THERMIC REACTIONS IN VACUO.* NEWS 3C when ignited with a fuse commences to react and By FRANK E. WESTON, B.Sc., and H. RUSSELL ELLIS, B.Sc. proceeds throughout the mass, when the reaction is carried out in a Hessian crucible in the air. PART I. IN a paper read before the International Congress of Applied Chemistry, in June, 1909, the authors gave a synopsis of a number of reactions similar in nature to the well-known "Thermite" reaction, and also an account of experiments carried out with these mixtures in vacuo. As the paper appears only in abstract in the Proceedings of the Congress, and as the work has been further extended Watts and Breckenridge (Trans. Am. Elect. Soc., 1908, xiii., 103) state that they were able partially to reduce silica by means of an alloy of Mg-Ca-Al when the mixture was heated in a steel vacuum tube to a high temperature; also that a mixture of MnO2 and SiO2 when heated with the same alloy in the steel vacuum tube produced an alloy of Si-Mn (22 per cent of the theoretical yield). It thus appears from these isolated cases that certain "Thermitic" reactions which take place readily in the air FIG. 1. TO PUMP TO PUMP FIG. 2. since that time, it was thought advisable to give further details. The authors are only able to find reference to two attempts where similar reactions have been tried in vacuo. Pring (Fourn. Chem. Soc., 1908, 2108) heated Al and C in an electric vacuum furnace and found that at 640° C., the melting-point of Al, combination begins to take place and that at 1400° C. it proceeds rapidly. From the description of the experiment it appears that the reaction once started does not proceed throughout the mass, whereas, as shown by the authors, a mixture of 4A1 and * A Paper read before the Faraday Society, May 31, 1910. are only partially reproduced in vacuo, and only then when continuously heated to a high temperature. Experimental Methods. The following methods have been used by the authors with varying degrees of success. Method 1.-The mixture is placed in a mild steel testtube, 18 in. long by in. diameter; the tube is connected both to a mercury manometer and a Töpler pump. It is heated in a Fletcher muffle furnace to a temperature of 1100° C. The projecting part of the test-tube is cooled by cold water in order to preserve the connection between the steel tube and the glass tube; the cementing material con sisted of an intimate mixture of fine asbestos wool, litharge, and water-glass (see Fig. 1). 1. Mixtures of Mg and Al2O3 when heated in this apparatus reacted with explosive violence. 2. Mixtures of Mg and CaO when heated in these tubes did not react, but the Mg distilled from the mixtures and condensed in the cool part of the tube. After several trials with different mixtures it was found that the method of heating the mixtures was unsatisfactory and so the method was abandoned. | bedded in the mixture, a channel was immediately formed on heating the wire. 2. Thermite reacted in this apparatus, but invariably broke the tube. 3. All mixtures of Al and Na2O2 tried reacted with separation of metallic Na, but invariably broke the tube. Method 3.-After several trials with small bottles, using Pt wire and Fe wire as the initial source of heating, it was found that certain fuses could be easily fired by Fe wire heated by an electric current, and finally, the following TO PUMP FIG. 3. Method 2.-The mixture, contained in a porcelain boat, was placed in a Jena glass tube 12 in. long by 1 in. internal diameter (see Fig. 2). A coil of thin platinum wire lying in the mixture was connected to two stout copper wires, which were embedded in glass tubes to the length of about 6 in. by means of Faraday's cement; these tubes passed through a rubber cork and could be connected to a source of electric current. The other end of the Jena tube was connected to a mercury manometer and the Töpler pump. It was very soon ascertained that the heat produced in the Pt wire by the passage of the electric current was insufficient to start the reaction of the mixtures, even when the Pt was raised to its melting-point (1780° C.), except in a very few cases, and then the tube invariably broke through the sudden great evolution of heat, e.g. 1. Mixtures of Mg and CaO were repelled from the heated Pt wire, and even when the Pt wire was well em apparatus was designed and found to give the best results. A wide-mouth bottle of three litres capacity was fitted with a rubber cork carrying three glass tubes; two of these tubes carried stout copper wires, which were embedded in Faraday cement for a length of 6 in. and at the top end in 2 in. of paraffin wax. The third tube was connected to a Töpler pump and a mercury manometer (see Fig. 3). The bottom of the bottle was covered with a layer of sand mixed with small lumps of quicklime-this mixture had been previously heated in a muffle furnace and finally cooled in a desiccator. A Hessian crucible holding from 20 to 30 gims. of any of the mixtures was embedded in the sand, and