摘要:

JANUARY, 19yI THE ANALYST Vol. 79, No. 934 Editorial ON December 17th, 1953, three Special Resolutions or groups of Resolutions were passed by an Extraordinary General Meeting of The Society of Public Analysts and Other Analytical Chemists. The first Resolution changed the name of the Society to “The Society for Analytical Chemistry.” The second group of Resolutions modified the Memorandum of Association to delete the “professional” clauses and the third group altered the Articles of Association to provide for a Junior Membership for chemists aged between 18 and 27 years. These are the latest steps in the development of the Society, and it is of interest to look back for a moment as well as forward into the future. The Society of Public Analysts was founded at a meeting held on August 7th, 1874; Dr.Theophilus Redwood, whose portrait is the central feature of the Society’s badge, became the first President. In those early days membership was restricted to practising public analysts and their assistants, but in course of time analytical chemists other than public analysts were also admitted as members. When, on August 7th, 1907, the Society was incorporated, its extended membership was recognised by the adoption of the title “The Society of Public Analysts and Other Analytical Chemists.” Even then there were doubts about the suitability of this title, and to-day the Society has far outgrown its original purpose of looking after the professional needs of the public analysts. The continued growth of the Society and the varied interests of its members have made it impossible for the Council to speak with one voice for all classes of analysts on professional matters, and the Council felt very strongly that its functions in the matter of the professional needs of analysts would be very much better discharged by the newly formed “Association of Public Analysts” and by the Royal Institute of Chemistry.The new Professional Association is receiving the full support of our Society, and financial and other help has been extended to them. The membership of the Society has increased by over 50 per cent. during the last ten years. The spirit of the Society has not changed, but there is no doubt that the quarters into which its interest has been directed have altered. The developments in analytical chemistry have been so vast and striking in this period that the Society has had to grow in order to accommodate them.The inevitable feeling of regret at losing the words “Public Analysts” from the title of the Society after so many years will be softened by the knowledge that the public analysts themselves will still be with us. Whilst the Society’s interest extends over the whole range of natural and manufactured products, there will still be the same platform for the discussion of investigations into the composifion of food and drugs. The growth of analytical chemistry was reflected in 1950, when the ever-increasing number of abstracts led to their being taken out of The Analyst. For four years British Abstracts C, published by the Bureau of Abstracts, was sent to all members and subscribers with The Analyst. With the closing of the Bureau, the Society is once more undertaking their production under the title of Analytical Abstracts. The Council has for some time past considered the needs of young chemists and, with the object of encouraging them, a class of Junior Membership has been instituted. Junior Members will pay a considerably reduced subscription and no entrance fee, and will receive The Analyst. The Society for Analytical Chemistry will continue in the tradition, established in 1874 and upheld ever since then, of encouraging, assisting and extending the knowledge and study of all questions relating to the analysis, nature and composition of natural and manufactured materials generally. There is every intention that the character of The Analyst and of the ordinary meetings of the Society will be maintained unchanged. It is hoped they will attend meetings. 1

ISSN:0003-2654    

DOI:10.1039/AN9547900001

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

2 PROCEEDINGS [Vol. 79 PROCEEDINGS OF THE SOCIETY OF PUBLIC ANALYSTS AND OTHER ANALYTICAL CHEMISTS ORDINARY MEETINGS AN Ordinary Meeting of the Society, organised by the Physical Methods Group, was held at 5 p.m. on Friday, October 23rd, 1953, in the Lecture Hall, Southampton University. The Chair was taken by the President, Dr. D. W. Kent-Jones, F.R.I.C. The subject of the meeting was “Paper Electrophoresis” and the following papers were presented and discussed : “The Analysis of Inarganic Compounds by Electromigration and Electrochromatography,” by F. H. Pollard, B.Sc., Ph.D. ; “The Use of Paper Electrophoresis in the Study of Nucleic Acids,” by Roy Markham, M.A., Ph.D. ; “Paper-Strip Electrophoresis of Serum Proteins,” by A. L. Latner, M.Sc., M.D., F.R.I.C. The meeting was preceded by an afternoon visit to the Esso Refinery at Fawley.AN Ordinary Meeting of the Society was held at 7 p.m. on Wednesday, November 4th, 1953, in the Meeting Room of the Chemical Society, Burlington House, London, W.l. The Chair was taken by the President, Dr. D. W. Kent-Jones, F.R.I.C., and visitors included members of the Iron and Steel Institute, the Institution of Mining and Metallurgy, the British Iron and Steel Research Association and the British Non-Ferrous Metals Research Association. The subject of the meeting was “The Determination of Niobium in Minerals and Mineral Dressing Products,” and the following papers were presented : “The Absorptiometric Deter- mination of Niobium in Some African Low-grade Minerals and Mineral Dressing Products,” by G.W. C. Milner, B.Sc., F.R.I.C., A.Inst.P., and A. A. Smales, B.Sc., F.R.I.C. (A.E.R.E., Harwell) ; “The Absorptiometric Determination of Niobium in Some African Low-grade Ores,” by A. E. 0. Marzys, BSc., A.R.I.C. (Uganda Development Corporation Ltd.) (read by Dr. K. A. Williams, F.R.I.C., A.1nst.P.) ; “Spectrographic Determination of Niobium and Tantalum in Sukulu-type Soils,” by C. S. Campbell, M.A., and D. Nicholas (Fulmer Research Institute Ltd.) ; “Inorganic Chromatography on Cellulose. Part XIV. A Shortened Method for the Determination of Niobium and Tantalum in Minerals and Ores,” by R. A. Mercer and R. A. Wells, B.Sc., A.R.I.C. (Chemical Research Laboratory, Teddington) ; “The Colori- metric Estimation of Niobium and Tantalum with Pyrogallol,” by E.C. Hunt, B.Sc., A.R.I.C., and R. A. Wells, BSc., A.R.I.C. (Chemical Research Laboratory, Teddington) ; “Inorganic Chromatography on Cellulose. A Rapid Method for the Determination of Niobium in Low-grade Ores,’’ by E. C. Hunt, B.Sc., A.R.I.C., and R. A. Wells, B.Sc., A.R.I.C. (Chemical Research Laboratory, Teddington). Contributions to the discussion following the papers were made by E. A. Hontoir (Rio Tinto Co. Ltd.) ; L. E. Gardner (British Iron and Steel Research Association) ; W. Ramsden (British Non-Ferrous Metals Research Association) ; A. Bowes and T. Burchell (Murex Ltd.) ; A. R. Powell (Johnson, Matthey & Co. Ltd.); W. H. Bennett (Colonial Geological Surveys); B. Bagshawe (Brown-Firth Laboratories). Part XV. AN Ordinary Meeting of the Society was held at 7 p.m.on Wednesday, December 2nd, 1953, in the Meeting Room of the Chemical Society, Burlington House, London, W.l. The Chair was taken by the President, Dr. D. W. Kent-Jones, F.R.I.C., and a Lecture on “Recent Advances in Medical Chemistry” was given by Professor C. H. Gray, M.D., D.Sc., M.R.C.P., F.R.I.C., and was illustrated by lantern slides and exhibits. The Lecture was followed by a discussion. EXTRAORDINARY GENERAL MEETING AN Extraordinary General Meeting of the Society was held in London on December 17th, 1953, the President occupying the Chair. A Special Resolution was proposed that the name of the Society be changed to The Society for Analytical Chemistry and was carried by an overwhelming majority, there being 357 votes for the motion and 18 against.Of those present a t the meeting, 48 voted for the motion and 3 against. Special Resolutions were proposed for altering Clauses 3 (B) and 3 (J) of the Society’s Memorandum of Association and for deleting Clauses 3 (c) and 3 (D). These were carriedJan., 19541 PROCEEDINGS 3 by overwhelming majorities. Further Special Resolutions for altering Articles 5, 6, 10, 27 and 28 of the Society’s Articles of Association were also carried by overwhelming majorities. The effect of these alterations to the Memorandum and Articles of Association is (a) to remove the “professional clauses”; (b) to permit the transfer of a sum of money to the newly formed Association of Public Analysts, which will look after the professional interests of public analysts; (c) to establish the new class of Junior Membership of the Society and (d) to confirm the power of the Council of the Society to determine the subscription and entrance fee to be paid by members.CHANGE OF EDITORSHIP OF THE ANALYST THE Council has appointed Mr. J. B. Attrill as Editor of The Analyst from January lst, 1954. Mr. F. L. Okell will remain on the editorial staff as Advisory Editor. Mr. B. J. Walby has been appointed Assistant Editor. ANALYTICAL ABSTRACTS WE are pleased to announce the appearance concurrently with this issue of The Analyst of the first number of AnalyticaZ Abstracts, dealing with all branches of analytical chemistry, published monthly by the Society. The Council has appointed Dr. Norman Evers as Editor and Mr. B. J. Walby as Assistant Editor. NEW MEMBERS Peter Frank Sedgeley Cartwright, B.Sc. (Lond.) ; Gordon Ivor Carver, B.Sc.(Lond.), A.R.I.C. ; Alan James Dobbins; Edward C. Dunlop, A.B. (Westminster), M.S., Ph.D. (Illinois) ; Claude Douglas Essex, A.R.I.C., A.M.1nst.B.E. ; Stanley Hill, M.A. (Cantab.), F.Inst.P., F.S.S. ; Ahmad Mohammad Jawad, B.A. (Beirut); Einhart Kawerau, B.A., M.B., B.Ch., BrA.O., L.M., M.Sc., A.R.I.C. ; Basheshar Dass Kochhar, DSc. (Punjab), F.R.I.C. ; Stanley Vasa Gray Lindwall; Leonard Charles Sears Mallery ; Richard McCotter, M.Sc. (Q.U.B.) ; Arthur Thomas Ness, B.S., Ph.D. (Washington) ; Frank Henry Singleton, B.Sc. (Lond.), A.R.I.C. ; David Alan Tame, BSc. (Lond.), A.R.I.C. ; Silvio J. Tassinari, M.S. (St. Michael’s, Vermont) ; John Robert Windass, B.Sc. (Lond.), A.R.I.C. ; Douglas Arnold Yoxall, B.Sc.(Lond.), A.R.I.C. BIOLOGICAL METHODS GROUP A JOINT Meeting of the Group with the Crop Protection Panel of the Agricultural Group of the Society of Chemical Industry, the Association of Applied Biologists and the Pharma- cological Society was held an Friday, October 2nd, 1953, in the Large Chemistry Lecture Theatre, Imperial College, London, S.W.7. The meeting, which was in two sessions, took the form of a Symposium on “Organo- Phosphorus Insecticides.” Professor V. B. Wigglesworth, C.B.E., M.D., F.R.S., occupied the Chair at the morning session, which began at 10.30 a.m. The following papers were presented : “Insecticidal and Anti-Esterase Activity of Organo-Phosphorus Compounds,” by Dr. K. A. Lord and Dr. C. Potter; “Toxic Action of Organo-Phosphorus Insecticides in Mammals,” by Dr. J. M. Barnes. A discussion followed, which was opened by Dr. B. A. Kilby. The afternoon session began at 2.15 p.m. and the Chair was taken by Dr. J. R. Nicholls, C.B.E., F.R.I.C. The following papers were presented: “The Behaviour of Organo- Phosphorus Systemic Insecticides in the Living Plant,” by Dr. G. S. Hartley; “Some Hydrolytic Aspects of Organo-Phosphorus Compounds,” by Dr. P. R. Carter ; “Bio-Assay of Organo-Phosphorus Insecticides,” by Mr. J. F. Newman. A discussion concluded the meeting.

