravimetric present % % % % % % Sample Relative Niobium of tracer niobium results for t7A-, number y-activity Recovery, found, .in steel, niobium Titanium Tungsten 1 .. 688 96.9 1.04 1-01 1.17 0.1s - 2 .. 673 94.8 1-28 1.17 1.40 . 0-41 - 3 .. 702 98.9 0.66 0-63 0-69 0.10 - 4 .. 677 96.4 1-03 0.98 1.16 0.06 0.05 6 .. 667 94-0 144 1.36 1.62 0-01 0.10 B.C.S. 246 707 99.6 0.80 0.79 0.82* - 0-22 * Certified gravimetric figure. Counting Standard 710 A comparison of the results for niobium found by these methods showed that those by the gravimetric procedure were generally higher than those by the absorptiometric procedures. This lack of agreement was rather disconcerting, and was most serious for samples containing more than 1 per cent.of niobium. It appeared that either the results by the absorptiometric method were erroneous, especially for the steels with the higher niobium contents, or alterna- tively that the tantalum content of these steels was significant. The behaviour of the absorptiometric method for steels containing about 1 per cent. of niobium or more was tested by taking 0.20-g portions of the steel Containing 0.82 per cent. of niobium and adding a standard niobium solution to give samples with increasing amounts of niobium up to a maximum of about 1.6 per cent. The recovery of the added niobium was always complete, which showed that the absorptiometric procedure was perfectly satisfactory for steels with the higher niobium content.The discrepancies between the absorptiometric and the gravimetric results in Table I indicated the presence of up to 0.1 per cent. of tantalum in some samples. Before attempting to confirm the presence of tantalum in these samples, however, we sought further evidenceJuly, 19541 NIOBIUM IN STAINLESS STEEL 429 in support of the absorptiometric figures in Table I. Bagshawe and Elwell's gravimetric method was applied to 5-g quantities of the steel samples and an accurately weighed quantity (30 mg) of the oxide mixture from each sample was then taken, fused with 2 g of potassium bisulphate, and extracted with 20 ml of a solution containing 5 g of tartaric acid; the resulting solution was diluted to 100 ml with water.A 10-ml portion of this solution was then diluted to 500ml with a 1 per cent. tartaric acid solution and the niobium - thiocyanate colour was produced from a 10-ml portion of this solution in the normal way. The niobium content of the 30-mg amount of mixed oxides was next determined from the optical density measure- ments and the niobium percentage was finally calculated from a knowledge of the total weight of mixed oxides found for each sample. The results by this procedure completely substantiated those given in column 4 of Table I. The tantalum content of the steels was therefore determined by the simplest available method, which was by radioactivation. This method had been used previously for relatively high levels of tantalum by Longs and Eichholz,' while one of us* has described its application to the binary mixture niobium pentoxide - tantalum pentoxide, when, for example, there was no difficulty in determining the tantalum pentoxide content of Specpure niobium pentoxide 90 steel No. 1 (0.4 per cent.).Beydon and FisherQ have also recently discussed this determination. The simplest way of applying the activation method to the steels was to irradiate weighed quanti- ties of about 20 mg of the mixed oxides produced in the gravimetric analysis carried out by Bagshawe and Elwell's procedure. These mixed oxides should contain only niobium and titanium in addition to tantalum, but because small amounts of other elements may be present, y-counting was done with a sodium iodide crystal spectrometer and a single-channel kick-sorter operating in a narrow channel centred on the 111-day half-life tantalum-182 photopeak at 1.1 MeV.This counting technique makes the direct-activation method for tantalum much more nearly specific than that previously described, but even so it is still possible to have interference from elements giving rise on irradiation to significant levels of radionuclides having y-energies greater than or near 1.1 MeV, for example, cobalt. The ideal method is, of course, to have a radiochemical separation of the tantalum after irradiation, when the activation method can be made completely specific, but this was unnecessary. The mixed oxides were irradiated in polythene tubing in the Harwell pile for two hours together with tantalum pentoxide standards.The y-spectra of an irradiated standard430 MILNER AND SMALES [Vol. 79 and sample are shown in Fig. 1. The results are shown in Table 11, together with the absorptiometric niobium figures and those for niobium plus tantalum by Bagshawe and Elwell’s method. TABLE 13 DETERMINATIONS OF TANTALUM AND NIOBIUM BY DIFFERENT METHODS Tantalum Sample determined by number activation, 1 .. 0.10 2 .. 0.10 3 .. 0-06 4 . . 0-08 5 .. 0.11 B.C.S. 246 . . 0.03 % Niobium determined by absorptiometry, including tracer, 1.04 1-25 0.65 1-03 1.44 0430 % Niobium + tantalum determined by gravimetric methods, 1.17 1.40 0-69 1.16 1.62 0-82 % CONCLUSIONS It is clear from Table I1 that the absorptiometric method with a tracer permits an accurate determination of niobium and that the apparent discrepancy between the absorptio- metric and gravimetric figures is explained by the tantalum content of the steel samples. If a specific niobium determination is of interest to metallurgical analysts, the thiocyanate absorptiometric method has a decided advantage in speed over the gravimetric method and, as is shown in Table I, even without the use of tracer niobium, results of better than 95 per cent. accuracy are attainable. REFERENCE s 1. 2. 3. 4. 5. 6. 7. Eichholz, G. G., NucZeonics, 1952, 10, No. 12, 58. 8. 9. ATOMIC ENERGY RESEARCH ESTABLISHMENT United Steel Companies Ltd., “Standard Methods of Analysis of Iron, Steel and Ferro-Alloys,” Bagshawe, B., and Elwell, W. T., J. SOC. Chem. Ind., 1947, 66, 398. Ikenberry, L., Martin, J. L., and Boyer, W. J., Anal. Chem., 1953, 25, 1340. Milner, G. W. C., and Smales, A. A., Analyst, 1954, 79, 315. Freund, H., and Levitt, A. E., Anal. Chem., 1951, 23, 1813. Long, J. V. P., Analyst, 1951, 76, 644. Smales, A. A., Proceedings of the Isotope Techniques Conference, Oxford, July, 1961, Vol. 11, Beydon, J., and Fisher, C., Anal. Chim. Acta, 1953, 8, 538. Fourth Edition, Lund Humphries and Co., Ltd., London, 1961. H.M.S.O. ANALYTICAL CHEMISTRY GROUP HARWELL, NR. DIDCOT, BERKS. January 4th, 1954

ISSN:0003-2654    

DOI:10.1039/AN9547900425

出版商:RSC    

年代:1954

数据来源: RSC