ii. 28 ABSTRACTS OF CHEMICAL PAPERS. Inorganic Chemistry. Preparation and Testing of Hydrogen of a High Degree of Purity. J u ~ r u s DAVID EDWARDS (J. Ind. Eng. C h m . 1919 11 961-963).-An apparatus has been devised as a substitute for the Kipp generator. It consists of two connected cylinders into one of which the zinc is introduced whilst a tube through a rubber cork in the bottom is connected with the acid flask. The outlet tube has a side connexion which dips into mercury and serves as a pressure indicator and as a safety valve. Air is exhausted from the cylin-INORGANIC CHEMISTRY. ii. 29 ders the acid flask nearly filled with dilute sulphuric acid and a few fragments of zinc introduced to sweep out the air before the flask is attached to the cylinders. The acid is drawn up into the gsnerator by turning a tap a t the top and opening a vent in the stopper of the acid flask.Pure hydrogen is thus a t once obtained and fresh acid introduced without admitting air. Tests made in c6mparison with pure electrolytic hydrogen by means of the inter- ferometer (A 1315 ii 478) showed that hydrogen produced by this apparatus contained less than 1 part in 10,000 of impurities. I n similar tests of the gas produced by a Kipp generator 83'86% of hydrogen was found to be present seventeen minutes and 99*377L fGI'ty-tWO minutes after starting. [See also J. Soc. Chem. Id. 1920 19A.1 C. A. M. New Experiment to Demonstrate the Minute Mass of a Hydrogen Atom. FR. B ~ ~ R K I (Helv. Chim. Acta 1919 2 703-704).--A yellowish-green fluorescence is distinctly observable in a darkened room when a beam of light is passed through 0.5 C.C.of a solution prepared by dissolving 0.4 mg. of fluorescein in 10 litres of water. The total quantity of hydrogen which can be furnished by the weight of fluorescein present in this solution is 6 x 10-11 grams which therefore gives an upper limit for the mass of the hydrogen atom. The figure is greatly in excess of the gener- ally accepted value. H. W. Composition of the Atmosphere. AUGUST KROGH (Math. fysiske illeddelelser 1919 1 No. 12 1-13).-Our knowledge of the composition of the atmosphere is still quite uncertain especially in regard to hydrogen and the relation between oxygen and nitro- gen. Until very recently little attention has been paid to the varia- tion in the mixture composing the atmosphere at different altitudes.Wsgener claims that a t a height of 70 km. hydrogen must be the chief constituent of the earth's atmosphere. Benedict has shown by 200 analyses extending over a period of nine months that the percentage of oxygen a t the surface or laboratory level never varied more than kO.01 from the average and that of carbon dioxide never more than *0*005. The author describes a new gas analysis appaiatus with three gas burettes one for moving the air sample to and from the absorption pipettes and for saturating the air with moistwe a second convenient in size for measuring the air before and after the ahsorption of carbon dioxide and a third which holds the residual gas after the absorption of oxygen. The author dis- cusses general means for reaching mare accurate and constant results.The heated platinum wire method has been applied for the detection of combustible gases and the conclusion is drawn that hydrogen or other combustible gas is present in an amount which is probably less than 0.002%. The percentage of carbon dioxide in the streets of Copenhagen is usually increased by 0.001-0.007. The need of a thorough research to establish the absolute average com- position of pure atmospheric air with more certainty and a studyii. 30 ABSTRACTS OF CHEMICAL PAPERS. of its variatious especially iii the higher strata of the atmosphere is strongly urged. Accurate analyses of samples taken simultane- ously from aeroplanes a t various heights up to 5 or 6 km. would give valuable inforniat,ion in regard to mixing by vertical currents.Automatic sampling by balloons is suggested. U p to the present time the absolute composition of atmospheric air a t the earth's surface has been found t o be CO 0'030%; N 79.022%; and 0,. 