ISSN:0003-2654    

DOI:10.1039/AN9547900002

出版商:RSC    

年代:1954

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95479BX003

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

4 STILL, DAUNCEY AND CHIRNSIDE CONDUCTIMETRIC [Vol. 79 Conductimetric Determination of Carbon in Metals BY J. E. STILL, L. A. DAUNCEY AND R. C. CHIRNSIDE The metal sample is burnt in oxygen in a normal combustion tube and resistance furnace. The carbon dioxide evolved is absorbed in sodium hydroxide or barium hydroxide solution and measured by the change in electrical conductance of this solution. An improved loading device allows introduction of the sample and gives a good view of the combustion process, without admission of atmospheric carbon dioxide. A conductivity cell made of Perspex provides for continuous circulation of the absorbing solution and for measurement of the conductance at any time without stopping the oxygen flow. Except when the highest precision is required for extremely low carbon contents, sodium hydroxide solution is preferred to barium hydroxide solution.A special conductivity bridge, ordinary 50-cycle A.C. and a vibration galvanometer give high precision of measurement over a wide range; provision is made to compensate for the capacitance of the cell. The apparatus requires the minimum of maintenance, both in preparation for a series of analyses and in use. The method gives adequate precision and accuracy on 1-g samples of standard steels and cast irons containing from 0.03 to 3 per cent. of carbon, the only change necessary over this range being an alteration in solution concentration. A wide variety of samples of metals and alloys, weighing from 0.1 to 3 g and containing from 8 per cent. down to 0.01 per cent.or less of carbon, have been analysed satisfactorily. THE work described in this paper was initially undertaken in an attempt to determine carbon in pure iron at levels well below 0.02 per cent. Although carbon determinations are by no means an everyday requirement in our laboratory, the range of samples submitted to us includes cast iron, many varieties of steel, nickel and nickel alloys, chromium, cobalt, tungsten and titanium alloys, and also iron and iron alloys that have been subjected to various experimental treatments such as decarburising. The infrequency and variety of the determinations has, in the past, raised two other problems besides that of attaining sufficient precision for very low carbon contents. The first of these was the large proportion of time spent in setting up, testing and main- taining apparatus, which frequently amounted to more than that spent in analysing samples.This trouble was likely to be intensified if the determination of very low percentages required an apparatus of a type entirely unsuitable for the higher ranges. The second subsidiary problem was to find a way of decomposing the sample so that the carbon was completely recovered. With the steady improvement in quality of combustion furnaces, tubes and boats, the direct combustion of even the most refractory of alloyshas now become possible, although there are still a few alloys that need special treatment. Attention had already been focused on the need, in the steel industry, for the determina- tion of small percentages of carbon because of the stringent requirements of the electrical engineer for silicon steel of low carbon content.Two main techniques have been developed: that known as the “low pressure” method and that involving the use of a large sample, say 20 g, for the combustion, and subsequent weighing of the carbon dioxide on the usual solid absorbent. In the low pressure method, the carbon dioxide produced from the combustion of 0.5 to 2 g of sample is condensed in a liquid oxygen trap; it is then expanded into a previously evacuated vessel of known volume and its pressure is measured. Modifications of this general method have been described by a number of ~orkers~y2,3,~y~y~,~,a,~ in the U.S.A. and in this country. It is fairly rapid and has for some years been established as the most accurate method available.The large-sample methodlo is of obvious value when the sample material is freely available and the equipment is in regular use. It shares with the conventional combustion method from which it was developed the disadvantage of being inconvenient for occasional use, and it has the more serious limitation for our work that the provision of sufficient sample would be often difficult and sometimes impossible.Jan., 19541 DETERMINATION OF CARBON IN METALS 5 A third type of method,11y12J3J4 less commonly used, involves conductimetric determina- tion of the carbon dioxide produced in the combustion. For our purpose, this seemed to have advantages because- (2) A conductivity cell and bridge would form a compact unit and would need little maintenance.(ii) The sensitivity and precision of measurement would be related to the electrical techniques available, and ought therefore to be high. (iii) The carbon dioxide could be measured as and when it is evolved, without dis- connection and removal of the absorption vessel. This is an important feature in an apparatus used for a great variety of samples, as there may often be some doubt whether all the carbon dioxide has been liberated within a specified time. One such method is that of Gardner, Rowland and Thomas.l* Their apparatus includes a special device for introducing the sample without letting air into the combustion tube, and a conductivity cell made from a Friederich-type gas washing bottle with a long helical path for the gas bubbles.The carbon dioxide is absorbed in a weak solution of barium hydroxide containing a non-ionic wetting agent (Lissapol N). APPARATUS Experience over a few months with apparatus similar in all but minor details to that described by Gardner et a2. gave results generally supporting the claims made for it, and confirmed our view that this method offered a useful compromise between the simplicity of the large-sample gravimetric method, and the accuracy of the low pressure method. Partly from this experience, and partly from consideration of our own special problems, a number of improvements in the apparatus and technique seemed desirable. For example- A reduction in the number of components in the oxygen purifying train would reduce maintenance work still further.The arrangement for introducing the sample, although adequate in preventing the entry of carbon dioxide, could be improved by eliminating the greased joint and hence the risk of grease contaminating the boat, and by providing a clear view of the interior of the combustion tube. As the temperature coefficient of conductance is about 2 per cent. per " C , the extra expense and maintenance of a thermostat for the conductivity cell would probably be repaid by increased accuracy of measurement. The conductivity cell could be improved in several ways, the most important one being the avoidance of the troublesome necessity for stopping the oxygen flow and blowing the solution back into the electrode chamber before each conductance reading. Besides the saving in time and trouble, this would obviate some minor faults such as the drying out of the electrodes between readings and the occasional erratic readings caused by persistence of gas bubbles in the electrode chamber.The provision of adequate circulation of absorbing solution in the helix would be another major improvement. The incomplete absorption of carbon dioxide noted by Gardner et al. (only 98 per cent. of the theoretical under the best conditions and for not more than 096mg of carbon) seemed to merit further investigation, with possibly a radical change in the absorbing solution. If the range of the method could be extended very considerably, say to as much as 30 mg of carbon, this would make it applicable to all our carbon-in-metals deter- minations instead of its being restricted to the very low values that were of primary interest when the work was started.All these improvements were eventually effected, together with some minor ones that are apparent in the following detailed description of our apparatus. OXYGEN PURIFYING TRAIN- Four components are necessary, and sufficient, to supply the combustion tube with the required stream of pure oxygen: these are a control valve, a flowmeter, a preheater and a carbon dioxide absorption tube. In addition, a reservoir, as included in the apparatus of Gardner et al., is useful for smoothing out variations of flow during combustion of the sample.6 STILL, DAUNCEY AND CHIRNSIDE : CONDUCTIMETRIC [Vol. 79 Control valve-This is a small medical-type needle valve, mounted close to the entrance end of the combustion tube.It is supplied with oxygen at a pressure of about 10lb per sq. inch. FZowmeter-This reads in litres per hour up to a maximum of 15. Preheater-A silica tube 12 inches long and 8 inch in internal diameter is packed with platinised asbestos for a length of 6 inches; the packed section is maintained at 400" to 450" C in a small electric furnace. 0- ri nzs Swinging Con clamp Section on line A-A Glass wtcdcw Oxygen inlet' 4 Fig. 1. Combustion tube closure Reservoir-A 2i-litre bottle, with a three-hole rubber stopper, has an inlet tube passing to the bottom and an exit tube extending just below the stopper, which also carries an open mercury U-tube manometer to indicate the pressure applied to the combustion tube. Purifying tube-A glass U-tube about 8 inches long and inch in diameter is packed with soda asbestos, the first part being large granules to prevent choking of the tube, and the remainder 14 to 20 B.S.S.mesh. All these components are connected with the minimum lengths of elastic PVC tubing,* with glass connecting tubes where necessary. COMBUSTION TUBE AND FITTINGS- The mullite combustion tube is 30 inches long, of 1-inch bore, and is without a reduced The middle portion is heated in a platinum-wound resistance furnace end or a side tube. 15 inches long. * Supplied by Portland Plastics Limited.Jan., 19541 DETERMINATION OF CARBON IN METALS 7 The brass end-piece at the entrance end (Fig. 1) fits closely over the mullite tube; the joint is covered by a thin polythene sleeve, and this in turn by a longer rubber sleeve cut from a 1%-inch cycle inner-tube of good quality.A second rubber sleeve over the whole adds to the firmness of the joint and protects the inner one from perishing. All other closures, including both those that are opened each time the apparatus is used, are effected by com- pressing neoprene O-rings against glass or brass surfaces. On removal of the three large knurled nuts (2 B.A.), which run on stainless steel screws, the whole end-cap can be slid off the screws, allowing insertion of the boat with no danger of contamination. After the cap is replaced and the tube swept out with pure oxygen, the small knurled screw is loosened to enable the clamping bar to be swung aside and the plug beneath it to be removed. The small hole thus opened allows the insertion of a thin stainless steel rod for pushing the boat into the hot zone of the tube, while the oxygen supply escapes around the rod, preventing the entry of atmospheric carbon dioxide.The flat glass window gives a good view of this operation, and also of the subsequent combustion. The fitting has the additional advantage of incorporating the side tube for the delivery of oxygen, which makes the combustion tube easier to replace. The exit end of the combustion tube is closed by a rubber bung carrying a glass tap; the bung is protected from radiant heat by a platinum baffle inside the tube about 6 inches from the bung. The replacement of this bung by another metal fitting incorporating a valve has been considered, but no trouble has yet been caused by the bung.Some form of tap is necessary at this point to prevent solution from being drawn out of the conductivity cell by a momentary reduction of the pressure in the combustion tube to below that of the atmosphere. Such a reduction in pressure can occasionally be caused by the combustion of a vigorously- burning sample, unless the reservoir is made inconveniently large. A small glass tube containing a few grams of granular precipitated manganese dioxide is connected between the glass tap and the conductivity cell t o retain oxides of sulphur produced in the combustion. CONDUCTIVITY CELL- The basic principle of a long helical gas passage on the outside of a cylindrical cell, with the electrode chamber in the centre, was retained as being the best arrangement of these main parts.In addition, the following requirements were kept in mind- The gas path should be longer than that in the Friederich wash-bottle used by Gardner et al. in case the incomplete absorption noted by them was caused by too short a path. The electrode chamber must always be full of solution. This ensures that (a) readings of conductance can be taken at any time with the minimum of trouble, (b) the oxygen flow need not be stopped in taking readings, with consequent extension of the sweeping time, and (c) the electrodes do not dry out between readings. The solution must circulate continuously between the absorbing helix and the electrode chamber. This ensures (a) continuous renewal of the absorbing solution in the helix, and (b) continuous mixing of the whole body of solution without a separate operation.There must be no dead spaces where solution of a concentration different from that of the main bulk can accumulate. Gas bubbles must not pass through the electrode chamber, and any bubbles trapped there in filling the cell must be able to escape without clinging to horizontal electrode surf aces. Although both the helix and the electrode chamber must be full of solution all the time, the total solution volume must be kept to a minimum to avoid reducing the sensitivity . All solution that passes back into the gas entry tube when it is open to the air during the filling of the cell (or when the pressure in the combustion tube drops) must be expelled into the cell before the gas current starts, otherwise a serious error will be caused by absorption of carbon dioxide in this solution.The cell must be robust, easy to dismantle, and extremely easy to empty and refill with fresh solution.8 STILL, DAUNCEY AND CHIRNSIDE CONDUCTIMETRIC [Vol. 79 Materiad-It would be almost impossible to satisfy the above list of requirements with a glass apparatus, because of difficulties of fabrication. Perspex was chosen because of its trans- parency and workability; preliminary tests showed that this material was without measurable effect on the conductance of solutions of barium hydroxide or potassium chloride with which it was left in contact. Early fears that, ‘under the conditions imposed on it, the material might not be sufficiently stable dimensionally have proved groundless.General constrzcction (see Fig. 2)-There are three separable parts, all made of Perspex. The outer wall is a piece of tube 2 inches in diameter and of &-inch wall thickness; this is Rubber washer Plan view 3xvgen Oxygen outlet Solution Oxygen Solution Electrodes Vertical chamber ,ode Rubber washer Solution holes Part section in direction Z showing routes taken by the gas and by the solution into the base 0; the helix Fig. 2. Conductivity cell cemented to the base. The annular member, on the outside of which is turned the helical bubble passage, is made from rod 1.75 inches in diameter; it is bored out centrally to a diameter of 1 inch for the lower 2 inches of its length to form the electrode chamber, and above this to $ inch in diameter.This member is fastened to the base with three 2 B.A. screws having their centres on a circle 18 inches in diameter, so that it can be removed, if necessary, to give access to the gas distributor holes. The remaining part is the lid, to which is cemented the central stern of &inch rod carrying the electrodes. To empty the cell, or to remove the electrodes, it is only necessary to take out the knurled stainless steel screw from the centre of the base; this screws into the central rod, and when tightened pulls it down into close contact with the base, a thin rubber washer between theJan., 19541 DETERMINATION OF CARBON I N METALS 9 surfaces making a gas-tight fit. The length of the outer Perspex tube is so adjusted that when this bottom joint is tight the thicker and softer rubber washer under the lid is also slightly compressed; this ensures that no atmospheric carbon dioxide can enter. Because the correct functioning and convenient operation of the cell depend very much on some of the smaller details of its construction, the making of some parts is described rather minutely below.Some care was taken to avoid any operations needing elaborate equipment; the only machine used in making the cell was a 3Q-inch screw-cutting lathe. Electrode system-The electrodes are two bands of platinised platinum (;g inch wide, 0.003 inch thick, 4 inch apart) surrounding the central rod near its bottom end, and slightly recessed into its surface. The ends of these bands are folded over into a vertical milled slot, 1$ inches long, & inch wide and & inch deep.They are secured by a slightly tapered block of Perspex forced into this slot and retained by three 6 B.A. stainless steel screws, both ends of each screw being afterwards covered with cemented Perspex plugs to protect them from the platinising solution. Connections are made to the two-pin socket mounted on the lid through two $-inch holes drilled down the length of the centre rod to the levels of the electrodes. A little mercury in the bottom of each hole makes contact with the platinum through a &-inch horizontal hole, and two thick copper wires, soldered into the socket, dip into the mercury. Some cotton wool, forced into the holes around the wires, serves to retain the mercury if the cell is inverted. Gas passages-The gas inlet to the cell is a third &-inch hole passing down the central rod.At the bottom of this, it communicates through a hole in the rubber washer with a passage in the base passing outwards to a point immediately under the bottom of the helix. Correct registering of these holes in assembly is ensured by a 6 B.A. cheese-headed screw in the bottom of the central rod, on the side opposite to the gas passage. The head of this screw locates in a recess in the base, and serves both to position the two parts and to hold the washer on to the rod during assembly. inch below the surface; it is drilled from the outside before the base is cemented to the outer wall of the cell, and the outer end of the hole is filled with a Perspex plug. At the required position for the gas distributor, a hole is made from the upper surface with a &-inch end mill, which just runs into this horizontal hole.(Care is necessary to ensure that no dead spaces that could retain even a small quantity of liquid are left in the gas passages). A recess, concentric with the +inch hole and & inch or less in depth, is now made with a &inch end mill, and a *-inch diameter Perspex plug is cemented into this recess, cut off, and filed flush with the surface. The gas inlet is thus left closed by a thin disc & inch in diameter. A pattern of seven fine holes is then made in this disc by rotating an ordinary sewing needle in the lathe at a fairly high speed, and pressing the desired spot of the disc gently on to its point by means of a block or bracket fixed to the top slide or tool-post.The first hole is taken in a few thousandths of an inch at a time until it is just possible to blow through it when it is tested by mouth pressure. Then the other six are pierced to exactly the same depth, as measured by the graduations of the top slide feed screw. Drilled holes, even if made with the smallest drill commonly available, are too large to give small bubbles, but the above process is quite easy, and the holes can be cleared, if this should ever be necessary, with a similar needle mounted in the end of a thin brass rod, a little longer than the cell, that can be twirled between the fingers. A pressure of less than 1 inch of mercury should be sufficient to pass 3 litres per hour of oxygen through the cell, most of this pressure being required to overcome surface tension, As only one or two of the holes are in use at this rate of flow, only a small increase in pressure is necessary to increase the flow to 10 litres or more per hour for rapid sweeping of the tube before the analysis.Immediately above the gas distributor holes is a small vertical chamber about $ inch high, and of about the same cross-section (& inch wide x t inch deep) as the helical passage. This is cut with an end mill before the helix is turned and the helix itself starts from its upper end, This vertical part of the bubble passage ensures that when the first bubbles appear they rise into the helix instead of passing through the solution holes (mentioned below) into the electrode chamber. The helix is cut at 35 turns per inch, and the bubble path is more than 55 inches long.The gas inlet and outlet holes in the lid are provided with suitably directed nozzles made from &-inch Perspex rod drilled with a +inch axial hole; such tubes are easily bent if held The horizontal part of the passage in the base is & inch in diameter and less than10 STILL, DAUNCEY AND CHIRNSIDE : CONDUCTIMETRIC [Vol. 79 a short distance above a small gas flame. A short length of glass tubing may be attached to the exit tube to ensure that no air has time to diffuse into the cell during the few seconds while the oxygen flow is stopped. A soda-lime guard tube is unnecessary, and also undesirable, because the moist oxygen will soon cause it to become choked, and excess pressure in the cell may then force solution into the holes carrying the electrode connections.Solution circulation-The volume of solution put into the cell is such that when the helix is filled with gas bubbles there is & inch or more of solution above the top of the annular member. The solution carried up the helix by the gas bubbles can thus pass across to the inner annular space surrounding the central rod and down this space into the electrode chamber. At the bottom of this chamber the annular member is drilled with four horizontal holes, nearly equally spaced around the circumference and arranged to clear the 2 B.A. fixing screws previously mentioned. One of these holes leads directly into the small chamber iid - I 33 fi I 1 Terminals for external standard Ct =< 0.050 pF c2 = 0.020 ' c3 = 0.020 .c4 = 0.010 Cs = 0.005 cb = 0'002 c7 = 0002 ' c* == Q.oOi I. Fig. 3. Circuit diagram below the helix, and the others into a collecting channel of about ,],-inch square cross-section communicating with this chamber. About 45 ml of solution are required, and the whole of this is circulated through the cell about once a minute by the lift-pump action of the gas bubbles in the helix. THERMOSTAT- The cell is clamped by its upper end so that the part containing solution is immersed in water at 25" C. Any small thermostatically controlled bath at least 6 inches deep will serve; oiirs is a 1300-ml beaker, stirred by a stream of air from a sintered-glass diffuser andJan., 19541 DETERMINATION OF CARBON I N METALS 11 heated by a home-made 50-watt glass immersion heater.15 The control is a commercial mercury-in-glass contact thermometer with a suitable relay.A small glass cooling coil carrying mains water is used in hot weather to keep the thermostat functioning properly. MEASURING BRIDGE- This was specially made for the purpose, and comprises a control box containing the necessary switches, rheostats and so on, a four-decade resistance box of good quality for the actual measurements, and a galvanometer. The control box has flexible leads to the resistance box, to the conductivity cell, to the terminals of the galvanometer and also to its lamp transformer. Circuit-This is shown in Fig. 3. I t is basically a simple A.C. bridge arranged to read directly in micromhos of conductance up to 11,110 on the resistance box, R,.The balance point is detected by the 50-cycle vibration galvanometer, G (sensitivity 10- per micro- ampere).