on top of the mixture was placed o'i grm. of the fuse contained in a very thin piece of tissue paper; on inserting the cork, a loop of very thin iron wire, connected to the two copper leads, just lay embedded in the fuse. The jar was first exhausted with a water-pump, and then finally by a Töpler pump; then the fuse was fired by passing an electric current through the iron wire from one to two seconds. The reactions that took place always did so quietly and steadily. In no case did the crucible crack and a very small amount of the mixture was always left unchanged. The pressure in the apparatus, after the reaction was completed, was usually from 50 to 80 mm.due to combustion of the piece of paper used with the fuse. (To be continued). the investigation of the duration of calcium rays confirm the conclusion drawn from the iron rays, viz., that the duration is approximately proportional to the intensity. In some cases the duration of the rays indicates the nature of the impurities present in a substance. The author thus found that the calcium he used contained traces of strontium, magnesium, aluminium, and manganese. Photographic Method of Counting a-Particles.William Duane.-The author describes a method of counting the a-particles emitted by a radio-active substance. The particles enter the ionisation chamber through CHEMICAL NOTICES FROM FOREIGN a thin wall of mica, and the ionisation due to each particle SOURCES. Comptes Rendus Hebdomadaires des Séances de l'Academie des Sciences. Vol. cli., No. 2, July 11, 1910. Density of Radium Emanation.-Sir William Ramsay and Robert Whytlaw Gray.-The most conclusive way to establish the atomic weight of a gaseous element is to determine its density. If the element is monatomic both the atomic and molecular weight are obtained by doubling the density. Radium emanation is without doubt one of the inert gases, and it is only necessary to find its density in order to fit it into the Periodic Table. This the authors have done by direct weighings, using a quartz balance (sensitive to half a millionth of a milligrm.). The following table gives the results of five determinations:— Thus the emanation is the second member of the series of inert gases, after xenon. Since it is undoubtedly an element its former cumbrous designation might conveniently be dropped, and the authors suggest as a suitable name Niton (shining), symbol Ni. Critical Constants of Acetylene and Cyanogen.Ettore Cardoso and Georges Baume.-The mean values of the critical constants of acetylene and cyanogen are as follows: = = = 61.5 at. 59.6 at. = C2H2.tc= 35.5° (308 5 abs.), pc C2N2.tc 128.3° (4013 abs.), pc Determination of Methylamine in Presence of Large Amounts of Ammonia. — J. Bertheaume. Methylamine replaces ammonia in its salts, and this property can be used as the basis of a method of determining it in presence of large quantities of ammonia. The mixed gases are led through a series of flasks containing hydrochloric acid, and thus the amines are concentrated in the first flasks. Fixation of Trioxymethylene by the Magnesium Derivatives of the Homologues of Benzyl Bromide. -P. Carré.-The bromide of mesityl magnesium reacts with trioxymethylene in the same way as the chloride of benzyl magnesium, giving mesitylcarbinol and a fairly large proportion of the corresponding ether oxide. No. 3, July 18, 1910. Action of Ultra-violet Rays on Gelatin.-A. Tian. -The ultra-violet rays do not appear to have any action on gelatin nor on its dilute solutions. The same rays destroy gels, causing their liquefaction or, in presence of water, their solution. The waves producing these changes have wave-lengths below 3000 A. Thus they belong to the same spectral region as those which coagulate the albumens. Relative Duration of Calcium Rays in the Selfinduction Spark.-G. A. Hemsalech.-The results of is registered by means of a sensitive electroscope or a quadrant electrometer. A photographic record of the movements of the gold leaf enables the effect of each particle to be seen. Thorium Sulphate.-M. Barre. The specific resistance and the freezing-point of thorium sulphate in aqueous solution agree perfectly with the laws of Bonty and of Raoult. The measurement of the conductivity and of the freezing-point of solutions containing 1 per cent of potassium sulphate indicate the formation of a stable double salt, and thus the results of determinations of the solubility are confirmed. Action of Ether Salts of Fatty Monobasic Acids F. Bodroux.-When the ether salts of monobasic fatty on the Monosodium Derivative of Benzyl Cyanide.— acids act on the monosodium derivative of benzyl cyanide CN substances of formula are obtained. C6H5-CH-CO-R The author has prepared and investigated the properties of some of these nitriles, e.g., phenylpropanal nitrile. They are unstable bodies which decompose in dilute alkaline solution at the ordinary temperature. Absorption of Iodine by Solid Substances.