20.948%. CHEMICAL ABSTRACTS. Nitrogen Generator for Laboratory Use. W. L. BADGER (./. 1n.d. Lng. C!/ieni. 1919 11 1052-1053).-A widenecked 2-litre bottle is closed with a rubber stopper through which passes one end of a Liebig condenser jacket (without the condenser tube) ; the lower water-inlet is sealed off and the upper one is connected wit,h a side- tube which extends through the stopper to the bottom of the bottle. A bulb with delivery tube is sealed on the upper end of the jacket. The boltle is packed with copper wire and as much animon(a (1 1) saturated with ammonium chloride is poured in as the bottle will hold.'The jacket is filled with copper turn- ings. Air is blown in through a tube reaching nearly to the bottom of the bottle; the oxygen is absorbed rapidly and practically pure nitrogel? escapes from the delivery tube. The passage of the air (or nitrogen) causes the solution to circulate slowly up the jacket and through the side-tube into the bottle again. w. P. s. The Synthesis of Ammonia at Very High Pressures. GEORGES CLAUDE (Compt. rend. 1919 169 1039-1041).-Having previously shown the possibility of using very high pressures for industrial purposes (ibid. 649) the aut,hor has studied the synthesis of ammonia a t these high pressures and finds that a t a pressure of 1000 kilos./cni.2 and a t a temperature of 5 3 6 O the yield of ammonia is over 40%.The zone of temperature within which this reaction takes place a t a reasonable velocity under such a high pressure is 500-700'. W. G. Behaviour of the Hydronitrogens [Nitrogen Hydrides] and their Derivatives in Liquid Ammonia. VI. Electro- lytic Nitridation of Various Anodes in a Solution of Ammonium Trinitride. A. W. BROWNE M. E. HOLMES and J. S . KING ( J . Anaer. Chem. S O C . 1919 41 1769-1776. Compare A. 1911 ii 1084 1085; 1913 ii 583).-Solutions of 1 gram of ammonium a i d e in 36 C.C. of liquid ammonia were electrolysed a t - 67O using anodes of copper silver cadmium aluminium lead antimony iron and nickel. The gas evolved a t both anode and cathode was measured and the loss of weight of the anode deter- mined. It is shown t h a t copper anodes undergo electrolytic corro- sion to an extent indicating the formation of some cuprous azide CuN with cupric azide CuN as the main product.No gas was liberated a t the anode. With silver. cadmium lead and antimony anodes the corrosion resulted in the formation of normal azides,INORGANIC CHEMISTRY. ii. 31 AgN CdN PbN and SbK9 without liberation of gas a t tlie anode. Aluminium iron and nickel anodes undergo corrosion accompanied by the liberation of nitrogen. The aluminiuni anode became coated with a bulky pyrophoric scale of varying colour and texture. Deep red ferric azide FeN was obtained in solution when a n iron anode was employed but this product was ammono- lysed and yielded an ammono-basic ferric azide.A pink deposit was formed on the nickel anode presumably an ammono-basic nickel azide. J. F. S. Some Properties of Nitrogen Trioxide Pure or in Solution in Nitrogen Peroxide. GEORGES BAUME and MARIUS HOBERT (Compt. rend. 1919 169 968-970).-The authors have studied the melting-point and vapour-pressure diagrams of the system nitrogen trioxide-nitrogen peroxide. The melting-point diagram is normal and has a single eutectic in the 'neighbourhood of the freezing point of pure nitrogen trioxide. From the vapour-pressure curves between -80c and 3 5 O the author deduces the b. p. of pure nitrogen trioxide to be - 2 F / 7 6 0 nun. Pure nitrogen trioxide can only exist a t very low temperatures in the solid state or in the liquid state under pressure of nitric oxide.At temperatures above - l O O c it dissociates. It is not possible to distil it in a vacuum because of the immediate formation of an atmosphere of nitric oxide a t temperatures a t which distillation is possible. W. G . Silicon Hydrides. VII. Protosiloxane O:SiH,. ALFRED STOCK and KARL SOMIESKI (Ber. 1919 52 [R] 1851-1860. Com- pare A 1918 ii lll).