* The power supply is 12 volts A.C., obtained from a small transformer, which was originally built into the control box. Some anomalous behaviour of the bridge was traced, after considerable trouble, to currents induced in the control box circuits by the magnetic field of this transformer, so the transformer was removed a short distance and connected to the control box by another flexible lead. Four additions to the simple bridge circuit are described in the sections that follow. Sensitivity switch-This is a spring-off switch, S,, which protects the galvanometer by inserting a large resistance, R,, in the supply circuit while an approximate balance is attained.Standardising switch-A three-way change-over switch, S,, allows the bridge to be connected to the cell, to an internal standard conductance, R, (a good quality radio-type 100-ohm wire-wound resistor), or to an external standard by way of two terminals on the top of the control box. Standardisation is seldom necessary except when the range of the bridge is changed (see below) ; it is carried out by putting S, in the appropriate position, setting up the known value of the standard conductance on the resistance box (9452 micromhos for the 100-ohm internal standard at present in use) and bringing the galvanometer to zero by adjusting the standardising rheostat, R,. Range-changing switch-This is a double-pole switch, S,, which, when closed, increases the conductance readings by a factor of ten; in other words, the resistance box then reads in tens of micromhos up to 111,100.The standardising connections of the bridge are altered at the same time so that the same rheostat, R,, can still be used with adequate precision for standardisation (the internal standard being set up as 945 units, for instance). The setting of R, is different for the two ranges. Capacity compensation-Two decades of capacitors, C, to C,, giving a total of 0-1 10 micro- farads in steps of 0.001, are provided for this purpose. They are connected as required to balance the capacity of the cell by eight small tumbler switches on the control box. Small silvered-mica condensers are used, specially supplied? to a tolerance of -+ 1 per cent.; these are much smaller and cheaper than boxed decade capacitors, and are quite adequate for the purpose. Performance-Up to about 4000 units, there is no difficulty in reading the conductance to the nearest unit; from about 7000 up to the maximum of 11,110 there may be uncertainty amounting to f.1 unit. In fact, the precision of reading is better than the accuracy of the resistance box unless this is of quite exceptional quality. In terms of carbon content of the sample, the above figures mean that a small carbon percentage is read to the nearest 0.00005 per cent. on a 1-g sample, i.e., to the nearest half microgram of carbon, and that throughout the range the error of the electrical measurement is insignificant compared with the total of errors from other causes.The only external supply required is a connection to the A.C. mains. REAGENTS Barium hydroxide solution-A solution containing 5 g of Ea(OH),.SH,O per litre. Dilute sodium hydroxide solution-A solution containing 5 g per litre. Strong sodium hydroxide solution-A solution containing 50 g per litre. Lissapol NS-A 2 per cent. v/v solution. * Model SS1 made by H. Tinsley & Co., Ltd. 'f Supplied by the London Electrical Manufacturing Co., Ltd. $ Supplied by Imperial Chemical Industries Limited.12 STILL, DAUNCEY AND CHIRNSIDE : CONDUCTIMETRIC [Vol. 79 The exact strength of the solution in the conductivity cell need not be known; this is another advantage of the conductimetric method as regards maintenance. The solids can be weighed to the nearest 0.1 g, or to the nearest pellet of sodiumhydroxide, and the solutions can be kept in corked bottles without special precautions.ADJUSTMENT TO WORKING STRENGTHS- The above strengths of solutions are all more than five times those used in the conductivity cell. In filling the cell, the necessary amount of alkali solution (about 8ml) is put in by means of a graduated pipette, together with the necessary few drops of Lissapol solution; the cell is then counterpoised on a rough balance and filled up to the correct weight with distilled water. Two counterpoises are used; one equal in weight to the empty cell, and another equal to the weight of solution required. If a new determination of the cell constant shows it to have altered slightly, this second counterpoise can be adjusted so as to alter the volume of solution used and to keep the factor of the determination the same as before.Similarly, the cell counterpoise can be separately adjusted if any alteration is made affecting the cell weight without altering the constant, e g . , fitting a new rubber washer. ADDITION OF LISSAPOL- The quantity of Lissapol solution required is not highly critical, and is best found by trial; it varies from 1 to 5 drops. Too little alXows the bubbles to coalesce too early on their way up the helix, leading to alternate fast and slow bubbling and erratic conductance readings; too much causes persistence of the foam in the top part of the cell, and possible loss of solution through the gas exit tube. Rather less Lissapol is required for sodium hydroxide solutions than for barium hydroxide solutions.PROCEDURE It is assumed that sodium hydroxide solution is to be used in the cell; the modifications necessary when barium hydroxide solution is used are mentioned on p. 15. FILLING THE CELL- Remove the knurled screw from the bottom of the cell, and raise the lid a millimetre or so to allow the used solution to drain out. Replace the bottom screw and remove the filling plug from the lid. Put in the necessary few drops of Lissapol solution and about the right volume of sodium hydroxide solution to bring the conductance near to the top of the bridge range. Use the weak solution and the lower bridge range for less than about 3 mg of carbon, and the strong solution and the higher range for quantities between 3 and 30 mg.Place the cell on the pan of a rough balance and fill it with distilled water to balance the counterpoises already mentioned, with 2 or 3 drops in excess to allow for the evaporation that occurs during the analysis period. Replace the filling plug, put the cell in place in the thermostatically controlled bath and connect it to the combustion tube. COMBUSTION OF THE SAMPLE- Weigh a sufficient quantity (usually from & to 2 g) into a suitably prepared boat, together with any flux or igniting material necessary (see p. 14). Remove the cap from the combustion tube, put the boat just inside, and replace the cap. Increase the oxygen flow to 8 or 10 litres per hour for 5 minutes, then reduce it to 3 litres per hour and measure the cell conductance as described below.When the conductance has remained steady for 2 or 3 minutes, remove the small plug from the end cap of the tube, push the boat into the hot zone of the tube, and replace the plug. Increase the oxygen flow as necessary during the burning of the sample, observing the reservoir manometer as well as the burning process meanwhile. If, in spite of an increased oxygen supply, the reservoir pressure falls to atmospheric pressure or below, close the tap at the exit end of the combustion tube for a few moments to prevent solution from being sucked out of the cell. When combustion ceases, set the oxygen flow at 3 litres per hour, and take the con- ductance reading when it has been steady for several minutes. This is usually 25 to 30 minutes after pushing the sample into the hot zone.Jan., 19541 DETEKMINATION OF CARBON I N METALS 13 MEASURING THE CELL CONDUCTANCE- To take a reading, see that switch S, is in the position connecting the cell to the bridge, and adjust the first two dials of the resistance box to give a minimum reading on the galvano- meter.Then depress the sensitivity switch S, and adjust the resistance box and the condenser switches alternately so as to reduce the galvanometer reading to zero. When the conditions are unknown, several alternate adjustments may be necessary, but after some experience the correct capacity will be approximately known, and a very close balance can be attained quickly with the resistance box, the capacity being adjusted by a unit or two only to improve the sharpness of the zero.The capacity varies from about 0.02 microfarad with a weak barium hydroxide solution to 0.09 or more with strong sodium hydroxide solution. CALCULATION- When the conductance reading is steady, check the weight of the cell before refilling it. If it is found to be more than 0-1 g in error, e.g., because of a long idling period before the analysis, apply the appropriate small correction to the result. As the volume of solution is about 45 ml, 0-1 g represents 1 part in 450. Manipulations involving an unnecessarily large number of decimal places, particularly in the factor determinations described later, are avoided by considering the factor to be the number of micrograms of carbon per micromho change in cell conductance. This is about 0.9 for sodium hydroxide and about 0.5 for barium hydroxide solution. Then- micromhos change x factor sample weight in grams x 104 Amount of carbon, per cent.= In our experience, the factor remains constant for months at a time, even when the apparatus is not in use, provided that the cell is left filled with solution or with water. EXPERIMENTAL THE COMBUSTION PROCESS- Although the primary object of the work described was the development of the method of measuring carbon dioxide rather than investigation of the combustion procedure, some attention had to be given to the latter, largely because of the considerable variety of samples encountered. No experiments have yet been made with a high-frequency furnace, but there seems to be no reason why the conductimetric method of measurement should not be used in conjunction with such a furnace if desired.Combustion tube temperatzcre-A temperature of about 1200" C was found suitable for cast irons, iron powders and carbon steels. Stainless steels and other alloy steels required at least 1260" C, and this temperature also appeared suitable for a considerable variety of alloys, including iron, nickel, chromium, cobalt and tungsten as major constituents. Titanium was found to give low and variable results by any method that caused vigorous combustion; carbon remained in the slag and was slowly given up over a period of more than 6 hours. However, this metal gave satisfactory and reproducible results under the following conditions. The sample was introduced when the tube was at 1100" C and the temperature was raised to 1300" C.The metal oxidised steadily without igniting, and evolution of carbon dioxide ceased after about one hour. Combustion boats-These have been the source of considerable trouble, especially with low-carbon alloys where a low and consistent blank is particularly desirable. In general, boats of a relatively porous nature that do not crack on heating tend to give large and variable blanks, and non-porous boats, which give low blanks, tend to crack on heating or when the sample burns, so allowing the tube to be contaminated with slag. Our usual method is to use a small non-porous boat inside a larger one; it is then seldom that both fail badly enough for any slag to reach the tube.* For low carbon contents, it has been found essential to preheat the boats to the operating temperature in oxygen in the combustion tube, or in a similar tube, for about an hour immediately before use, and to expose them to air for the minimum period required to put in the sample. both are said to be made of aluminous porcelain, but their textures differ considerahly.* The small boats axe sold by A. Gallenkamp & Co., Ltd., the large ones by C. V. Brindley of Shefield;14 STILL, UAUNCEY AND CHIRN SIDE CONUUCTIMETRIC [Vol. 79 Fluxes and igniters-Some of the metals and alloys dealt with will burn well without any addition, and the results are not improved by the use of fluxing or igniting material. These samples include cast irons, iron powders, plain carbon steels if in a finely-divided condition, and also titanium.Metallic lead and tin have both been found to be free from large blank errors, and results have been satisfactory with either or both of these additives on all the other alloys so far analysed. Lead appears to act primarily as a flux, and is most useful with iron-rich alloys. Tin acts mainly as an igniter, and is most useful for alloys containing much nickel, cobalt or tungsten; with high-nickel alloys it has a fluxing effect as well. It was found impossible to get satisfactory blank values from analytical reagent grade lead foil by any method of cleaning with abrasives, acids or solvents. On the other hand, a very low blank value (of the order of 5 pg of carbon) is obtained by simply igniting the desired amount of lead (usually 0.7 g) to a bright red heat in a small refractory boat in a gas flame, and tipping the molten globule on top of the sample.VARIABLES AFFECTING THE CALCULATION- If the assumptions (not necessarily justified) are made that the whole of the carbon in the sample has been released into the gas stream as carbon dioxide, and that other gases and vapours that could alter the cell conductance are absent or are removed by the manganese dioxide tube, there still remain at least eight variables that can affect the relation between the weight of carbon dioxide released by the burning sample and the conductance change observed in the cell. All these have been separately investigated. Oxygen blank-If oxygen is passed through the apparatus with the combustion tube cold, the only change in the conductance of the cell, after it has reached its correct tempera- ture, should be a slow rise due to evaporation (see below).The theoretical rise in conductance is observed when the cell contains potassium chloride solution, but with barium hydroxide solution the rise is a little slower. If this difference is caused by carbon dioxide remaining in the oxygen, which is probable, the blank error from this cause does not amount to more than 0.5 pg of carbon during the period of an analysis. Tube blank-The oxygen blank is always appreciably higher when the combustion tube is hot than when it is cold. The difference might be caused by traces of combustible matter escaping the preheater, but, as it is unaffected by an increase in the preheater temperature and it increases with the combustion tube temperature above about 1000" C, we believe it to be due to diffusion of carbon dioxide through the tube wall from the higher partial pressure in the atmosphere outside.The additional blank from this cause amounts to between 1 and 2 pg of carbon during the analysis period. Boat blank-By use of suitable boats and correct pre-treatment as described above, this can be reduced to the equivalent of 4 or 5 ,ug of carbon. FZux blank-If the lead is added in the molten condition as described, or if pure tin is cut up with a clean milling cutter and subsequently washed with acetone in a Soxhlet extractor, this blank is reduced to the same order as that for the boat. These four blanks are conveniently determined together in an ordinary blank analysis ; they are subtracted from the observed change in conductance before the remainder of the calculation.Their sum is of the order of 10 pg of carbon, or 0-001 per cent. on a 1-g sample; this is reasonably small even for samples contahing little carbon, and becomes negligible in the higher ranges. The remaining variables affect the factor for the determination and cannot be included in a subtracted blank. Evaporation from the cell-This could theoretically be avoided by saturating the oxygen with water vapour before introducing it into the combustion tube, but this expedient would involve further difficulties ; for instance, the manganese dioxide tube and the connecting tubes would have to be kept warm to avoid condensation in them. The error produced is small, amounting to an increase in conductance of about 0-15 per cent.per hour, if the oxygen is assumed to be saturated on leaving the cell; it is easy to check the cell weight after an analysis, if necessary. A bsor$tion efficiency-Except when the strong barium hydroxide solutions mentioned below were used, we have no evidence that carbon dioxide is incompletely absorbed in the present cell. Although the theoretical conductance change is not quite attained even in weak barium hydroxide solution, the factor as determined is not altered by a decrease in oxygen flow-rate or by an increase in Lissapol concentration. Thesc changes, giving respectively aJan., 19541 DETERMINATION OF CARBON IN METALS 15 longer contact time and smaller bubbles, would be expected to improve absorption to a significant extent, if it were incomplete under normal conditions.Furthermore, with sodium hydroxide solutions containing a little less Lissapol than usual, it is possible to increase the oxygen flow to 6 litres per hour throughout the analysis, without measurably increasing the fact or. Cell constant-This has been determined over the whole available conductance range with potassium chloride solutions of nine different strengths from 0.001 to 0.4M. When this work was first carried out, a large rise in the cell “constant” was observed as the con- centration was increased; the cause of this was eventually found to be in the connecting leads to the electrodes, whose resistances were too high. The resistance was reduced to the lowest reasonable value, but there still remains a small error of the same nature, the cause of which is not certainly known.The cell “constant” rises steadily from 0.414 at a con- ductance of 350 micromhos to 0.457 at 110,000 micromhos. I t should be noted that these are the extreme limits of the usable range of the bridge; not more than about one-tenth of this range is used for a single analysis. A slight change in the cell constant sometimes occurs when the electrodes are re-platinised, but because the electrode stem does not have to be removed from the cell in normal use, this operation is seldom necessary. The only other occurrence that has been found to affect the cell constant is the accumulation of barium carbonate in the cell, which is caused by performing a number of analyses with barium hydroxide solution without washing out the carbonate with acid.Most of the carbonate is precipitated in the lower part of the helix, and none appears to adhere to the walls of the electrode chamber; nevertheless the constant changes slightly if more than two or three batches of barium hydroxide solution are used successively without acid rinsing. To avoid this effect, before each batch of barium hydroxide solution is emptied, about 0 6 m l of diluted hydrochloric acid (1 + 9) is put into the cell while the oxygen is still passing through it. This dissolves the carbonate in a minute or so, after which the cell is drained and washed out with distilled water before it is refilled. Equivalent conductivity-Even if the cell “constant” were truly a constant, the con- ductance change would still not be proportional to the weight of carbon dioxide absorbed, over any considerable range of concentrations of barium hydroxide solution.With a change of concentration from 1 to 5 milli-equivalents per litre, the equivalent conductivity changes from 247.0 to 240.8. This represents a proportional change about twice that due to the alteration in cell constant over the same range. The two changes affect the factor in the same direction, so that the theoretical factor rises from 0.467 to 0.487 over this range. When a similar calculation is made for sodium hydroxide solution the result is much different. It is assumed for this purpose that the conductivities of the hydroxide and carbonate, in a mixed solution, can be regarded as added together; this assumption may not be strictly justified, but it is largely borne out by experiment.The equivalent conductivities of both hydroxide and carbonate, like that of barium hydroxide, fall with increasing concen- tration, but that of the carbonate falls much the more rapidly. Hence the difference between the two equivalent conductivities (the quantity that is a measure of the carbon dioxide absorbed) rises with increasing concentration, and so affects the factor in the opposite direction to the effect of the cell constant change. As the two changes are of about equal magnitude, the “theoretical factor” (derived on the above assumption that the conductivities are additive) only alters from 0.946 to 0.938 over a tenfold increase in concentration (from 20 to 200 milli-equivalents per litre).FACTOR DETERMINATION- The factors for barium hydroxide solution and for both strengths of sodium hydroxide solution have each been determined in at least two ways. The methods used included heating small weighed amounts of calcium carbonate in the tube, burning small quantities of sucrose either weighed direct or weighed in the form of a dilute solution that was dried before burning, and evolving small amounts of carbon dioxide gas directly into the oxygen stream by measuring small volumes of standard sodium carbonate solution into a gas wash- bottle containing strong sulphuric acid through which the oxygen was passing. Platinum boats were used for heating sucrose or calcium carbonate in the combustion tube, because these give a negligible blank even with barium hydroxide solution.The changes in the factors expected on theoretical grounds were generally realised in practice. The factor for barium hydroxide solution certainly changes with concentration16 STILL, DAUNCEY AND CHIRNSIDE : CONDUCTIMETRIC [Vol. 79 at approximately the rate predicted; but the average at a particular concentration is between 1 and 2 per cent. higher than that predicted; that is, the conductance change is always a little less than expected. With sodium hydroxide, on the other hand, the factor is about 4 per cent. lower than that predicted (this is not surprising in view of the assumptions that have to be made), but it is, as expected, constant within less than 10 parts per thousand over the whole range available. COMPARISON OF BARIUM HYDROXIDE AND SODIUM HYDROXIDE SOLUTIONS- Barium hydroxide solution containing about 1 g of the base per litre was used for all the earlier work on the method, because interest was mainly centred around attaining the highest possible precision in measuring very small carbon contents. Later, when it was required to extend the range upwards, a solution containing about 3.5~: of barium hydroxide per litre was used; the conductance of this solution was as great as could be measured at that time.To our surprise, the results with this solution were low and variable, and it appears that 1 g per litre is about the maximum concentration of barium hydroxide that will absorb carbon dioxide quantitatively under our conditions. We regard 1 mg as being about the maximum amount of carbon that can be satisfactorily determined with barium hydroxide solution in the type of cell described here.Below this limit, barium hydroxide has the advantages of higher sensitivity (nearly twice that of sodium hydroxide) and somewhat higher precision of the electrical measurement (because the measure- ment is made in the lower part of the bridge range). Nevertheless, it is more troublesome to use, because the cell must be washed out frequently with acid instead of merely being drained and refilled, and the factor is more difficult to determine and to apply; instead of a single figure, a graph is required, which must be checked at a number of points. This graph is plotted with the factor as ordinate against the average conductance during the analysis as abscissa.Sodium hydroxide is necessary for all analyses in which more than 1 mg of carbon is to be determined, and unless the highest possible precision is required, we also regard it as preferable to barium hydroxide in the lower ranges. It has an additional advantage in that by using a little less Lissapol than usual a steady oxygen flow can be maintained at as much as 5 or 6 litres per hour. This decreases the sweeping time after the combustion to 15 or 20 minutes, with only a small loss of precision in the conductance measurement. TIME REQUIRED FOR ANALYSIS- The time that elapses between placing a sample in the tube and taking the final reading varies from 25 to 40 minutes, except in the rare instances when the carbon dioxide is released relatively slowly.To this must be added the time required for bringing the cell to equilibrium TABLE I CARBON IN STANDARD STEELS AND CAST IRON Carbon present according to B.A.S. A I \ Average, Range, % % Carbon found, % 0-029 0.0275 to 0.032 0.0289, 0-0283, 0.0283, 0.0285 0-366 0.360 to 0.370 0.361. 0.363, 0.363, 0.362 2-88 2.83 to 2.93 2.90, 2.89, 2.88, 2.89 with the thermostatically controlled bath and for pre-treating the boats. This can be as much as an hour if the solution is put into the cell at a temperature much below 25" C, and if the boats are to be ignited in the combustion tube used for the analysis, but the time can be much reduced by using solution and water at about the right temperature; the same boats can often be used for more than one analysis. Further time can often be saved by using the same solution for two or more analyses. (Barium hydroxide solution can be used down to 20 per cent., and sodium hydroxide down to 65 per cent., of the original conductance without loss of carbon dioxide.)Jan., 19541 DETERMINATION OF CARBON I N METALS 17 RESULTS- The results shown in Table I were obtained on two standard steels and one standard cast iron from the Bureau of Analysed Samples Ltd. All these determinations were made with sodium hydroxide solutions in the cell; the factors were the means of those from a number of determinations with both calcium carbonate and sucrose. Sample weights of both 0.5 g and 1 g were included in each set of four; no result has been discarded. The method has been used for a great variety of metals and alloys containing from 8 per cent. of carbon down to less than 0.1 per cent.; sample weights have varied from 0.1 to 3 g. In several instances the results have been checked by independent gravimetric determinations, always with good agreement. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Yensen, T. D., Trans. Amer. Electrochent. SOC., 1920,.37, 227. Ziegler, N. A., Ibid., 1929, 56, 231. Wooten, L. A., and Guldner, W. G., Ind. Eng. Chem., Anal. Ed., 1942, 14, 835. Murray, W. M., and Ashley, S. E. Q., Ibid., 1944, 16, 242. Gurry, R. W., and Trigg, H., Ibid., 1944, 16, 248. Murray, W. M., and Niedrach, L. W., Ibid., 1944, 16, 634. Stanley, J. K., and Yensen, T. D., Ibid., 1945, 17, 699. Naughton, J. J., and Uhlig, H. H., .4nal. Chem., 1948, 20, 477, Wells, J. E., J . Iron and Steel Inst., 1950, 166, 113. Bagshawe, B., “Brit. Iron and Steel Res. Ass., Proc. Fourth Chemists’ Conference,” 1950, p. 7. Cain, J. R., and Maxwell, L. C., Ind. Eng. Chem., 1919, 11, 852. Ericsson, G., Jernkontor. Ann., 1944, 128, 579. Bennet, E. L., Harley, J. H., and Fowler, R. M., ,4naZ. Chem., 1960, 22, 445. Gardner, K., Rowland, W. J., and Thomas, H., Analyst, 1950, 75, 173. Still, J. E., Chem. & Ind., 1944, 63, 294. THE RESEARCH LABORATORIES THE GENERAL ELECTRIC COMPANY LIMITED WEMBLEY, MIDDLESEX August 19th, 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900004