-Marcel Guichard. The fixation of iodine at the surface of a solid is a specific property, and many substances exhibit the property to a very slight extent only. The power of absorption varies with the circumstances and with the preparation of the solid in question. MISCELLANEOUS. The Birkbeck College.-The 88th Session of the College will commence on Wednesday, September 28th. Opening Address will be given in the Theatre, at 7.30 p.m., by Prof. M. E. Sadler, M.A. The Class Rooms, &c., will afterwards be open for inspection, and there will be an Exhibition in the Art School. Visitors are invited. The College is conducted in relation with the University of London; Classes are held both in the Day and Evening; twenty-nine Members of the Staff are Recognised Teachers of the University. The courses of study provides for Degrees in the Faculties of Arts, Science, Laws, and Economics. There is a very complete curriculum for Chemistry, Physics, Mathematics, Botany, Zoology, and Geology, The Laboratories are well equipped with modern apparatus and appliances, and Research Work is encouraged in all the Science Departments. The Arts Courses include English, Classics, French, German, Italian, History, Logic, Mathematics, Economics, and Geography, all conducted by Recognised Teachers of the University. The Law Courses comprise Mercantile and Common Law, Equity, Conveyancing, and Patent Law, Classes in English and Roman Law, and Constitutional Law and History. According to the Calendar more than 140 Students passed some Examination of the University during the last Session; forty-five took Degrees in Arts or Science, twenty with Honours, and several Students gained distinction at other Universities. The College possesses an Art School, which provides instruction in the various Branches of Artistic Work (Painting, Modelling, Life, Design, &c.). NEWS THE CHEMICAL VOL. CII., No. 2653. SCANDIUM.* PART II.t trace of scandium remains in the filtrate; the precipitate NEWS. contains only a very small amount of yttrium earths which can be removed by repeated precipitation with thiosulphate (cf. pp. 147-149 of the first article). When this separation was carried out quantitatively 44 grins. of rare earths gave 22 grms. of Sc2O3 (with small amounts of ThO2), and 17 grms. of yttrium earths; the residue of 5 grms. consisted of iron, manganese, calcium, lead, &c.; apart from these "not rare earths" the mixture of earths of wolframite from Zinnwald consists of 564 per cent Sc2O3 and 43'6 per cent of other rare earths. The mixture of oxides obtained by direct precipitation with sodium silicofluoride is much richer in scandium. To investigate it 48 gims. of the nearly pure white product were converted into chloride and precipitated with sodium thiosulphate. In the filtrate 32 grms. of yttrium earths were dissolved. Hence the oxide from the silicofluoride precipitation contains 93.3 per cent Sc2O3 and 6.7 per cent of other earths. By R. J. MEYER and HERBERT WINTER. For completeness sake the quantitative analysis of the wolframite from Zinnwald is given below. The material consisted of black glittering pieces with single crystal surfaces. It was carefully picked out for the analysis with a microscope :Wolframite from Zinnwald. WO3 FeO .. PbO SnO2 Per cent. Per cent. 75°41 9'34 9.36 14'00 14'15 0.18 0'35 TiO2 + Ta205 Occurrence and Preparation in the Pure State. In the first article on scandium, to supplement G. Eberhard's spectrographic experiments (Sitzungsber. Kgl. Preuss. Akad. Wissensch., 1908, xxxviii., 851), new methods were described which had led to the separation of scandium from the wolframite of the Ore Mountains in Saxony. These methods essentially depended upon the insolubility of scandium fluoride in presence of free mineral acid, and upon the fact that scandium can be precipitated with sodium thiosulphate. (See also D.R.P. 202523 and 208355, Class 40 A, Group 51). The initial material consisted principally of the residues of iron and manganese oxide, which remain undissolved after boiling up the alkaline melt of wolframite. If this mixture of oxides is dissolved in hydrochloric acid, so that a nearly neutral solution is obtained, the rare earths contained in the wolframite can be precipitated from this solution by means of a large excess of oxalic acid. The precipitate also contains manganese, calcium, iron, and lead oxalates. To purify it further, this product in which the scandium has accumulated is precipitated in acid solution with hydrofluoric acid. A second and simpler method consists in the direct precipitation of the hydrochloric acid solution of the Thus the Zinnwald wolframite contains about o'I per original mixture of oxides with sodium silicofluoride. cent of pure scandium earth. Wolframite containing When fresh scandium has repeatedly to be prepared from scandium is also found in the Sadisorf Copper Mine near wolframite residues this method is found very convenient. Schmiedeberg in the Ore Mountains (Eberhard," Uber die The hydrofluosilicic acid separates the earth in a very pure weite