-In the earlier paper it was shown t h a t di- bromomonosilane reacts with water to form polymerides of protosil- oxane 0:SiH2. The unimolecular form has now been obtained as a gas by the action of the required amount of water vapour on dichloro- inoiiosilane in a very large Aask under greatly reduced pressure. It has a n extraordinary tendency to polymerise much more so than the analogous carbon compound formaldehyde in consequence of which the flask must be perfectly clean and smooth.Liquid and solid polymerides are formed immediately on condensation. The liquid ones are like benzene and can be obtained conveniently as a solu- tion by shaking a benzene solution of dichloroinonosilane with water. These benzene solutions are fairly stable towards water but reduce silver nitrate in the cold. They correspond roughly with (SiH,?),. The solid polymerides are insoluble. All the polymerides react with sodium hydroxide according to the equation SiH,O + 2NaOH = Na,SiO + 2I-I,. Constitutional Formulae of Kaolinite and other Silicates. RUDOLF WEGSCHEIDER (Zebtsch. Elektrochem. 1919 25 352).-It is shown t h a t the constitutional formulz of polysilicates must be wiitten with the silicon atoms joined through oxygen and not directly with one another.Compounds with the silicon atoms directly united are reducing agents and unstable. Further if the polysili- J. C. W.ij. 32 ABSTRACTS OF CHEMICAL PAPERS. cates are considered as having the silicon atoms directly united it will follow that the oxygen atoms must also be directly united which again will point t o instability. The combination of a strongly reducing group -Si-Si- with a strongly oxidising group -0-0- will of necessity give a compound of an extremely unstable charac- ter which is certainly not the case with the polysilicates. [See also Simmonds T. 1904 85 681; Pukall A . 1910 ii 780; Manchot A. 1910 ii 1060.1 Theory of Binary Mixtures. V. Vapour Pressure and Molecular Constitution of Liquid Argon and Argon-Nitrogen Mixtures.F. UOLBZALEK (Zeztsch. physzkal. Chem. 1919 93 585-595. Compare A. 1916 ii 132).-A theoretical paper in which using the experimental data previously published the asso- ciation constants are calculated from the saturation pressure of the mixture and also the molecular constitution of the mixtures and pure argon the saturation pressures of the liquid mixtures and the constitution of the rapour. It is shown that liquid argon .is partly composed of diatomic molecules and has an association con- stant a t 85’11O abs. of 11=0*20. Hence a t low temperatures argon possesses chemical affinity and consequently argon compounds are to be expected a t low temperatures. There is therefore no funda- mental difference between the inactive gases and other gases but only a difference of degree. When liquid argon is diluted with liquid nitrogen the diatomic molecules dissociate according to the law of mass action.The vapour tension of liquid argon-nitrogen mixtures as well as the constitution of the vapour may be calcu- lated a t all concentrations by means of the author’s theory of solution (A. 1909 ii 2 2 ) . J. F. S. J. F. S. Preparation of Metallic Potassium. FRITZ C. WICKEL and WALTER LOEBEL (D.R.-P. 307175; from Chem. Zentr. 1919 iv 361). -Accurately measured quantities of metallic sodium and potassium hydroxide are melted together with exclusion of air so that sodium oxide is formed and potassium is volatilised from the mixture (at about 670°) and suitably condensed.Hydrogen is obtained as by-product. The process is suitable for the manufacturing scale. H. W. The Autoxidation of Sodamide. HANS SCHHADER (Zeitsch. anorg. Chem. 1919 108 44-48).-The amides of the alkali metals are known to undergo autoxidation with the formation of nitrite hydroxide and ammonia. When finely divided sodamide is exposed to air in presence of a little water a’ yellowish-red oxida- tion product is formed which is now shown to be a peroxide probably of the formula NaNH,,O,. The peroxide formed was estimated by adding the product to a cold saturated solution of barium chloride when barium peroxide was precipitated and was estimated in the usual way It was found that a sample of sodamide gave in fifty-eight days a t ordinary temperature 0.44INORQANIU CHEMISTRY. ii.33 mol. % peroxide and 6.9% nitrite. I n dry air autoxidation does not take place a t the ordinary temperature but a t looo to llOo the peroxide is slowly formed. The peroxide is stable in dry air but in moist air is changed into a white substance the aqueous solution of which gives the peroxide reaction. Solubility and Fusion Relations at High Temperatures and Pressures. GEORGE W. MORBY ( J . Eng. Club Philadelphia 1919 35 509-519).