出版商:RSC    

年代:1954

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95479BP009

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

Jan., 19541 DETERMINATION OF CARBON IN METALS The Determination of Silicon in Tungsten and Titanium Metal Powders, Carbide Sintering Alloys, Tungstic Oxide and Tungstates BY B. BAGSHAWE AND R. J. TRUMAN A method is described for the determination of silicon in such tungsten and titanium based compounds and materials as tungstic oxide, ammonium tungstate, tungsten metal and carbide, titanium metal and carbide and mixtures of them with molybdenum carbide, tantalum and cobalt. It depends upon calcination to oxide, fusion with sodium carbonate, extraction of the fusion product under conditions that prevent adsorptive and hydrolysis losses of silicic acid and determination by an adaptation of the molybdisilicic acid - molybdenum blue reaction. Optimum conditions for the application of this reaction to the materials under review are given, and results are shown that establish the validity of the method for a series of synthetic mixtures.DURING recent years the need has arisen for an accurate means of determining the total silicon content of tungsten and allied metal powders, their carbides, and mixtures of these with other elements, such as those used in the manufacture of hard sintered alloys. Small percentages of silicon and, particularly, sodium silicate can have a pronounced effect on the properties of hard metal carbides. The silicon can be responsible for excessive brittleness and porosity, and it can also interfere very considerably with the control of grain size of the finished material, which has to be controlled very closely to ensure that these products have their required physical properties.18 BAGSHAWE AND TRUMAN THE DETERMINATION OF [Vol.79 The amount of silicon is usually small, ranging from a few tenths of 1 per cent. for some carbides to amounts below 0.01 per cent. for the purest tungstic oxide and for tungsten metal powders produced by reduction with hydrogen. It is virtually impossible to determine these small amounts of silicon with sufficient accuracy by the usual chemical separation methods, and results by these methods are usually of doubtful authenticity. Hydrolysis of silicic acid a t low silica concentrations is incomplete, and as silica is usually a minor component of a major residue of hydrolysis, its evaluation in these large residues by measurement of the loss of weight resulting from hydrofluoric acid treatment is questionable, being further complicated by the high volatility of other oxide components of the residues, notably molybdic and, to a lesser extent, tungstic oxides.In addition, precipitation of silica is usually incom- plete from solutions containing tartrate, citrate, and so on, as hydrolysis inhibitors for the other metals present. We were led, therefore, to examine the possibilities of applying the molybdisilicic acid - molybdenum blue colour reaction to the materials under review. EXPERIMENTAL Preliminary experiments in roasting the metal or carbide to form the oxide and in fusing the oxide with sodium carbonatel were carried o ~ t with the object of producing in the acidified fusion extract a complex of tungsten with phosphate, oxalate or tartrate.All these reagents, however, were shown to interfere with the formation of the molybdisilicic acid complex. Further experiments with alkali carbonate fusions of pure tungstic oxide showed that the stability of the resulting alkali tungstate was such that its acidified solution could be manipulated without the occurrence of hydrolytic decomposition and precipitation of tungstic acid. The equivalent of 0.5 g of tungsten could be rendered stable in this manner after fusion with 4 g of sodium carbonate and extraction. The stability of these solutions is attributable to the formation of complex tungstates such as Na,W,O,,, which are not readily decomposed in weak acid solution. Hence we were able to make determinations by the molybdisilicic acid - molybdenum blue reaction in the presence of all the tungsten on aliquots of suitable size from fusions of 0.5-g samples of tungsten metal or oxide. EFFECT OF OTHER ELEMENTS- The introduction of elements such as iron, cobalt and titanium into the tungsten based mixture causes low results owing to adsorption of silica in the precipitate from the aqueous fusion extract.Table I shows the results for a sample of tungsten metal powder containing 0-18 per cent. of silicon and iron in the form of ferric oxide, which was added before fusion with sodium carbonate. In each determination the iron precipitate was filtered from the aqueous extract of the fusion, again fused with sodium carbonate, and the silica was determined separately on both the primary and secondary extracts. The results show clearly that double fusion is insufficient to free the ferric hydroxide precipitate from occluded silica, the yield of silicon is never the maximum and the error increases with increasing iron additions.TABLE I RECOVERY OF SILICON FROM TUNGSTEN CONTAINING IRON AND 0.18 PER CENT. OF SILICON Silicon found Amount of Single Repeat Silicon not iron, treatment, fusion, Total, recovered, r L \ % 0-02 10 0.08 0.025 0.105 0.07 20 0.04 0.025 0.065 0.115 Y O 0.16 % % YO 5 0.145 0.015 Losses of silica in the precipitate from the fusion extract are shown in Table I1 for tungsten based materials containing 5 per cent. of cobalt or 5 per cent. of titanium as titanium carbide. It is shown also that silica is lost independently of whether the precipitate con- taining the titanium or cobalt is filtered from the alkaline fusion extract or is filtered after its acidification.Jan., 19541 SILICON I N TUNGSTEN AND TITANIUM MET.4L POWDERS 19 Sample Tungsten metal + 6 per cent. of cobalt . . . . .. Tungsten metal + 5 per cent. of titanium .. .. TABLE I1 LOSSES OF SILICA IN FUSION EXTRACTS Silicon found in fusion extract; cobalt or titanium residue removed after acidification f 1 Single Repeat treatment] fusion, % % 0.16 0.02 0.17 0.02 Silicon found in fusion extract : cobalt or titanium residue removed before acidification Single Repeat treatment] fusion, f A 7 % % 0-165 0.02 0.15 0.05 Repeated fusions were a roided by digesting the acid fi sion extract at boiling point for several minutes and oxidising with permanganate, when cobalt and most of the iron dissolve.Titanium and tantalum remain insoluble, but they do not occlude silica (see Tables IV and VI). CONDITIONS FOR COLOUR DEVELOPMENT- The conditions we finally derived for the formation of a stable molybdenum blue colour in solutions of samples rich in tungsten were similar to those put forward by Gentry and Sherrington for steels2; we utilised a combination of oxalic acid and ferrous sulphate to reduce the yellow molybdisilicic acid to molybdenum blue. We considered this reduction to be more suitable for solutions rich in tungstic acid than one based on stannous chloride, in which interference from the blue tungsten reduction product is possible. We were able to confirm in general the conditions postulated by Gentry and Sherrington, but their claim for stability of the molybdenum-blue complex is one that requires some qualification.We noted a slow but progressive deterioration of colour, which limited the permissible period for reading at maximum intensity, and we found ammonium oxalate to be more effective than its equivalent of oxalic acid in promoting colour stability, although the colour was slightly less intense for a given concentration of silicic acid. We found that a particular solution with ammonium oxalate giving a reading of 0.830 units within 5 minutes of colour development showed no change over 15 minutes and that the reading had only fallen to 0.81 units after 60 minutes, whilst its counterpart solution with oxalic acid gave an initial reading of 0.995 units after 5 minutes and had fallen to 0.925 after 60 minutes.This degree of instability is equivalent in our application of the method to a loss of about 0.015 per cent. of silicon on a 0-2 per cent. silicon content over the 60-minute period, whilst with ammonium oxalate the equivalent error is reduced to 0.005 per cent. of silicon over the same period. We found, as did Gentry and Sherrington, that colour development is not critically related to the concentration of ferrous ammonium sulphate used in the reduction, but we noted that the background colour produced in the compensating fraction increases with increase in ferrous ammonium sulphate concentration, and so for this reason we have limited the amount used to rather less than half the amount they specify.This amount has proved to be sufficient for full reduction over the whole range of products discussed in this paper. REAGENTS- with acid or water to a specific gravity of 1.125. METHOD Szclphuric acid, sp.gr. 1.125-Add sulphuric acid, spgr. 1.84, to water, cool, and adjust Potassium permanganate solution, 1 per cent.-Prepare the solution in cold water. Ammonium molybdate solution-A 5 per cent, aqueous solution of ammonium molybdate, (NH,) ,Mo,02,.4H20. Ammonium oxalate solution-A 3 per cent. aqueous solution. Ferrous ammonium sulphate solution, 5 per cent.-Dissolve crystalline ferrous ammonium sulphate, FeS0,.(NH,),S0,.6H20, in water containing 0.1 ml of sulphuric acid per 100 ml. All reagent solutions should be freshly prepared in minimum amounts for each series of determinations, as solutions stored in bulk take up silica from the containing vessels.Only the purest of reagents are suitable for this class of work.20 BAGSHAWE AND TRUMAN : THE DETERMINATIOK OF [Vol. 79 PROCEDURE- Weigh 0.5 g of sample (B.S. mesh size <loo) into a platinum dish and ignite the sample, cautiously at first, at a low temperature and finally at about 1000" C until the metal or carbide is completely converted into oxide. Roasting is facilitated if the sample is stirred at intervals with a platinum spatula or wire (Note 1). Fuse the ignited residue with 4 g of anhydrous sodium carbonate and extract the fusion products in 50 ml of water in a nickel beaker, digest for 5 minutes, add 30 ml of sulphuric acid, sp.gr. 1-125, from a burette or pipette, and immediately transfer the solution to a 260-ml beaker made of resistance glass.Stir briskly to facilitate removal of carbon dioxide (Note 2). Heat the solution, which at this stage should not exceed 120 mi, to boiling. If precipitates of iron, titanium, cobalt, tantalum and so on, are present at this stage, heat the solution at its boiling point for 5 minutes. This digestion is unnecessary for pure tungsten and molybdenum compounds. Add 1 per cent. potassium permanganate solution, dropwise, until a slight excess is present, digest for a minute or two and then discharge the excess of permanganate by adding the minimum number of drops of 5 per cent. ferrous ammonium sulphate solution. Cool the solution, transfer it to a 200-ml graduated flask, add a small amount of macerated filter- paper pulp, make up to volume with water and mix.Collect about 100 ml of clear filtrate in a dry flask by passing the solution through a dry Whatman No. 40 filter-paper. If the filtered portion is cloudy, as it occasionally is with samples containing much titanium, re-filter through another Whatman No. 40 filter-paper. Take two 40-ml fractions by pipette from the clear filtered portion and transfer them to 100-ml calibrated flasks. Treat the two fractions separately as follows. Test sohtion-Add 5 ml of 5 per cent. ammonium molybdate solution, mix and set aside for 6 minutes at 20" to 25" C. Add 10 ml of 3 per cent. ammonium oxalate solution, mix and dilute to about 90 ml with water. Add 2.5 ml of 5 per cent.ferrous ammonium sulphate solution, dilute exactly to the mark with water, mix and set aside for 5 minutes. Compensating solution-Add 10 ml of 3 per cent. ammonium oxalate solution, mix, rinse the walls of the vessel with water, add 5 ml of 5 per cent. ammonium molybdate solution and dilute to about 90 ml with water. Add 2.5 ml of 5 per cent. ferrous ammonium sulphate solution, dilute exactly to the mark with water and mix. Measure the difference of absorption between the test and compensating solutions on a photo-electric absorptiometer-a Spekker absorptiometer with a mercury-vapour lamp, 4, 2 or 1-cm cells andIlfordSpectrum yellow No. 606 and H697 or Calorex H503 filters, is suitable. The difference in readings is a measure of the silica from the sample plus minor amounts of silica introduced from the reagents or from glassware.The amount of extraneous silica introduced in this way is normally extremely small, if freshly prepared solutions of pure reagents are used. Correction for it is made by means of a blank fusion on 4 g of sodium carbonate and subsequent determination through all stages of the procedure. Evaluate the corrected readings by reference to a standard calibration graph prepared from pure reagent solutions that have been passed through all stages of the procedure, as is done for the blank solution, but which contain suitable additions of standard silicate solution (Note 3). NOTES ON PROCEDURE- 1. (a) Tungstic oxide requires only the minimum of ignition to be freed from water and traces of lower oxides.(b) Ammonium paratungstate must be ignited cautiously in the initial stages until water and ammonia are completely volatilised. (c) Finely divided metallic powders, such as tungsten and titanium powders, sometimes are oxidised violently, so the ignition should be carefully controlled in the initial stages to prevent .mechanical loss of sample or explosive attack on the platinum dish. If the powders are liable to contain hydrogen, either free or as hydride, the ignition is best performed in a nickel crucible. This should be at the lowest temperature at which oxidation is complete, following which the sample must be transferred to a platinum dish before fusion with sodium carbonate. (a) Molybdenum based materials, eg., metal, oxide, carbide and carbide mixtures containing molybdenum carbide as one of their components, should be finally ignited atJan., 19541 SILICON IN TUNGSTEN AND TITANIUM METAL POWDERS 21 about 10oO" C until the characteristic white fumes of molybdenum trioxide cease to be evolved.All or most of the molybdenum in molybdenum-based compounds is removed in this way. 2. The decomposition of fusion products from materials containing much titanium may be somewhat protracted and continued digestion after transfer of the acidified extract to the glass beaker may be necessary. 3. The standard silicate solution is prepared from pure calcined Brazilian quartz (100 B.S. mesh) as follows. Fuse 0-200 g of the prepared quartz in a platinum dish with 5 g of sodium carbonate. Extract with water in a 500-ml nickel beaker, dilute to the maximum capacity of the beaker and filter through a paper-pulp pad.Wash the filter with water containing a small amount of sodium carbonate, reject the filter-paper and residue, cool and make up the filtrate to exactly 1000 ml in a calibrated flask (I ml of this solution is equivalent to 0.0002 g of silica). For calibrations with the 2 and 4-cm cells a weaker solution is required; this can be prepared by diluting a suitable quantity of the silicate solution to one quarter of its original concentration. For use in calibration tests prepare a series of blank 4-g sodium carbonate fusion extracts, and operate as in the initial stages of the procedure as far as the point at which the solution is transferred to a 200-ml calibrated flask. Then add suitable amounts of the appropriate silicate solution before adding macerated pulp and making up to volume.The number of calibration tests and the amounts of standard silicate solution to be added to each should be arranged to cover adequately the desired working range for the materials to be examined. Take the series of calibration tests through all subsequent stages of the procedure as described in the foregoing method and correct for reagent silica and "picked-up" silica by means of a blank calibration test, i.e., a test in which no addition of standard silicate solution is made. This solution is suitable for the calibration range of the 1-crn cell. RANGE- approximate calibration ranges for 0 to 100 drum divisions- The method for 0.1-g aliquots from an initial sample weight of 0.5 g gives the following Cell size .... .. 4 2 1 Range of silicon, per cent. , . 0 to 0.12 0 to 0.24 0 to 0.47 Relatively small amounts of silica, such as those found in pure samples of tungstic oxide, are, therefore, most accurately measured in a 4-cm cell, for which the accuracy of measurement is within +_0.002 per cent. and amounts of silica as low as 0.01 per cent. can readily be determined. RESULTS The greatest difficulty in assessing the quality of the values obtained from test samples whether as oxide, metal or carbide is associated with the complete absence of reference standards and the inadequacy of the classical methods for the purpose of standardising selected samples as reference standards. This means that any sample of tungstic oxide, TABLE I11 SILICON CONTENT OF TUNGSTIC OXIDE Silicon Test added, Yo Solution containing no WO, 0 0.005 0.010 0.030 0 0.006 0.010 0.030 Solution containing 0-5 g of woti Uncorrected absorptiometer reading 0.075 (blank) 0-12 0.16 0.345 0.175 (blank) * 0.220 0.265 0.420 Silicon found, % - 0.005 0.010 0.031 0.0 12 0-017 0.022 0.040 Silicon in tungstic oxide by difference, % 0.012 0-012 0.012 0.010 * Value inflated owing to silicon in tungstic oxide.22 BAGSHAWE AND TRUMAN : THE DETERMINATION OF [Vol.79 for instance, that may be used for testing the method carries an inherent small, but unknown, silica content whose value is beyond measurement by any alternative method. Table I11 shows that known additions of standard silicate solution can be determined by the method with the same precision and accuracy in the presence of the tungstic oxide as they are in pure solutions free from tungsten; from these values the inherent silica content of the tungstic acid is found to be 0-012 per cent.of silicon to close limits of accuracy. Having established the validity of the method on pure tungsten-base material, we analysed samples of tungstic oxide and tungsten metal powder and used them as a basis for addition of cobalt, iron, titanium, and so on. Table IV shows that the equivalent of 5 per cent. of cobalt, iron and titanium added to either tungstic oxide or tungsten metal samples does not affect the results. TABLE IV EFFECT OF COBALT, IRON OR TITANIUM ADDED TO TUNGSTEN-BASED SAMPLES Specification Silicon present, O/ Tungstic oxide- Nothing added .. .. .. .. .. 5 per cent. of iron added . . .. . . 10 per cent. of iron added . . .. .. Nothing added . . .. .. .. .. 5 per cent. of iron added . . .. .. Nothing added . . .. .. .. .. 5 per cent. of iron added . . .. .. 5 per cent. of cobalt added . . .. .. 5 per cent. of titanium added . . .. .. Tangsten metal- /O 0.15 0.15 0.15 0.06 0.06 0.17 0.17 0.17 0.17 Silicon found, % 0.15 0.15 0.145 0.06 0-07 0-17 0.17 0.17 0.18 The stability of the molybdenum blue in tungsten-based solutions is illustrated by the figures in Table V, which show that there is no loss of colour over a period of 60 minutes. TABLE V STABILITY OF MOLYBDENUM BLUE IN TUNGSTEN-BASED SOLUTIONS Sample Absorptiometer readings after r \ Silicon 5 minutes 60 minutes found, A O/ Tungstic oxide containing 0-15 per cent.of Tungstic oxide containing 0.15 per cent. of Tungstic oxide containing 0.06 per cent. of silicon and 5 per cent. of iron . . .. 0.69 silicon and 10 per cent. of iron . . .. 0.66 silicon and 5 per cent. of iron . . .. 0.35 0.69 0.67 0.35 /O 0.15 0.145 0-07 In applying the method to titanium based materials, we are able to make use of a sample of titanium carbide, Tic, containing 0.4 per cent. of silicon, on which we established an accurate independent value gravimetrically by the method of sulphuric acid dehydration. The following values show that the proposed method gives results of comparable accuracy and reproducibility- Silicon by gravimetric method, per cent. . . . . 0.39 0.40 0-40 0-40 0.41 Silicon by proposed method, per cent.. . . . 0.39 0-40 0.39 0-39 0.41 This sample of titanium carbide and samples of tungsten carbide powder, WC, and molybdenum carbide, Mo,C, were used as bases for the preparation of a series of synthetic carbide mixtures on which the validity of the method was finally proved. The results are shown in Table VI.Jan., 19541 SILICON IN TUNGSTEN AND TITANIUM METAL POWDERS TABLE VI RESULTS BY THE PROPOSED METHOD ON SYNTHETIC MIXTURES OF CARBIDE Mixture Tungsten carbide (WC) . . .. .. .. .. .. Titanium carbide (Tic) . . .. .. .. .. .. Molybdenum carbide (Mo,C) . . . . .. .. .. 60 per cent. of Tic + 40 per cent. of Mo,C . . .. .. 50 per cent. of Tic + 60 per cent. of Mo,C . . .. .. 20 per cent. of Tic + 80 per cent. of Mo,C 80 per cent. of Tic + 20 per cent.of Co (containing.0-10 pk; 40 per cent. of Tic 40 pe; cent: bf MoC + 20 per cent. of & 80 per cent. of WC + 20 per cent. of Ta as T+O, (containing 60 per cent. of WC + 20 per cent: ‘of Tic’+ 2 0 b r cent. of Ta 60 per cent. of WC + 20 per cent. of Tic + 20 per cent. of Ta . . cent. of Si) 0.004 per cent. of Si) . . .. .. + 10 per cent. of Co . . .. .. .. .. .. Silicon present, % 0.25 0.40 0.025 0.255 0.2 1 0.10 0-34 0.19 0.20 0.23 0.225 Silicon found, % 0-260 0.405 0.025 0.252 0.19 0.106 0.335 0.18 0.19 0.23 0.22 23 Typical values for silicon in various Tungstic oxide . . .. .. Ammonium paratungstate . . Hydrogen reduced tungsten metal Carbon reduced tungsten metal Tungsten carbide .. .. Titanium metal . . .. .. Titanium carbide .. .. Molybdenum carbide . . .. production samples were as follows- .. . . 0.009 to 0.017 per cent. of silicon .. . . 0 to 0.007 per cent. of silicon .. . . 0.006 to 0.08 per cent. of silicon .. . . 0.16 to 0-18 per cent. of silicon .. . . 0-005 to 0.007 per cent. of silicon .. . . 0.04 to 0-31 per cent. of silicon .. . . 040 per cent. of silicon .. . . 0-025 per cent. of silicon The authors wish to thank Dr. C. Sykes, F.R.S., for permission to publish this paper. Acknowledgment is also given to Mr. H. Burden of Firth Brown Tools Ltd. for his encourage- ment and interest throughout the course of the investigation. REFERENCES 1. Gentry, C. H. R., and Sherrington, L. G., “XIth Congress of Pure and Applied Chemistry”; Brit. 2. -,- , J . SOC. Chem. Ind., 1946, 65, 90. Abstr. C, 1948, 86. THE BROWN - FIRTH RESEARCH LABORATORIES PRINCESS STREET SHEFFIELD, 4 J d y 17tk, 1963