Verbreitung des Scandiums auf der Erde II.," state, and it is preferable to employ this process rather Sitzungsber. der Kgl. Preuss. Akad., April 21, 1910, xxii., than the combined oxalic acid and hydrofluoric acid 404). The wolframite from this place differs from that method, when the object is to obtain the scandium alone. found at Zinnwald-Altenberg in the fact that the proporBut if all the earths of wolframite are to be studied it is tion of iron to manganese is reversed. The specimens essential to perform the separation with oxalic acid. We which we had at our disposal could be separated from have recently given special attention to the qualitative and accessory constituents only with difficulty. Pieces that quantitative composition of the earths contained in had not been picked out with sufficient care gave on wolframite, and have found that the relative amount of analysis only about 70 per cent of tungstic acid, and 2 to 2'3 scandium oxide in it is not so high as was supposed when per cent of titanic, niobic, and tantalic acids, some molybdic these circumstances were not taken into consideration. acid, and a not inconsiderable amount of calcium phosFor further investigation we used a raw scandium oxide phate. Sadisorf wolframite contains rather more rare which had been obtained only by repeated oxalic acid pre-earths than that from Zinnwald, about o‘2 per cent being cipitations from the solution of the oxide residues, and found in it. The amount of scandium may be about the which must thus contain all the rare earths of the wolframite. same, but the earth isolated from the Sadisorf mineral with It was a yellowish powder, the solution of which in hydro- sodium silicofluoride and thiosulphate contains more chloric acid on spectroscopic examination showed very thorium oxide as the spectrographic examination shows. faintly the principal bands of didymium. The cerium The following analysis refers to a specimen picked out with reaction was correspondingly faint, but the bands of the the greatest possible care :erbium earths appeared fairly clearly. Thus it was a mixture of earths, consisting essentially of yttrium earths. fairly pure ore (Stuferz). The residues from the more impure "Schlicherz" and "Setzerz " are generally much poorer in scandium, corresponding to the smaller amount of tungstic acid in this material. Thus a specimen of wolframite from Zinnwald with 50 per cent of WO3 gave only o'04 per cent of Sc2O3. On the other hand, we have occasionally come across residues supposed to be derived from slick or small ores which contained the normal percentage (03) of Sc2O3. The different compositions of these products seem to depend upon accidents of locality, and also upon the kind of preliminary separation and working up. To test rapidly whether a material is worth working up for scandium, the hydrochloric acid solution of 100 grms. of the residue may be boiled with sodium silicofluoride, the precipitate obtained decomposed with sulphuric acid, and the solution of the sulphate residue precipitated with ammonia. The colourless oxide obtained from the hydroxide or oxalate should weigh not less than 0.3 grm. if the quality of the product worked up is good. As Eberhard has shown, besides certain wolframites from the Ore Mountains, the tinstone accompanying them contains much scandium. We have not studied this material ourselves, but have investigated fairly thoroughly the tin-slacks which are obtained from the slopes, often centuries old, of former tin works in the Ore Mountains. In the slacks obtained from the smelting of Zinnwald ores only the more important constituents were determined analytically. A historical description is given of the method of smelting employed hundreds of years ago in the Inaugural Dissertation by H. Winter, "Vorkommen und Reindarstellung des Scandiums," Berlin, 1910, in which urther information is to be found relating to the analysis of wolframites and tin-slacks. Tin-slack from Zinnwald. Per cent. 8:513.2 per cent met. tin 16.0 30'0 18.4 g'o Sn (met.) SnO2 SiO2 WO3 FeO MnO The residue consists chiefly of calcium phosphate, derived from the fluxes. To determine the amount of scandium present the slack was opened up with soda in an iron crucible, and after the silicic and tungstic acids had been separated with hydrochloric acid the earths were precipitated with oxalic acid or sodium silicofluoride, and then purified by a thiosulphate precipitation. With two different slacks the amount of scandium oxide was found in both cases to be o'09 per cent, which is about the same as that contained in wolframites. The high percentage of tin in the slacks shows how irrational were the old methods of smelting the tin ores of the Ore Mountains. scandium (see Eberhard, Sitzungsber. d. Kgl. Preuss. Akad., 1910, xxii., 413, No. 657 of the minerals examined). In answer to our enquiries, Sir William Crookes then most kindly informed us that the supply of wiikite containing scandium which was extracted for him at