-An address. The preparation of the com- pounds monopotassium aluminate K,O,Al,O large octahedral crystals and monopotassium ferrite KoO,Feo_Os large red octa- hedral crystals decomposed by water below 500° is mentioned. E. H. R. CHEMICAL ABSTRACTS. The Properties of Ammonium Nitrate. I. The Freezing REGINALD GEORGE EARLY Point and Transition-temperatures .and THOMAS MARTIN LOWRY (T. 1919 115 1387-1404). Action of Bromine on Calcium Carbide. EDWARD BARNES (Chem. News 1919 119 260-261).-Liquid bromine acts slowly on calcium carbide a t ordinary temperatures producing carbon hexabromide and calcium bromide. 4.5 Grams of finely powdered calcium carbide treated with 45 grams of purified dry bromine in a 50 C.C. bottle for five months gave a product which after removal of excess of bromine weighed 32.4 grams and contained 2 2 grams of hexabromoethane 8.8 grams of calcium bromide and 0.2 gram of unchanged calcium carbide. Calcium carbide and bromine heated in sealed tubes a t looo until no further action was apparent produced carbon and calcium bromide. Dry chlorine was found to have no action on calcium carbide exposed to it a t the ordinary temperature for two months.S. 8. A. Reduction of Barium Nitrate by the Alternating Current. P. WENGER and A. LUBOMIRSKI (Ann. Chim. anal. 1919 [ii] 1 339-342) .-With lead electrodes the yield of nitrite increases with the amperage but with mercury electrodes the maximum yield is obtained with 0'6 ampere ; the yield also increases with temperature in the case of lead electrodes and decreases with aluminium or zinc electrodes. Temperature has little if any effect when copper cadmium or magnesium electrodes are employed. The oxygen liberated from the nitrate combines with the electrode forming suboxides (copper mercury cadmium and tin electrodes) hydr- oxides (lead zinc aluminium and magnesium electrodes) or oxides (nickel and silver electrodes).The nitrate is not reduced by the Conditions in which Periclase or Crystalline Magnesia is Formed. J . MEUNIER (Bull. SOC. chim. 1919 [iv] 25 560-562).-Magnesium oxide may be freed from contained calcium by igniting the material a t a red heat and metallic electrodes unless the current is passing. w. P. s. Purification of Magnesia VOL. oxvm. ii. 2ii. 34 ABSTRACTS OF CHEMICAL PAPERS. subsequently after cooling extracting it with successive quantities of a 10% sucrose solution until the last extract does not give any turbidity with ammonium oxalate. If such material is then washed with water dissolved in hydrochloric acid the solution evaporated to dryness and the residue gradually heated in a covered crucible until it attains a red heat magnesium oxide is obtained in the form of crystals of periclase on the lid and sides of the crucible and on the surface of the amorphous oxide in the crucible.The material so obtained is very pure. Concentrated Thallium Amalgams their Electro- chemical and Thermochemical Behaviour ; Densities and Freezing Points. THEODORE W. RICHARDS and FARRINGTON DANIELS ( J . Amer. Chem. SOC. 1919 41 1732-1768).