ISSN:0003-2654    

DOI:10.1039/AN9547900017

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

24 €OOK, STITCH, HALL AND FELDMAN: THE [Vol. 79 The Fractionation of Urinary Neutral 17-Ketosteroids by Adsorption and Partit ion Chromatography BY E. R. COOK, S. R. STITCH, A. E. HALL AND MARY P. FELDMAN The fractionation of urinary neutral 17-ketosteroids by adsorption and by partition chromatography is described, and the close similarity of certain fractions obtained by both methods is illustrated. Despite the influence of the method of hydrolysing the 17-ketosteroid conjugates upon the chromato- graphic pattern, the constancy of the patterns for the urines from some subjects over a period of weeks is shown by each fractionation procedure. A METHOD for separating urinary 17-ketosteroids into eight fractions by adsorption chromato- graphy on columns of alumina has been described by Dingemanse, Huis in't Veld and de Laat,l and by Dingemanse, Huis in't Veld and Hartogh-Katz.2 The procedure is somewhat lengthy for clinical purposes, and micro-modifications of the technique have been described by Zygmuntowicz, Wood, Christ0 and Talbot,3 by Pond4 and by Wilkins and Carlson.5 Rubin and Rosenkrantzs use silica gel columns, and the principle of partition chromatography has been applied to the separation of 17-ketosteroids by Jones and Stitch.' The partition chromatographic method requires no change of eluting solvent system and can be used with an automatic fraction collector.This technique has been compared with the longer but otherwise satisfactory method of Pond.* REAGENTS- Benzene-Analytical reagent grade benzene is unreliable in quality for chromatographic requirements, and the solvent must be washed with water and fractionally distilled, the middle fraction being then stored over anhydrous analytical reagent grade sodium sulphate until required.Carbon tetrachEoride-Hopkin and Williams technical grade carbon tetrachloride gives satisfactory reagent blanks after being washed with water and fractionated, the middle fraction being then dried with anhydrous analytical reagent grade sodium sulphate. Other reagents are purified by the methods described by Cook.* PROCEDURE- Samples of urine are hydrolysed by the method of Callow, Callow, Emmens and Stroud.9 Add 37 ml of concentrated hydrochloric acid and 50 ml of carbon tetrachloride to 250 ml of urine and heat under reflux for 1 hour on a bath of vigorously boiling water.Cool, remove the carbon tetrachloride and repeat the procedure with two more 50-ml quantities of carbon tetrachloride. Combine the extracts and wash them once with 30 ml of 2 N sodium hydroxide, once with 30 ml of 2 N sodium hydroxide saturated with sodium hydrosulphite (shake very hard until the organic solvent layer is decolorised), once more with 30 ml of 2 N sodium hydroxide, and three times with 30 ml of distilled water. Dry over anhydrous sodium sulphate, filter, distil to dryness and apply suction from a water pump to remove the last traces of carbon tetrachloride. Dissolve the residue in 10 ml of benzene, use a pipette to put duplicate 0.4-ml aliquots into test tubes, evaporate the aliquots to dryness in an oil-bath at 100" to 105" C, and again apply suction.Estimate the steroid in the dry residue by means of the ZimmermanlO reaction as modified by Callow, Callow and Emmens.11 An aliquot of the extract containing from 1.5 to 2.0 mg of 17-ketosteroid is then chromato- graphed by the method of Pond.* It is found necessary to decrease the vacuum applied to the columns from 30 cm of mercury to 10 to 15 cm of mercury and to increase very slightly the volume of eluting liquid, developing the columns with- ADSORPTION CHROMATOGRAPHY 5 x 1Oml of benzene (including the extract volume), 14 x 10ml of benzene containing 0.05 per cent. of ethyl alcohol, 12 x 10ml of benzene containing 0.1 per cent. of ethyl alcohol, 12 x 10ml of benzene containing 0.5 per cent. of ethyl alcohol, 6 x 10 ml of benzene containing 1.0 per cent. of ethyl alcohol, 1 x 1 O m l of ethyl alcohol.Jan., 19541 FRACTIONATION OF URINARY NEUTRAL 17-KETOSTEROIDS 25 The eluates are evaporated to dryness and the residue is determined by the Zimmerman reaction.Under these conditions, reproducible chromatograms can be obtained (Table I), although only one batch of alumina has so far been used. By analogy with the method of Dingemanse, Huis in't Veld and Hartogh-Katz,2 it appears probable that the 17-ketosteroids eluted in each peak are- Fraction I-Hydrolytic artefacts 3/3-chloro-androst-5-en-l7-one, and androst- 3 : 5-dien-17-one (formerly 3-chloro-A5-androstenone-17 and A3'5-androstadienone-17, respectively). Fraction 11-3 : 5-cycZoAndrosten-6-ol-17-one (i-androstan-6-01-17-one).Fraction 111-The /3-hydroxy-17-ket osteroids, androst-5-en-3/3-01-17-0ne and androstan-3/3-01-17-one (dehydroisoandrosterone and isoandrosterone) . Fraction IV-Androstan-3a-ol-17-one (androsterone). Fraction V-Testan-3a-01-17-one (etiocholanolone). Fraction VI-Androstan-3a : 1 la-diol-17-one (1 1-hydroxyandrosterone) . Fraction VII-Testan-3a : 1 la-diol-17-one (1 1-hydroxyetiocholanolone) . Fraction VIII-Unidentified. Our own studies with pilot substances confirm the elution of these 17-ketosteroids in fractions I , 11, 111, IV and V of Pond chromatograms, but the 11-hydroxy-17-ketosteroids have not been available for pilot studies of fractions VI and VII. PARTITION CHROMATOGRAPHY The technique used is that of Jones and S t i t ~ h , ~ which is briefly described here.An aliquot containing 5 mg of the crude total 17-ketosteroids is purified by the Girard T separation described by Reiss, Hemphill, Gordon and Cook,12 and approximately 2.5 mg of the ketonic fraction are evaporated to dryness and dissolved in 0.2 ml of pure alcohol and 5 ml of the mobile phase described below. This solution is then loaded on a previously prepared silicic acid column containing nitromethane as stationary phase, and the chromato- gram is developed with a mobile phase consisting of 3 per cent. v/v of chloroform in light petroleum, boiling range 60" to 80" C, saturated with nitromethane. The eluate is collected in 5-ml fractions by the automatic receiver changer of Grant and Stitch.13 The fractions are evaporated to dryness in an oil-bath at 100" C and then reduced pressure is applied from a water pump to remove the last traces of nitromethane; the dry residue is estimated by the Zimmermann reaction.Studies with pilot substances show that 3/3-chloro-androst-5-en- 17-one is eluted in peak I , 3 :5-cycZoandrosten-6-ol-l7-one and androstan-3a-ol-17-one in peak 111, androstan-3/3-01-17-one in peak IV, androst-5-en-3/3-01-17-0ne and testan3a- 01-17-one in peak V, and androst-5-en-3:17-dione in peak VI. 3 : 5-cycZoAndrosten-6-ol-l7-one is only encountered in extracts of neutral urine and may be an artefact formed by the decomposition of androst-5-en-3/3-~1-17-one sulphate (Jones and Stitch7). There is also a partial separation of androst-5-en-3/3-01-17-one and testan-3a-ol-17-one in peak V, especially if the /3-17-ketosteroid content is high.THE COLORIMETRIC DETERMINATION OF 17-KETOSTEROIDS Crude urinary extracts contain non-specific chromogenic material that intensifies the colour produced in the Zimmermann reaction. These chromogens are almost entirely non-ketonic in nature; either they can be removed by the Girard T chemical separation (Girard and Sandulesco,l4 Talbot , Butler and MacLachlan,lS Pincus and Pearlman,l* Cook* and Brayl7) or their interference can be allowed for by means of a correction factor, which makes use of the different absorption maxima shown by the non-ketonic fraction and pure 17-ketosteroids (Gibson and Evelyn,l* Fraser, Forbes, Albright, Sulkowitch and Reifenstein,lg Talbot, Berman and MacLachlan,20 Engstrom and Mason,21 Hamburger22 and Allen23).Jones and Stitch' use a purified 17-ketosteroid extract for fractionation, and correction for interfering chromogens is then unnecessary. The non-specific material present in the eluates from the Pond technique must be allowed for by colour correction methods; to reduce the calculations required to a minimum, a nomogram has been constructed and has been described in detail by Cook and Rooks.24 PROCEDURE- 2.5 N potassium hydroxide to the dry residue in the test tube. Add 0.4 ml of alcoholic 1 per cent. m-dinitrobenzene solution and 0.2 ml of alcoholic Make up a similar reagent26 COOK, STITCH, HALL AND FELDMAN: THE [Vol. 79 blank. Shake the tubes thoroughly, stopper them firmly and incubate them in the dark at 25" 0.1" C for 60 minutes. Dilute the reaction mixture with 10 ml of absolute alcohol, and determine the colour extinction against the reagent blank in a photo-electric absorptio- meter with an Ilford No.604 spectrum green filter. An additional reading must be made with an Ilford No. 601 spectrum violet filter for eluates from the Pond columns. The 17-ketosteroid content of each tube can be calculated by reference to a standard graph or a nomogram. RESULTS The reproducibility of chromatograms formed by each chromatographic method was examined by carrying out duplicate fractionations of the same urinary extracts; the results, as Table I shows, were in close agreement. TABLE I REPRODUCIBILITY OF CHROMATOGRAPHIC PATTERNS OBTAINED BY THE METHODS OF POND4 AND JONES AND STITCH' Distribution of 17-ketosteroids in fraction Chromatographic h 'I Urine technique I I1 111 IV V VI VII VIII A Pond 4.4 - 19.5 39.0 24.8 11.4 - 0.9 A Pond 4.9 - 19.9 37.2 26-1 11.0 - 0.9 A Dingemanse ef ul.*v2 6-1 - 19.4 33.4 32.0 8.6 - 1.5 % % $10 % % Y O YO % I3 Pond B Pond 3.2 42.9 2.4 27-3 15.9 7.4 0.2 0.7 3.9 42.8 2.7 27.2 14.1 7.6 0.5 1.2 C Jones and Stitch 41.5 7.6 20.0 - 26.8 4.1 C Jones and Stitch 44.0 5.9 16.7 - 26-7 6-7 Other urine specimens were divided into aliquots, which were then separately hydrolysed and extracted by the method of Callow, Callow, Emmens and S t r o ~ d , ~ previously described, and the extracts were chromatographed by both adsorption and partition techniques.In some instances urine aliquots intended for fractionation by the adsorption chromatographic method were extracted by the neutral procedure of Dingemanse, Huis in't Veld and Hartogh-Katz2 before acid hydrolysis.Table I1 shows the chromatographic patterns obtained for ten urines. TABLE I1 DISTRIBUTION OF 17-KETOSTEROIDS IN CHROMATOGRAMS PREPARED BY THE ADSORPTION TECHNIQUE OF POND AND THE PARTITION TECHNIQUE OF JONES AND STITCH Fraction by Pond technique A r Urine I I1 I11 IV V VI % % % % % % A1 5-0 4.8 2.9 63.0 17.6 6.2 A2 4-8 10.0 2.3 58.8 16.7 7.0 A3 6-1 9-2 2.9 58-5 13.4 9.9 B 10.2 4-2 1.4 37.7 36.3 9.3 C 3.3 - 9.9 19.2 51-0 12.9 n 4.4 - 8.9 30.9 35.2 16.8 E 7.4 5.4 6-1 27.2 39.1 10.5 F1 9.6 - 13.7 34.6 24.9 11.3 F2 14.5 - 18.1 28.6 29.4 8.1 F3 13.4 2.1 12.7 32.4 26.4 9.9 VI I % - 1.7 1.6 2.0 1.9 - Fraction by Jones and Stitch technique VIII I I1 I11 IV v VI 0.5 20.6 3.2 48.0 - 23.1 5.1 0.4 14.6 3.0 51.5 - 25.0 5.9 24.8 4.7 40.5 - 24.9 5-1 2.0 7.6 14.5 24.0 - 51.2 2.7 2.2 15.7 14.0 38.2 - 28.2 3.9 2.3 10.8 2.9 28.8 - 51-0 6.5 4.0 8.5 7.5 33.6 - 42-5 7-9 1.3 15.8 5.1 32.8 - 40.9 5.4 3.1 17.8 6.0 37.3 - 31-6 7.3 -- r h \ % % % % % % % - 0.9 35.1 11.4 32.5 - 31.0 - DISCUSSION OF RESULTS It is well known that the severe acid hydrolysis necessary to liberate free 17-ketosteroids from their water-soluble conjugates may cause considerable destruction or transformation , especially of androst-S-en-3fl-ol-17-one (formerly called dehydroisoandrosterone) mostly by the substitution of the 3/3(0H)-group with chlorine (Talbot, Ryan and Wolfe,25 Bitman and Cohen,26 Landau, Knowlton , Lugibihl and MunsonJ2' Butenandt and Dannebaum,28 Butenandt, Dannebaum, Harrish and Kundszu~2~ and Venning, Hoffman and Browneao).Jan., 19543 FRACTIONATION OF URINARY NEUTRAL 17-KETOSTEROIrbS 27 Transformations may take place (a) during the process of deconjugation, (b) in the hot aqueous acid suspension of the free steroid and (c) to a very slight extent by the action of aqueous acid on an organic solvent containing the extracted free 17-ketosteroid. Many methods, differing widely in their hydrolysis and extraction conditions, have been published for the determination of total 17-ketosteroids, and we have found that five of these techniques give similar results.However, the technique of hydrolysis and extraction becomes important when fractionation is required, and Table 111 shows the difference in chromatographic patterns that can be obtained by different hydrolysis procedures.The figures for urines A and B in Table 111 show clearly that the method of independent hydrolysis followed by solvent extraction, used by many workers, produces a high proportion of artefact material, mostly by attack on the p-17-ketosteroids in fraction 111 (Dingemanse technique). TABLE I11 INFLUENCE OF THE HYDROLYTIC METHOD ON THE CHROMATOGRAPHIC PATTERN Chromato- Distribution of 17-ketosteroids in fraction Method of hydrolysis graphic I- A \ Urine of urine technique I I1 I11 IV V VI VII VIII A A B B C C Yo % Yo Yo % % Yo x (a) Heat under reflux with Dingemanse 18-9 - 16.3 35.4 18.6 8.1 - 2.7 16 per cent. v/v of conc. HCl for 16 minutes, cool, extract with ether 26 per cent. v/v HC1 and heat under reflux for two periods of 6 hours with 40 per cent.v/v of benzene et al. Y Y 1.2 ( b ) Add 16 per cent. v/v of 12.0 - 26.1 34.8 17.6 8.3 - (a) As in A (a) (b) Add 16 per cent. v/v of conc. HCl and heat under reflux for three periods of 1 hour with 20 per cent. v/v of carbon tetrachloride Y Y 3 Y 26.6 - 5.5 37.1 26.0 4.5 - 0-3 11.1 - 17.9 36.4 25.9 7.8 - 0.9 3.7 41.2 2.0 3i.3 17.6 7.6 - 0.6 (a) Reflux neutral urine for two periods of 6 hours with 40 per cent. v/v of benzene, then hydrolyse as in B (b). Combine extracts Pond ( b ) As in B (b) >7 15.1 - 25.1 26.9 18.3 13.2 - 1.4 Alternative and gentler methods of hydrolysis by the use of /3-glucuronidase prepara- tions are still being studied; a t present they require a minimum of 50 to 60 hours for complete deconjugation of the 17-ketosteroid glucuronides (Buehler, Katzman, Doisy and D ~ i s y , ~ ~ Bitman and Cohen,s2 Buehler, Katzman and Doisy,33 Kinsella, Doisy and Glicka and Cohens). Cohen= states that these extracts contain considerably less non-specific chromogenic material than do extracts from acid-hydrolysed urines.Recently, Henry and T h e ~ e n e t ~ ~ and Henry, Thevenet and Jarrige3' have described the total hydrolysis of urinary steroids and of the potassium salt of androst-5-en-3/3-01-17-one with enzyme preparations from the edible snail (Helix pomatia) and shown that the artefact peak was absent when steroid extracts prepared by this method were chromatographed on alumina. Stitch and Halker~ton~~ have shown by partition chromatography that there is a marked absence of artefact material with a concomitant increase of the fraction containing androst-5-en-3jI-ol-17-one and testan-3a-ol-17-one in ketosteroid extracts from urine hydrolysed by enzyme preparations of the common limpet (PateZZa vzdgatu).These workers also indicated the presence in these preparations of a sulphatase able to split sulphates conjugated through alcoholic hydroxyl groups, such as occur in the 17-ketosteroid sulphates. However, although Beher and Gaeblef19 have demonstrated that abnormal constituents of pathological urines may introduce considerable losses of 17-ket osteroids during the course28 COOK, STITCH, HALL AND FELDMAN: THE [Vol. 79 of boiling with acid, it seems probable that for some time acid hydrolysis will remain the method of choice for clinical work.We have preferred to use the slightly longer lower- temperature (about 78" to 80" C) simultaneous hydrolysis and extraction procedure of Cdlow, Callow, Emmens and Stroudg in order to reduce transformations by process (b) to a minimum, and also to decrease the quantity of non-specific chromogens present in the extract (Cooks). Dingemanse, Huis in't Veld and Hartogh-Katz2 have shown that by heating neutral urine with benzene under reflux, it is possible to obtain 3 :5-cycZoandrosten-6-01-17-0ne, an acid-labile 17-ketosteroid that is easily converted by hydrochloric acid into 3P-chloro- androst-5-en-17-one and thence to androst-5-en-3/3-01-17-0ne by heating. Mason and Engstrom@ suggested on theoretical grounds that 3 :5-cycZoandrosten-6-01-17-one may be produced from the sulphate of androst-5-en-3p-ol-17-one, and this has since been confirmed by Jones and Stitch.' It may be that 3:5-cycZoandrosten-6-ol-17-one is produced as an artefact by the Dingemanse, Huis in't Veld and Hartogh-Katz2 technique of neutral hydrolysis.The figures for urine C in Table I11 show the chromatographic patterns obtained from a urine hydrolysed (a) by the Dingemanse neutral procedure and then by Callow acid hydrolysis and (b) by Callow acid hydrolysis. The absence of the p-17-ketosteroids in fraction I11 (Pond technique) after neutral extraction is noticeable, and in such chromatograms it is probably more correct to add fractions I1 and I11 together in order to get an approxima- tion to the P-17-ketosteroid value.This figure is undoubtedly lower than the true content of the urine specimen, since artefact peaks, corresponding to the transformation products of b-17-ketosteroids, occur in the chromatograms obtained by both adsorption and partition techniques. When comparing the parallel chromatograms in Table 11, it must be remembered that 3 : 5-cycZoandrosten-6-ol-17-one does not normally appear in fraction I11 of the Jones and Stitch chromatograms, and this peak is then equivalent to fraction IV of the Pond method. There is excellent agreement between the two sets of values, except for urines A1 and A3. Fraction V of the Jones and Stitch procedure, in which is eluted testan-3a-ol-17-one and androst-5-en-3/3-01-17-one, can be compared only with the total of fractions 11, I11 and V of the Pond technique, and here the agreement is not so pronounced, especially for urine D.However, in view of the inherent difficulties of chromatography and the use of two different chromatographic principles, the general agreement is remarkably close. TABLE IV CONSTANCY OF CHROMATOGRAPHIC PATTERN OF 17-KETOSTEROID EXTRACTS OF URINE COLLECTED FROM SOME MALE SUBJECTS AT DIFFERENT TIMES Chromato- Subject hydrolysis technique Method of graphic -4 Neutral procedure of Pond Dingemanse et aZ.,2 followed by Callow s t al.9 B 97 7 9 C Callow e2 ~ 1 . ~ 93 D 3> Jones and Stitch E 3Y 39 Days 0 22 56 61 0 85 0 25 29 0 2 25 0 120 Distribution of 17-ketosteroids in fraction A I - \ I I1 111 IV v VI VII VIII 7.8 2.2 1.5 35.2 44.8 7.1 1.4 - 3.1 - 5.4 36.6 45.0 8.8 - 1.1 10.2 4.2 1.4 37.7 36.3 9.3 - 0.9 4.1 - 7.9 35.1 40.7 10.4 0.9 0.9 3.6 41.4 1.9 23.8 18.8 9-7 0.3 0.5 3.5 42.6 4.4 22.0 14.0 13.0 - 0.51 9.8 - 16.0 31.5 32.6 8.7 - 1.5 12.8 - 14.5 34.4 29.0 7.3 - 2.0 10.2 - 13.3 32.9 32.1 9.4 - 2.2 20.6 3.2 48.0 - 23.1 5.1 14.6 3.0 51.5 - 25.0 5.9 24.8 4.7 40.5 - 24.9 5.1 17.3 4.1 33.2 2.1 31-3 12.0 21.1 5.1 30.0 - 33.1 10.7 Yo Yo Yo Yo Yo Yo Yo Yo Comparatively little is known of the variations in chromatographic patterns found for normal subjects.Zygmuntowicz et aZ.3 have published data for 6 men and 4 women, and Robinson and Goulden*l have reported the percentage composition of ketosteroid excretion patterns of 19 normal men. The determination for one subject was repeated after 12 months and no change in pattern was observed.Dingemanse, Huis in't Veld and Hartogh-Katz2 have given the ranges found in repeated determinations for 13 women and 14 men. They reported their values in milligrams of the particular fraction per 24 hours and stated that theJan., 19541 FRACTIONATION OF URINARY NEUTRAL 17-KETOSTEROIDS 29 daily variations in excretion patterns for a single subject were so large that repeated deter- minations on one individual had the same significance as the same number of determinations on different subjects. Our own work shows that the individual pattern of normal and some mental patients, reported as percentage distribution of the various fractions in the 24-hour specimen, remains remarkably constant over manv weeks, as illustrated in Table IV. It is hoped 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. to report this work inlmore detail elsewhe&. REFERENCES Dingemanse, E., Huis in't Veld, L. G., and de Laat, B. M., J . Clin. Endocrinol., 1946, 6, 535. Dingemanse, E., Huis in't Veld, L. G., and Hartogh-Katz, S. L., J . Clin. Endocrinol. and Met., Zygmuntowicz, A. S., Wood, M., Christo, E., and Talbot, N. B., J . Clin. Endocrinol., 1951, 11, Pond, M. H., Lamet, 1951, ii, 90q. Wilkins, R. B., and Carlson, L. D., J . Clin. Endocrinol. and Met., 1952, 12, 647. Rubin, B. L., and Rosenkrantz, H., Fed. Proc., 1952, 11, 180. Jones, J. K. N., and Stitch, R. S., Biochem. J., 1953, 53, 679. Cook, E. R., Analyst, 1952, 77, 34.Callow, N. H., Callow, R. K., Emmens, C. W., and Stroud, S. W., J . Endocrinol., 1939, 1, 76. Zimmerman, W., Hoppe-Seyl. Z., 1935, 233, 257. Callow, N. H., Callow, R. K., and Emmens, C. W., Biochenz. J., 1938, 32, 1312. Reiss, M., Hemphill, R. E., Gordon, J. J., and Cook, E. R., Ibid., 1949, 44, 632. Grant, R. A., and Stitch, S. R., Chem. and Ind., 1951, 230. Girard, A., and Sandulesco, G., Helv. Chim. Acta, 1936, 19, 1095. Talbot, N. B., Butler, A. M., and MacLachlan, E. A,, J . Biol. Chem., 1940, 132, 595. Pincus, G., and Pearlman, W. H., Endocrinology, 1941, 29, 413. Bray, B. M., Analyst, 1952, 77, 426. Gibson, J . G., and Evelyn, K. A., J . Clin. Invest., 1938, 17, 153. Fraser, R. W., Forbes, A. P., Albright, F., Sulkowitch, H., and Reifenstein, E.C., J . Clin. Endo- Talbot, N. B., Berman, R. A., and MacLachlan, H. L., J . Bid. Chem., 1942, 143, 211. Engstrom, W. W., and Mason, H. L., Endocrinology, 1943, 33, 229. Hamburger, C., Acta Endocrinol., 1948, 1, 375. Allen, W. M., J . Clin. Endocrinol., 1950, 10, 71. Cook, E. R., and Rooks, M. E., Analyst, 1952, 77, 525. Talbot, N. B., Ryan, J., and Wolfe, J . K., J . Bid. Chem., 1943, 148, 593. Bitman, J., and Cohen, S. L., Ibid., 1949, 179, 455. Landau, R. L., Knowlton, K., Lugibihl, K., and Munson, P. L., Endocrinology, 1951, 48, 506. Butenandt, T. A., and Dannebaum, H., Hoppe-Seyl. Z., 1934, 229, 192. Butenandt, T. A., Dannebaum, H., Harrish, G., and Kundozus, H., Ibid., 1937,237, 57. Venning, E. H., Hoffman, M. M., and Browne, J. S. L., J . Biol. Chem., 1942, 146, 369. Buehler, H. J., Katzman, P. A., Doisy, P. P., and Doisy, E. A., Proc. SOC. Exp. Biol. Med., 1949, Bitman, J., and Cohen, S. L., Fed. Proc., 1950, 9, 152. Buehler, H. J., Katzman, P. A., and Doisy, E. A,, Ibid., 1950, 9, 157. Kinsella, R. A., jun., Doisy, R. J., and Glick, J. H., jun,, Ibid., 1950, 9, 150. Cohen, S. L., J . Biol. Chem., 1951, 192, 147. Henry, R., and Thevenet, M., Bull. Chem. SOC. Biol., 1952, 34, 886. Henry, R., Thevenet, M., and Jarrige, P., Ibid., 1952, 34, 897. Stitch, R. S., and Halkerston, I. D. K. H., J . Endocrinol., 1953, 9 , xxxvi. Beher, W. T., and Gaebler, 0. H., Anal. Chem., 1951, 23, 118. Mason, H. L., and Engstrom, W. W., PhysioE. Rev., 1950, 30, 321. Robinson, A. M., and Goulden, F., Brit. J . Cancer, 1949, 3, 62. 1952, 12, 66. 578. crinol., 1941, 1, 234. 72, 297. THE BIOCHEMICAL AND ENDOCRINOLOGICAL RESEARCH DEPARTMENT BRISTOL MENTAL HOSPITALS BARROW HOSPITAL BARROW GURNEY, BRISTOL J u n e 29th. 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900024