a place in Finland was exhausted. But, as appears from the more recent researches of Eberhard (loc. cit.), many different minerals, varying in appearance and colour, are often described under the name wiikite, when their chemical nature is not sufficiently well known, and when possibly they belong to different species. They are all found in Finland, on the Northern shores of Lake Ladoga in the parish of Impilaks, and the two most important localities in which they are found are the islands Lokansaari and Nuoliniemi on the coasts (Bull. de la Commission Géologique de Finlande, Helsingfors, 1907, No. 19, “Die Erzlagerstätten von Pitkäranta am Ladogasee von Otto Trüstedt "; with a geological chart 1: 40,090), Among the specimens of these minerals investigated spectrographically by Eberhard (No. 647-656) some contain considerable percentages of scandium. It would be interesting to examine them analytically in order to be able to form an opinion concerning their chemical character and their relations to one another. Preparation of Pure Scandium. Separation of Thorium. By using the methods of separation and purification described in the first article, we obtained a raw scandiumi the atomic weight of which was about 46. The result of the spectrographic and chemical examination showed unmistakably that of known impurities, only thorium was present. Repeated experiments have convinced us that the earth precipitated with sodium silicofluoride, and then repeatedly purified with sodium thiosulphate, no longer contains any foreign yttrium earths; in the arc spectrum neither yttrium nor ytterbium can be detected. This result is all the more remarkable because the ytterbium earths are extremely persistently retained in the last frac tions by the scandium earths in the method used by Nilson, Cleve, and Crookes of the fractional decomposition of the nitrates. This may be explained by the very small difference of basicity of the two earths. The problem still to be solved is thus the separation of thorium from scandium. At the end of the first article allusion was made to the difficulty of effecting this separation. Even from a qualitative point of view thorium in presence of a large excess of scandium shows certain deviations from its normal behaviour. Thus it is not possible to precipitate the thorium from scandium containing thorium in neutral solution by means of hydrogen peroxide a reaction which enables thorium to be separated quantitatively from the other earths. By synthetic experiments we have shown that the amount of thorium salt present must be very considerable in order that precipitation with hydrogen peroxide may occur. This hindering of a reaction can only be explained by the well marked tendency of scandium to form strong complexes. We have finally found only one single reaction by which thorium can be detected in presence of scandium with certainty; it depends on the action of iodic acid in strongly acid solution (see R. J. Meyer and M. Speter, Chem. Ztg., 1910, No. 35). The solution in sulphuric acid or nitric acid is treated with an excess of iodic acid in nitric acid. Scandium periodate, which is simultaneously precipitated, is again dissolved if sufficient acid is added, while the very difficultly soluble thorium iodate remains undissolved even in a very great excess of acid. Moreover, very small amounts of thorium in the scandium are revealed by the fact that the neutral solution of the chloride, when it is strongly diluted, becomes turbid after some time, owing to the separation of thorium oxychloride, while solutions free from thorium remain clear. The articles of Sir William Crookes (Proc. Roy. Soc., Series A, 1908, lxxx., No. A 541, p. 516; in full in Phil. Trans. Roy. Soc., Series A, 1908, p. 208; translated in Zeit. Anorg. Chem., 1909, Ixi., 349; also CHEMICAL NEWS, vol. xcviii., p. 274 et seq.) directed our attention to the Finnish wiikite which was said to contain more than I per cent of Sc2O3. According to the specified composition this mineral must be allied to the euxenites, which, when obtained from Norway, are said to contain o‘06 cent of scandium earth, as Hauser and Wirth have quite recently stated (Ber., 1909, xlii., 4445). The examination of wiikite, 500 grms. of which we obtained from Herr A. H. Petra, of Helsingfors, gave the astonishing result that hardly any scandium could be detected in it. In agreement with Crookes we found, after the separation of the acid metals (TiO2, Ta2O5, Nb2O5) and of the silicic acid, about 13 per cent of rare earths, viz., 9 per cent of cerite and yttrium earths, and 4 per cent of thorium oxide, but scandium could not be detected with certainty in the part precipitated with thiosulphate. The spectrographic examination of the earths also revealed only a small amount of We will now describe the experiments which were undertaken to separate thorium from scandium. The initial material for this research consisted of about 70 grms. of scandium oxide of mean atomic weight 48 0. |