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

30 ARCHER: A SEMI-MICRO WET COMBUSTION METHOD [Vol. 79 A Semi-micro Wet Combustion Method for the Determination of Carbon BY E. E. ARCHER The method involves digesting the sample with a wet combustion mixture, aspirating the evolved gases by a slow stream of air through a heated silica tube containing a silver spiral into an evacuated Buchner flask containing barium hydroxide solution, neutralising the excess of barium hydroxide to thymolphthalein and titrating the precipitated barium carbonate directly by means of standard acid with bromophenol blue as indicator. A correction is made €or the amount of acid needed to change the acidity of the solution from that required for the colour change of thymolphthalein to that for the change of bromophenol blue. Thus the carbon dioxide evolved is directly estimated, without isolation of the precipitated barium carbonate, and errors caused by volatile mineral acids are avoided.The method is applicable to hydrocarbons , aliphatic carboxylic acids, alcohols, ketones and various sulphur and chlorine compounds. Results are quantitative €or glucose. Dry sucrose give low and irregular results, but results are quantitative with aqueous solutions of sucrose. Aqueous solutions, in general, can be successfully analysed when they contain not less than 0.5 per cent. of carbon. MANY methods have been described for the determination of carbon by wet combustion. A comprehensive review of the literature is given in a paper by Houghton.1 Weak mixtures of sulphuric acid and dichromate as used by some authors2y3 permit simple gravimetric or titrimetric estimation of the evolved carbon dioxide, but many substances are incompletely oxidised.Stronger oxidation mixtures1 ~ 4 9 ~ increase the range of substances oxidised completely, but special precautions1 or difficult techniques4~6 are necessary to avoid interference from evolved acids. In the method described a simple double titration technique avoids this interference. An evacuated flasks containing barium hydroxide is used for absorbing evolved carbon dioxide; titration is carried out in the precipitation flask, so that transference errors are avoided. In accordance with statements in the literature, it was found in this laboratory that simple mixtures of dichromate and sulphuric acid did not completely oxidise many substances, notably aliphatic acids, and although the addition of silver2 and ceric' salts effected some improvement, the range of substances completely oxidised was still limited.It was decided to use the powerful oxidation mixture devised by Van Slyke and Folch: as this completely oxidised a wide range of substances. However, volatile hydrocarbons were not completely oxidised and it was necessary to place a heated silica tube after the main reaction flask.8 To avoid bleaching of the indicators by halogens a silver spiral was placed in the silica tube.* The researches of Boivins showed that glucose, fructose, lactose and sucrose in the dry state gave low results in wet combustion methods owing to the production of carbon monoxide. Van Slyke and Folch4 claimed quantitative results for dry dextrose; the method described below also gave quantitative results for dextrose.On the other hand, results were low and irregular for dry sucrose, despite the use of the auxiliary furnace; thus it seems probable that carbon monoxide is evolved so rapidly that insufficient oxygen is present to oxidise it to carbon dioxide. Results were quantitative for sucrose when it was added as a dilute solution. METHOD APPARATUS- The apparatus is shown in Fig. 1. The combustion apparatus is a modification of that used for alkoxyl estimations by Shaw.lo The bulb on the inlet is of 10-ml capacity, and the inlet from the Winchester quart bottle packed with soda lime should be sufficiently wide to admit a I-ml pipette. The Winchester bottle packed with soda lime allows a rapid stream of air freed from carbon dioxide to be drawn through the apparatus.Jan., 19541 FOR THE DETERMINATION OF CARBON 31 The silica tube, which is supported by a sheet-iron shield, ensures that volatile substances not readily oxidised in a wet combustion process are completely converted to carbon dioxide before reaching the absorption flask.Two spirals can conveniently be mounted in one shield and, if a Y-piece is connected to the outlet of the soda-lime tower, two sets of apparatus can be used side by side. Volatile hydrocarbons gave results about 6 per cent. too low when the silica tube was omitted. For many substances its use is unnecessary, but in this work it was retained throughout. If the substance under examination contains chlorine or bromine, a few strands of silver wire are put in the bend of the silica tube to absorb Winchester bottle containing soda-lime Micro burner Mckcr burner 10 an Fig.1. Apparatus assembly for determination of carbon by the wet-combustion procedure (drawn t o scale) the halogen, which otherwise would bleach the indicators used in the titration procedure. A long jet is fitted to a 26-ml burette, so that a rubber bung that fits the Buchner flask can be slid over it. REAGENTS- Barium hydroxide solution-Dissolve 12 g of barium hydroxide in 100 ml of hot water; make up to 1 litre with cold boiled water. Keep the solution in an aspirator bottle fitted with a tap at the bottom and a soda-lime tube at the top. If the solution is left overnight, any sediment of barium carbonate will conveniently settle below the level of the tap outlet.Connect the tap to a 20-ml automatic pipette, the overflow of which is connected to a trap closed by means of a soda-lime tube. When the pipette is not in use, close its tip with a rubber cap. Combzcstion mixtzcre-Pour 85 ml of sulphuric acid, sp.gr. 1-85, and 16 ml of phosphoric acid, sp.gr. 1-75, into a 250-ml conical flask fitted with a ground-glass stopper. Add 20 g of phosphorus pentoxide, 10 g of chromium trioxide and 1 g of potassium iodate. Heat the mixture to 150" C, stir at this temperature for 2 minutes, cool and stopper the flask. Hydrochloric acid, 0.05 N-Dilute standard N acid with boiled and cooled distilled water. AZcohoZic thymolphthalein, 0.2 per cent.-Dissolve the solid in 95 per cent.ethanol. Alcoholic bromophen.ol blue, 0.1 per cent.-Dissolve the solid in 95 per cent. ethanol. PROCEDURE- This varies slightly according to the nature of the sample. Normally sufficient sample is taken to give about 6 mg of total carbon. Non-volatile solids and liquids-Add 10 ml of combustion mixture to the combustion fiask and assemble the apparatus as in Fig. 1, lubricating the ground-glass joints with a little phosphoric acid. Remove the Buchner flask, evacuate the apparatus by means of a water pump, close the tap, and re-assemble the apparatus. Open the tap to aspirate carbon32 ARCHER: A SEMI-MICRO WET COMBUSTION METHOD [Vol. 79 dioxide-free air through the apparatus. Repeat the operation to eliminate all carbon dioxide from the system and to fill the Buchner flask with carbon dioxide-free air.From the automatic pipette add 20 ml of barium hydroxide solution to the Buchner flask, replace the bung, and again evacuate. Weigh the sample in a weighing spoon that can be lowered through the condenser attached to the combustion flask. Break the joint leading to the silica tube, carefully lower the weighing spoon into the combustion flask, and replace the silica tube. Open the tap on the Buchner flask carefully so that a slow stream of air is aspirated through the apparatus. Heat the combustion flask fairly vigorously with a micro-burner, but avoid heating enough to cause heavy white fumes to pass the condenser. After 10 minutes remove the flame and increase the air flow so that in about another 5 minutes the pressure inside the Buchner flask is again atmospheric. After 15 minutes, with occasional swirling, the flask is ready for titration.Take 10 ml of the combustion mixture through the procedure exactly as above. Remove the bung from the Buchner flask, add a few drops of 0.2 per cent. alcoholic thymolphthalein and close the flask with the bung attached to the 25-ml burette containing 0.05 N hydrochloric acid. Titrate the solution to the thymolphthalein end-point, remove the burette, and add 1 ml of 0.1 per cent. bromophenol blue. Titrate until a reddish-yellow colour is reached. The difference between the two titrations is referred to as the “indicator blank.” In practice this is about 0.6 ml of 0.05 N hydrochloric acid, or slightly more than would be needed to take 20 ml of barium hydroxide solution directly delivered from a pipette through the same indicator changes.This difference is thought to be caused by traces of organic matter in the combustion mixture-possibly from the paraffin wax used to seal bottles of phosphorus pentoxide. Titrate the sample in exactly the same way as the “indicator blank,” until the barium carbonate has completely dissolved. Near the end-point replace the rubber bung in the flask, and, closing the tap, shake the flask well to dissolve any carbonate sticking to the sides. When the end-point is near, allow the flask to stand for 5 minutes before making the final adjustment. The end-point is reached when the sample solution and “indicator blank” exactly match. The match is best made by comparing the colours of the two solutions, viewed horizontally, against a white background.It is convenient to refill the burette after reaching the thymolphthalein end-point, as the value of the first titration is disregarded. For each batch of combustion mixture an “indicator blank” must be prepared. Then (titre - “indicator blank”) x 6 x 100 20,000 x weight of sample taken. the percentage of carbon in sample = Volatile liquids-The sample is weighed in a piece of capillary tubing, about 2$ inches long, drawn out to a very fine point at both ends. B.D.H. capillator tubing is ideal for the purpose. Weigh the empty tube. Manipulate the sample by means of a capillator bulb so that it occupies the middle section of the tube, leaving none at either end. Remove the capillator bulb, taking care not to break the fine tip and, after wiping the tube with filter-paper, re-weigh the tube.Temporarily break the connection to the soda-lime tower, drop the tube down the side-arm of the combustion flask and completely assemble the apparatus as before. Aspirate a slow stream of air through the apparatus, and at first heat fairly strongly. As the air in the capillary warms up, the sample liquid is forced up the tube until it reaches the tip. At this point increase the air stream slightly to prevent the carbon dioxide, which is rapidly evolved, from sucking back. When the reaction has subsided somewhat, close the tap temporarily so that the combustiori mixture is sucked back into the side-arm and the capillary is completely immersed in the hot combustion mixture.This ensures complete reaction of any less volatile constituents. The rest of the procedure is exactly as described for non-volatile solids and liquids. Aqueous solutions-Temporarily break the connection to the soda-lime tower, and run 1 ml of the sample from a pipette down the side-arm. Then proceed as before. It was found that with solutions of formic acid the reaction was extremely rapid, leading to losses of carbon dioxide whilst the sample was being added. This was avoided by first running a few drops of water down the side-arm to form a “buffer layer” between the combus- tion mixture and sample solution. It is then vaporised by the incoming air stream.Jan., 19541 FOR THE DETERMINATION OF CARBON 33 EXPERIMENTAL To test the absorption of carbon dioxide under the conditions described, a solution of pure sodium carbonate was used.The apparatus was set up as usual except that no com- bustion mixture was added to the flask. After the apparatus had been swept free from carbon dioxide, 10 ml of the solution were placed in the flask via the condenser by means of a pipette. The condenser was reconnected and 1 ml of water and then 1 ml of 5 N hydrochloric acid were added through the side-arm by means of a pipette. The carbon dioxide scrubber was reconnected, and air was aspirated through the system exactly as described for a sample, the mixture being gently boiled. In these experiments 0-4373g of sodium carbonate was made up to 100m1, so that each 10-ml aliquot contained 0.04373 g of sodium carbonate.On calculating the amount of absorbed carbon dioxide back to its equivalent weight of sodium carbonate in three experiments, the amounts of sodium carbonate recovered were 0.04380, 0.04377 and 0.04370 g. Hence, within normal experimental limits, the recovery of carbon dioxide is 100 per cent. RESULTS Table I shows the range of substances examined and the accuracy of the results obtained. TABLE I THE DETERMINATION OF CARBON BY THE PROPOSED METHOD Hychocavbons- n-Heptane . . 3-Methylpentane Naphthalene Sample .. .. .. . . .. .. . . .. Carbon 7 A \ found, theory, Yo % 83.7, 83.2 83.1, 83.1, 82.8 92.9, 93.6, 93.8 84.0 83-7 93.8 Carboxylic Acids- Acetic acid (by direct weighing) . . . . .. 40.5, 39.9, 40.2 40.0 Formic acid (dilute aqueous solution) .. .. 26.0, 25.9, 25-9 26.1 Butyric acid (dilute aqueous solution) . . . . 54-6, 54.6 54.6 Ca ybohydyates- Glucose . . .. .. .. .. . . 39.8, 39.8 Sucrose (weighed dry) . . . . .. .. 34.6, 34.4 Sucrose (aqueous solution) . . .. . . 42.0, 42.1 Ethanol (dilute aqueous solutiorl) . . .. . . 51.9, 51.9 Acetone (dilute aqueous solution). . .. . . 62.5, 61.8 Sulphanilic acid . . .. . . .. . . 41.4, 41.2 Polyvinyl chloride . . .. . . .. .. 38.4 act-Dichloropropionitrile (experimental sample) . . 294, 28.9 Vayious- 40.0 42.0 42-0 52-2 62- 1 41.6 38-4 29.0 The author wishes to thank Mrs. E. Jacquet, who assisted in working out the method, and the Directors of The Distillers Company Limited for permission to publish this article. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Houghton, A. A., Analyst, 1945, 70, 118. Nicloux, M., Bull. SOC. Chim. B i d , 1927, 9, 639. Messinger, J., Ber., 1890, 23, 2756. Van Slyke, D. D., and Folch, J., J . Biol. Chem., 1940, 136, 509. Farrington, P. S., Niemann, C., and Swift, E. H., Anal. Chem., 1949, 21, 1423. Ingleson, H., and Bentley, J. A., Analyst, 1946, 71, 328. Krough, A., Biochem. Z . , 1930, 221, 247. Cornfield, A. H., J . Sci. Food Agric., 1952, 3, 154. Boivin, A., Bull. SOC. Chim. Biol., 1929, 11, 1269. Shaw, J., J . SOC. Chem. Ind., 1947, 66, 147. THE DISTILLERS COMPANY LIMITED RESEARCH AND DEVELOPMENT DEPARTMENT GREAT BURGH, EPSOM, SURREY June 23rd, 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900030

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

34 WAGER: AN IMPROVED COPPER REDUCTION METHOD FOR [Vol. 79 An Improved Copper Reduction Method for the Micro-determination of Reducing Sugars BY HAROLD G. WAGER Errors in the method for the micro-determination of reducing sugars proposed by Nelson are shown to derive from (2) oxidation of the copper reagent by dissolved oxygen during heating to give reducing compounds that interfere with the determination, (ii) re-oxidation of the cuprous ions formed and (iii) instability of the blue colour formed by the arsenomolybdate reagent. These can be avoided by carrying out the whole test under oxygen-free conditions and recording the extinction coefficient at a standard time after adding the arsenomolybdate reagent. THE use of paper chromatography for the qualitative separation of sugars in extracts of plant tissues or plant products is now well established, but for the quantitative determination of the sugars so separated a good and convenient micro-method is required.If, as not infrequently occurs, one of the sugars is present in small amounts relative to the other sugars, the determination must be reliable over a wide range of concentration. The estimation could be carried out by a micro technique of the Linderstrom-Lag1 type, but a less specialised method was required. The colorimetric method proposed by Nelson2 seemed suitable for the quantitative determination of sugars eluted from paper chromatograms, but it was found to be subject to serious errors, especially at low concentra- tions of sugar. The causes of some of these errors are shown below, and an improved procedure is suggested by which 100 pg of reducing sugar can be estimated to -+_ 1 per cent.and 10 pg to k3 per cent. The modification of the Shaffer - Hartmann3 copper reduction method proposed by Somogyi4 in 1937 was adapted for use as a micro-colorimetric method by Nelson.2 The cuprous ions formed by the reaction with sugars were allowed to react with an arseno- molybdate reagent and the blue colour so formed was measured photometrically. In 1945, Somogyi6 proposed another copper reagent and recommended it for use with the arseno- molybdate reagent in the micro-method. A further paper by SomogyP has appeared since the work described below was done; in it he criticises the copper reagent proposed in his 1945 paper on the grounds that it interferes with the stability of the colour developed by the arsenomolybdate reagent, and he proposes in its place a reagent, buffered by sodium carbonate and bicarbonate, which therefore is of similar type to the Maskell and Narain copper reagents used in the present work.He further states that “tartrate measurably reduces copper, especially when the reagent is heated,” but apart from this no reference to the effect described below is made. EXPERIMENTAL Two copper reagents have been used- (1) The reagent described by S~mogyi.~ With this reagent the water blank value slowly increased with time. (2) The modified Shaffer - Hartmann copper reagent developed by Maskell and Narain* with the potassium iodate and oxalate omitted. These compounds are only required for the iodometric estimation of the cuprous ion.This reagent was found to be stable. * Professor E. J. Maskell has kindly contributed the following brief description of his method. “The modified Shaffer - Hartmann technique developed by Maskell and Narain in 1932 and described in R. Narain’s Ph.D. thesis, Cambridge, 1932, is unpublished except for a brief statement by Gawadi.’ The essential features of the modified technique are (1) the use of a copper reagent of lower alkalinity] which improved linearity and reduced the sensitivity to non-sugars, and (2) the mixture of copper reagent and sugar solution is, immediately before being boiled, swept out for 10 minutes with a stream of nitrogen gas, thereby greatly reducing reoxidation by the reduced copper oxide during boiling.The copper reagent consists of 0.02 M copper sulphate, 0.1 M sodium carbonate, 0.2 M sodium bicarbonate, 0.06 M sodium potassium tartrate, 0.1 M potassium oxalate and 0.0033 M potassium iodate.”Jan., 19541 THE MICRO-DETERMINATION OF REDUCING SUGARS 35 The arsenomolybdate reagent described by Nelson2 was used for the development of the molybdate blue colour. The tests were made with 2 ml of the copper reagent and 5 ml of sugar solution in a 6 x &inch test tube with a 25-ml calibration mark. The tube with the mixed solution was plunged into a bath of boiling water for 10 minutes, and then cooled in running cold water, 1 ml of arsenomolybdate reagent was added, the volume was made up to 25ml and the extinction coefficient of the molybdate blue was determined, a deep red filter such as Ilford No.608 being used. RESULTS- Considerable variations in the extinction coefficient of both the water blanks and tests with 100 pg of glucose were found in tests with the same solution of the Somogyi copper reagent on successive days, and results are shown in Table I. The variation was approxi- mately constant at all the concentrations of glucose tested and therefore the percentage errors in determinations of 5 to 10 pg of glucose were very large. TABLE I EXTINCTION COEFFICIENT OF THE WATER BLANK AND OF 100 pg OF GLUCOSE: SOMOGYI COPPER REAGENT Average increase for 100 pg of Date Water blank 100 pg of gIucose glucose 8.12.48 0.106, 0.108 0.562, 0.662 0.445 10.12.48 0.092, 0-098 0.562, 0.662 0.467 10.12.48 0.110, 0.108 0.633, 0,645 0.435 13.12.48 0.127, 0.116 Possible causes of these variations seemed to be- (i) The formation of reducing compounds by oxidation of the copper reagent.(ii) The re-oxidation at some stage of the cuprous ions formed during the oxidation To test the first possibility, water blank tubes were prepared with the Somogyi copper reagent, and oxygen, air or nitrogen was bubbled through the liquid for 10 minutes before heating. The results of this experiment are shown in Table 11. The values of the extinction coefficient of the water blank are clearly dependent on the amount of oxygen in the liquid. Heating the reagent in the presence of oxygen, therefore, appears to lead to the production of reducing substances that reduce either the cupric ion or the arsenomolybdate complex.of the sugar. TABLE I1 EXTINCTION COEFFICIENT OF THE WATER BLANK AFTER EQUILIBRATION WITH NITROGEN, AIR OR OXYGEN: SOMOGYI COPPER REAGENT Gas used for equilibration Extinction coefficient Nitrogen . . .. . . . . . . 0.044, 0-042 Air . . .. .. .. .. . . 0.128, 0.127 Oxygen . . * . .. .. .. . , 0-305, 0.308 Lower and less variable values for the water blank and for tests with glucose might therefore be expected if oxygen was removed from the test liquids by a stream of nitrogen gas before heating (this technique was used by Maskell and Narain, cf. Gawadi7 and footnote, p. 34). Many tests of the effectiveness of oxygen removal by a stream of nitrogen gas were made with the Somogyi copper reagent; the extinction coefficient of the water blank was always low but was still variable; most values fell between 0.02 and 0.03, but values as high as 0.057 were recorded.In consequence of this variability, the possibility of the re-oxidation of the cuprous ion during cooling or during the addition of the arsenomolybdate reagent was investigated. The mixture of copper reagent and water or sugar solution was freed from oxygen by a stream of nitrogen and was maintained under nitrogen both during the heating and during the cooling, and nitrogen was blown into the top of the test tube while the arsenomolybdate reagent was added. The effect of this treatment was to increase the value of the extinction coefficient36 WAGER: AN IMPROVED COPPER REDUCTION METHOD FOR [Vol. 79 of the water blank to about 0.06.Re-oxidation of cuprous ions must, therefore, have been occurring in the previous tests. This re-oxidation could be demonstrated by admitting a little air to the tubes before adding the arsenomolybdate reagent, when the value of the extinction coefficient fell again to 0.02. The reliability of the method with both reagents and with the modified technique just described was studied and results for the Maskell and Narain copper reagent are shown in Table 111. This table shows that the estimations made in 1949 agree very closely with those made in 1951, that 100 pg of glucose can be estimated to within 2 1 per cent., and that a control sugar sample in each run is unnecessary; such controls were required in the Nelson technique. A similar series of observations, which are not quoted, made with the Somogyi copper reagent showed that the extinction coefficient of the water blank increased slightly on successive days and for some unknown reason the scatter of values was greater; in 14 tests the range of the increase in the extinction coefficient of 100 pg of glucose over that of the water blank was 0.479 to 0.522.TABLE I11 EXTINCTION COEFFICIENT OF WATER BLANK AND OF 100 pg OF GLUCOSE: MASKELL Each figure is the average of two determinations in one experiment AND NARAIN COPPER REAGENT, UNDER NITROGEN THROUGHOUT Date 20.1.49 21.1.49 21.1.49 25.1.49 26.1.49 31.1.49 1.2.49 Water blank 0.073 0.065 0.065 0.052 0.061 0.060 0.050 Increase over water blank for 100 pg of glucose 0.59 1 0.580 0.585 0.586 0.579 0.592 0.592 Date 13.4.51 18.4.51 23.4.51 23.4.51 26.4.51 26.4.51 1.5.51 Water blank 0.051 0.041 0.056 0.060 0-064 0.065 04356 Increase over water blank for 100 pg of glucose 0.59 1 0.584 0.584 0.582 0.591 0.594 0.587 With the proposed method the extinction coefficient is directly proportional to the concentration of glucose over the range of 0 to lOOpg, as shown in Table IV.This table also shows that a reasonably reliable estimate of as little as 10 pg of glucose is possible with the modified technique. TABLE IV EXTINCTION COEFFICIENT FOR VARIOUS CONCENTRATIONS OF GLUCOSE : MASKELL AND NARAIN COPPER REAGENT, UNDER NITROGEN THROUGHOUT Each value is the average of two determinations in one experiment Glucose, Pg 10 20 100 Increase of the extinctions coefficient over the water blank for glucose 0.059, 0.059 0.119, 0.119 0.591, 0.594 An estimate of the reducing bodies formed, and of the extent of re-oxidation of the cuprous ion, can be made by comparing the extinction coefficients of solutions heated in equilibrium with either air or nitrogen and then cooled either under air or nitrogen (see Table V). The high value of the water blank of the Somogyi reagent when heated with air is, presumably, caused by oxidation of the reagent, and as there is only a slight increase in value when it is cooled under nitrogen, only a small amount of re-oxidation can be occurring as a result of the reduced solubility of oxygen caused by the presence of sodium sulphate.In the Maskell and Narain reagent, on the other hand, the presence of oxygen at any stage leads to much oxidation of the cuprous ion, as instanced by the low water-blank values when air is present at any stage.Further, the higher value given by 100 pg of glucose with the nitrogen - air treatment than with the air - nitrogen treatment shows that this re-oxidation of the cuprous ion is greater during the heating than the cooling phase. The productionJan., 19541 THE MICRO-DETERMINATION OF REDUCING SUGARS 37 of reducing bodies by direct oxidation of the reagent seems to be less in the Maskell and Narain reagent than in the Somogyi reagent ; this may be related to the different pH values of the two reagents. TABLE V EXTINCTION COEFFICIENTS OF WATER BLANKS AND 1OOpg OF GLUCOSE Each value is the mean of two determinations in one experiment UNDER DIFFERENT CONDITIONS OF AERATION Gas under which solution was & heated cooled Air Air Air Nitrogen Nitrogen Nitrogen Nitrogen Air Maskell and Narain copper reagent r A 1 Increase for 0.013, 0.112 0.444, 0.449 0.026, 0.024 0.452, 0.443 0.024, 0.048 0-561, 0.527 0-056, 0.060 0-584, 0.582 Water blank 100 pg of glucose Somogyi copper reagent r A 1 Increase for 0.121, 0.126 0.439, 0.434 0-143, 0.131 0-443, 0.476 0.044, 0.034 0.476, 0.471 0.067, 0.070 0.488, 0.502 Water blank 100 pg of glucose The blue colour produced by the arsenomolybdate reagent increases with time, as shown This rise in intensity of colour is undoubtedly one of the major sources of error in Fig.1. Time, hours Fig. 1. Change of extinction coefficient of molybdate blue with time, measured through a deep-red filter.The arsenomolybdate reagent was added at zero time. Curve M, 100 pg of glucose tested with the Maskell and Narain copper reagent; curve B, water blank value of the Maskell and Narain copper reagent; curve S, 100 pg of glucose tested with the Somogyi copper reagent remaining in the method. Modifications to the conditions of colour development were tried, but no increase in stability resulted. The use of a phosphomolybdate reagent was found to give a stable colour, but with the various conditions tested, an intensity of colour, resulting from added glucose, greater than about one-third of that produced by the arsenomolybdate reagent could not be obtained, and this led to a serious decrease of sensitivity. It was finally decided to determine the extinction coefficient at an exactly standard time after adding the arsenomolybdate reagent.This slowed the determinations, but the accuracy was considerably increased. The results in Tables 111, IV and V were recorded with this technique. PROPOSED METHOD APPARATUS- A series of units, only one of which is shown complete, are linked together by two systems of T-pieces. A convenient apparatus for carrying out the method is shown in Fig. 2.38 WAGER [Vol. 79 The test tubes stand in a rack that can be put successively into boiling water and into cold running water without stopping the flow of nitrogen. PROCEDURE- Place 6ml of sugar solution and 2ml of copper reagent by pipette into each of the graduated test tubes and fix them to the rubber bungs. Pass nitrogen into the liquid for about 10 minutes through the T-pieces, D, the capillary resistances, E, inserted to equalise the flow in all the test tubes, and the tube, F, of narrow bore to ensure a stream of fine bubbles. G Fig.2. Apparatus for heating and cooling under nitrogen The nitrogen escapes through the second system of T-pieces, C. After the liquids have been freed from oxygen, turn the two-way tap, G, to allow the stream of nitrogen to flow over the tops of the tubes. Put the rack with the tubes into boiling water for exactly 10 minutes, and then transfer it to cold water for 3 minutes. The immediate contraction in volume occurring on plunging the tubes into cold water is made good by nitrogen from the vessel, A. Then shut the tap, B, when nitrogen escapes through the pressure regulator and overflow, H.Lift each bung in turn just sufficiently to insert a pipette and add 1 ml of the arsenomolybdate solution; replace the bung and shake the tube. During this addition the whole stream of nitrogen escapes from the test tube, so effectually preventing entry of oxygen. Add the arsenomolybdate reagent to successive tubes at exactly 2-minute intervals and measure the extinction coefficient exactly 24 minutes later. These times are arbitrary, but they have been found convenient for eight or ten tubes and must be kept standard, as the duration of standing affects the calibration. Once the arsenomolybdate has been added, oxygen has no further effect and the liquids are made up to volume in air. I have used the Maskell and Narain technique for the semi-micro estimation of sugars for many years, and I should like to thank Professor E. J. Maskell for having given me particulars of it. The work described in this paper forms part of the programme of the Food Investigation Organisation of the Department of Scientific and Industrial Research. Mr. J. Howe carried out the experimental work. REFERENCES 1. 2. 3. 4. 5. - , Ibid., 1945, 160, 62. 6. - , Ibid., 1962, 195, 19. 7. Linderstrom-Lag, K., and Holter, H., C.R. Lab. Carlsberg, 1933, 19, No. 14. Nelson, N., J . Bid. Chem., 1944, 153, 376. Shaffer, P. A., and Hartmann, A. F., Ibid.. 1920, 45, 365. Somogyi, M., Ibid., 1937, 117, 771. Gawadi, A. G., Plant Physiol., 1947, 22, 438. Low TEMPERATURE STATION FOR RESEARCH IN BIOCHEMISTRY AND BIOPHYSICS and UNIVERSITY OF CAMBRIDGE DEPARTMENT OF SCIENTIFIC AKD INDUSTRIAL RESEARCH June 22n8, 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900034

出版商:RSC    

年代:1954

数据来源: RSC

摘要:

Jan., 19541 VENTURA AND WHITE 39 The Determination of Iron and Copper in Single Serum Samples BY S. VENTURA* AND J. C. WHITE A method is described for the spectrophotometric determination of iron and copper in single samples of serum. The metallic ions are liberated from protein with 6 N hydrochloric acid and determined as the 2-2’-dipyridyl- ferrous complex and as copper diethyldithiocarbamate, the latter being extracted into a mixture of ether and amyl alcohol. IN the study of iron and copper metabolism, it is convenient to measure the amount of metals in single samples of serum. Satisfactory splitting of the metallic ions from protein is an essential prelude to their spectrophotometric or colorimetric determination as coloured complexes. The ferrous-ion complex with 2-2’-dipyridyl and copper diethyldithiocarbamate prove most suitable, as they can be separated readily.SEPARATION OF IRON AND COPPER IONS FROM SERUM PROTEINS- Both metals are bound to protein in the serum. The iron-binding protein or siderophyllin has been identified as a &-globulin in the IV-7 fraction; but although this fraction binds both iron and copper in vztro, in vivo it appears to be concerned only with binding of iron (Surgenor, Koechlin and Strong1). Dejardin and LambrechW used sodium citrate, Kitzes, Elvehjem and Schuette3 and Cartwright, Jones and Wintrobe4 used warm trichloroacetic acid and Jones5 nitric acid. Hydrochloric acid at different strengths and temperatures is frequently used (Barkan and Walcher,6 Agner7 and Von Porat.8 Ventura and King9 have shown that simple 6 N hydrochloric acid extraction of serum, as used by Heilmeyer and PlotnerlO for the determination of serum iron, gives maximal reproducible values for both iron and copper (Tables I and 11).TABLE I EFFECT OF CONCENTRATION OF HYDROCHLORIC ACID ON SPLITTING OF IROX Each value is the mean of five determinations EXPERIMENTAL A variety of methods have been used for freeing the metallic ions. AND COPPER FROM SERUM PROTEINS Concentration of hydrochloric acid, N Iron released after a reaction time of 5 minutes, 10 minutes, 15 minutes, 20 minutes, A 7 7 pg per 100 ml pg per 100 ml pg per 100 ml pg per 100 ml 2 51 59 63 64 4 89 97 104 107 6 106 117 117 118 8 97 98 100 100 Copper released after a reaction time of 10 minutes pg per 100 ml 101 113 124 114 TABLE I1 DIFFERENCES IN AMOUNTS OF IRON AND COPPER SPLIT FROM SERUM BY VARIOUS REAGENTS AND 6 N HYDROCHLORIC ACID ACTING FOR 10 MINUTES Means of five samples Yo 6 N Hydrochloric acid .. . . .. I00 Nitric acid .. .. .. .. 91 Sodium citrate . . .. . . .. 85 Trichloroacetic acid? . . .. .. 104 Reagent Iron, (86-95) (75-93) (96-110) Copper, Yo 100 92 (88-98) 85 (76-91) 101 (98-1 03) 7 By heating the serum to 90” C and then treating it with trichloroacetic acid by the method of Kitzes et aZ.9 for 3 minutes. * Present address : Clinica Medica, Pavia, Italy.40 VENTURA AND WHITE: THE DETERMINATION OF [Vol. 79 PRECIPITATION AND SEPARATION OF THE PROTEINS- Deproteinisation after liberation of the metallic ions is commonly carried out with 20 per cent. trichloroacetic acid.Separation of the protein precipitate can be effected by centrifugation or filtration. The former method does not always yield an optically clear supernatant liquid, and for filtration, papers washed in iron-free N hydrochloric acid must be used. Five successive determinations on the same serum sample by both methods gave average values of 118 pg of iron per 100 ml (standard deviation 1-48; coefficient of variation 1-25 per cent.) by filtration and 119 pg per ml (standard deviation 1.84; coefficient of variation 165 per cent.) by centrifugation. FORMATION OF THE COLOURED IRON COMPLEX- The iron in the protein-free solution may be determined as ferric thiocyanate or as the stable complexes of ferrous iron with 2-2'-dipyridyl or o-phenanthroline (Table 111).Here, where copper is determined simultaneously, the 2-2'-dipyridyl reaction has proved to be the best method. TABLE 111 RECOVERY OF IRON ADDED TO SERUM Method Z-Z'-Dipyridyl o-Phenanthroline Thioc y anate A I - - - % z z t P 0 Iron added No. of of of of to sample, samples Recovery, variation, Recovery, variation, Recovery, variation, % 2.68 10 to 60 9 85.9 76to 160 9 96.6 2-59 98.9 1.08 98.9 1-26 176 to 300 9 99.4 1.01 99.7 0.84 99.6 2.02 Yo 974 % % 92.6 3.78 % 6.32 pg per 100 ml % With a solution containing 300 pg per 100 ml of iron", we have found that for complete reduction to ironn, 2 per cent. hydroquinone in 0-1 per cent. aqueous ascorbic acid solution is more reliable than sodium dithionite (Na,S,O,) (HilPl), hydrazine sulphate (Barkan and Walchers) , hydroxylamine hydrochloride ( Jones6) , sodium pyrosulphite (Agner7 and Von Porats) , or hydroquinone alone (Heilmeyer and PlotnerlO).The coloured complex with 2-2'-dipyridyl forms between wide pH limits (pH 3 to 9; Snell and Snell12), but the intensity and rate of formation vary. Standardisation is effected by buffering to a pH of 4.4. FORMATION OF THE COLOURED COPPER COMPLEX- The yellow-green of the complex of copper with sodium diethyldithiocarbamate does not vary greatly in intensity between pH values of 4.4 and 10.0. At pH 4.4 interference from ferrous ions is eliminated if these are fixed as the stable complex with 2-2'-dipyridyl. TABLE IV MEAN DENSITIES AND STANDARD DEVIATIONS OF THE COPPER - SODIUM SOLVENTS, EACH CONTAINING 365pg OF COPPER PER 100ml Ten determinations at a pH value of 4.4 DIETHYLDITHIOCARBAMATE COMPLEX BY A SINGLE EXTRACTION FROM VARIOUS Mixture of Mixture of Mixture of ether and ether and ether and amyl alcohol amyl alcohol amyl alcohol Solvent Ether Amyl alcohol (2 + 1) (2 + 6) (1 + 1) Mean density .. 0,286 0.247 0.276 0.270 0.268 Standard deviation . . 0.00205 0.00166 0.00268 0.00139 0.000641 The copper complex can be extracted from the mixture with solvents of low polarity. Amyl alcohol (Parker and Griffinla) and isoamyl alcohol cannot be used for the extraction at a pH of 4.4 as the iron complex is also partly removed. Ether extracts the copper complexJan., 19541 IRON AND COPPER IN SINGLE SERUM SAMPLES 41 selectively, but several extractions are required, and, if a high concentration is present, some may still remain in the aqueous phase.A mixture of equal volumes of ether and amyl alcohol removes the greater part of the copper complex in a single extraction. The optical densities of the resulting extracts are reproducible under standard conditions (Table IV) . None of the iron complex is removed and the intensity is not affected. With the o-phenanthroline - iron complex, however, some fading occurs. The formation of the respective complexes of iron and copper proceeds independently, and the molar extinction coefficients and absorption curves in the separated fractions are the same as when determined independently. Recovery of added copper and iron from the mixtures is shown in Tables 111 and V. TABLE V RECOVERY OF ADDED IRON AND COPPER FROM SERUM BY COMBINED DETERMINATION Metal added to sample, Recovery of iron, Recovery of copper, pg per 100 ml % % 50 96.9-1 0 1 * 0 934-101.0 (6 samples) (6 samples) 100 96.7-101.0 97.3-98.6 Mean 98-4 98.3 STOCK STANDARD SOLUTIONS- Iron-Dissolve 0.863 g of ferric ammonium sulphate and 5 ml of concentrated sulphuric Coj5j5er-Dissolve 0.1 179 g of AnalaR copper sulphate, CuS04.5H,0, in distilled water acid in distilled water and make up to 1 litre.Each millilitre contains 100 pg of iron. and make up to 1 litre. Each millilitre contains 30 pg of copper. METHOD PKOCEDURE- Mix 4 ml of serum with 2 nil of 6 N hydrochloric acid and set the mixture aside for 10 minutes. Then add 4 ml of 20 per cent. trichloroacetic acid, leave for another 10 minutes and filter. Add to 5 ml of filtrate in a separating funnel, 1 drop of p-nitrophenol solution.Carefully add concentrated ammonium hydroxide solution dropwise until the colour is yellow-green. -4djust the pH value by back titration with 0.1 N hydrochloric acid until colourless, add 0-6 ml of sodium acetate and acetic acid buffer solution (pH 4-4) and then 0-5 ml of 2 per cent. hydroquinone in 0.1 per cent. ascorbic solution. After shaking, add 5 drops of 1 per cent. 2-2’-dipyridyl solution, when the red colour develops at once. Five minutes later, add 1 ml of 0.1 per cent. sodium diethyldithiocarbamate solution and 5ml of a (1 + 1) mixture of ether and amyl alcohol. Extract the copper complex by shaking for 1 minute. Separate the phases. If the mixture of ether and amyl alcohol is deeply coloured, extract a second time.Measure the optical density of the copper complex a t 440mp on a spectrophotometer. Make the aqueous layer containing the red iron complex up to 10 ml and determine its optical density at 520 mp on the spectrophotometer. Simultaneously put a blank solution and 4 rnl of a standard containing 1-5 pg of iron per ml and 1.5 pg of copper per ml through the same procedure. APPLICATION The method described has the advantage of requiring only a small serum sample, but allows accurate separation and determination of the iron and copper contents. 2-2’-Dipyridyl for determination of iron has been suggested frequently (Hil1,ll Allport14 and Snell and SnelP2). It is not the most sensitive reagent for iron (Woods and Mellon15 and SandelP), but the complex is stable and is not extracted from the aqueous phase under the conditions used.We have found a molecular extinction coefficient (E) of 8950 for the coloured cation at 520mp and at a pH of 4.4, after 5 minutes. The complex contains 3 molecules of 2-2’- dipyridyl co-ordinated with 1 atom of ferrous iron (Feigll’). The figures given by Heilmeyer and Plotnerlo indicated that E was about 8630. There is close adherence to Beer’s law,42 VENTURA AND WHITE [Vol. 79 even with concentrations of less than 100 pg of iron per 100 ml of serum, which give optical densities of less than 0.032 in the final five-fold diluted coloured solution. With the copper complex, Cu(S.CSN.C,H,),, there is similarly a regular relation between concentration and density. The method has been used mainly with a Beckman spectrophotometer, at a slit-width of 0.4 mm.With the low densities frequently encountered in pathological sera, accuracy can be maintained best by use of this, or a similar sensitive spectrophotometer. A photo- electric absorptiometer is not so well suited to this purpose, but we have obtained good results with a Hilger “Biochem” absorptiometer and a Chance OGRI filter and also with a Hilger Spekker absorptiometer, a 4-cm cell for higher densities and an Ilford No. 604 spectrum green filter (transmission 10 per cent. at 520 mp). With the second of these instruments 30ml of coloured solution are required and the serum sample must be 12 ml in volume. Attention to the cleanliness of the optical cells is essential, so they must be washed several times in water doubly distilled from glass apparatus, and then drained before use.Representative results are shown in Table VI (see also Ventura and Whitelg). TABLE VI REPRESENTATIVE VALUES FOR IRON AND COPPER DETERMINED IN SINGLE SERUM SAMPLES IN HEALTH AND VARIOUS ANAEMIAS Group No. of cases Iron, Copper, pg per 100 ml standard deviation, 7.8) pg per 100 ml standard deviation, 5.5) Normal males . . .. .. 10 115-141 109-124 (Mean, 125; (Mean, 116; Iron-deficiency anaemia . . .. 8 29-85 Haemolytic anaemia . . .. 4 69-175 Other blood disorders . . .. 13 32-190 Pernicious anaemia. . .. .. 2 130-191 91-278 44-174 174-181 134-237 We are grateful to Professor E. J. King for his interest and help in this work, and to Dr. G. H. Beaven for criticking the manuscript.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. REFERENCES Surgenor, D. M., Koechlin, B. A., and Strong, L. E., J . Clin. Invest., 1949, 28, 73. Dejardin, J., and Lambrechts, A., Acta Biol. Belg., 1943, 14, 208. Kitzes, G., Elvehjem, C. A., and Schuette, H. A., J.. Biol. Chem., 1944, 155, 653. Cartwright, E., Jones, P. J., and Wintrobe, M. M., Ibid., 1945, 160, 593. Jones, F., Anal. Chem., 1949, 21, 1216. Barkan, G., and Walcher, G., J . Biol. Chem., 1940, 135, 37. Agner, K., quoted by Von Porat, B. T. D., Scand. J . Clin. Lab. Invest., 1950, 2, 106. Von Porat, B. T. D., Ibid., 1950, 2, 106. Ventura, S., and King, E. J., Biochem. J., 1951, 48, lxi. Heilmeyer, L., and Plotner, K., “Das Serumeisen und die Eisenmangelkrankheit,” Fisher, Jena, 1937. Hill, R., Proc. Boy. Soc., B, 1930, 107, 205. Snell, F. D., and Snell, C. T., “Colorimetric Methods of Analysis,” Volume 11, Van Nostrand & Parker, W. E., and Griffin, F. P.. Canad. J . Res., B, 1939, 17, 66. Allport, N. L., “Colorimetric Analysis,” Chapman & Hall Ltd., London, 1945, p. 53. Woods, J. T., and Mellon, M. G., Ind. Eng. Chem., Anal. Ed., 1941, 13, 554. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Second Edition, Interscience Feigl, F., translated by Matthews, J. W., “Qualitative Analysis by Spot Tests,” Nordemann Ventura, S., and White, J. C., Biochem. J., 1951, 48, lv. Co. Inc., New York, 1944, p. 310. Publishers Inc., New York, 1950, Chap. 23. Publishing Co., Amsterdam and New York, 1939, p. 93. POSTGRADUATE MEDICAL SCHOOL LONDON, W.12 DUCANE ROAD June 8th, 1953

ISSN:0003-2654    

DOI:10.1039/AN9547900039

出版商:RSC    

年代:1954

数据来源: RSC