-The electromotive force density specific volume specific heat heat of dilution and freezing point of thallium amalgams up to a Concentration of 40‘9% thallium have been experimentally determined. The heat of solu- tion of thallium in mercury and in various amalgams has also been determined. The E.M.F. measurements were made a t 20° 30° and 40° and are shown to be much greater than demanded by the simple concentration law.The temperature-coefficients of the E.M.F.’s have been calculated for the various amalgams and they are found to be less than is demanded by the gas law and vary as the concentration increases. The solution volume of thallium is remarkably constant varying only from 17.47 to 17.51 but it is somewhat larger than the specific volume (17.21) of thallium. Thallium therefore must expand on amalgamation. The coefficients of expansion were calculated from these results and found to diminish with increasing concentration. The heat capacities were found t o be distinctly larger than the sum of the heat capacities of the thallium and mercury in the amalgam. This excess is greater in dilute than in concentrated amalgams.The heat of dilution with mercury was determined a t 20° and from the results the values a t 30° and 3 2 5 O were calculated. Although the heat of dilution of thallium amalgam is a very different effect from the transfer of thallium from one amalgam to another it is shown that either of these effects may be calculated from the other. Within the limits of experimental error the heat effects in the cells of which the E.M.F. was measured may be calculated from either the heat of dilution of the amalgams or the heat of solution of thallium in the amalgams or the temperature-coefficient of the E.M.F. each of these three methods giving essentially identical results within the limits of experimental error. The freezing points of amalgams containing from 16 to 45 atomic % of thallium were found to give a curve indicating conclusively the existence of the solid compound Tl,Hg,. The single potential of pure electrolytic thallium was found to be about 2’5 m.v. more negative than that of saturated thallium a t the ordinary temperature. Nitrous-Nitric Complexes of Thallium.L. ROLLA and G. BELLADEN (Gazzetta 1919 49 ii 217-224).-Thallium nitrite W. G . J. F. S.IXORGANIC CHEMISTRY. ii. 35 like barium strontium and calcium nitrites (compare Vogel A. 19d3 ii 391) has a specific electrolytic conductivity below that of the corresponding nitrate and undergoes normal dissociation. The B.M.F. of the cell thallium amalgam 1 -IT/ 10-thallium nitrite 1 normal electrode is 0.667 volt a t 25O the same value being obtained if the nitrite is replaced by the nitrate.Peters found that lead nitrate and nitrite interact in solution. forming well-defined compounds containing both salts. Similarly thallium nitrite and lead nitrate react yielding compounds contain- ing Pb T1 NO NO and OH. With potassium nitrite and thallium nitrate however despite wide variations in the concentra- tions of the reacting solutions no product containing potassium could be obtained ; the nitrate-nitrites formed although they have compositions differing greatly from those of Peters' salts are stable and may be recrystallised unchanged. If the two reacting salts are taken in molecular proportions the resultant product has the percentage composition T1 76.65 NO 20.58 and NO 2.76 whilst in the proportion TlNO and 2KN0 the compound obtained has the formula Tl,N,O the ratio of NO to NO in the complex ion being 2 1.With 3 mols. of potassium nitrite and 1 mol. of thallium nitrate the product has the percentage composition T1 82.78 NO 6.64 NO 5.80 and OH 4.78 and with 4 mols. of the nitrite to 1 mol. of the nitrate the composition is T1 86.22 NO 1.27 NO 4.17. and OH 8.34. T. H. P. Investigations on some Rapid Low Temperature Reactions by means of Heating Curves. J. ARVID HEDVALL and NILS VON ZWEIGBERGK (Zeitsch. anorg. C'hem. 1919 108 119-136).- I n a previous paper (A. 1919 ii 26) it was shown that the decom- position of barium peroxide takes place a t a much lower tempera- ture in presence of all modifications of silica than when heated alone a certain amount of barium silicate being formed a t the same time.The behaviour of a great number of other oxides when heated with barium peroxide has now been studied by examining the heating curves of the mixtures in molecular proportions. Cuprous oxide reacts violently with barium peroxide a t about 1 30° and is completely oxidised to cupric oxide. The latter decomposes barium peroxide catalytically the optimum temperature being about 660O. Magnesium and calcium oxides start the decomposi- tion of the peroxide at 250° and 310° respectively whiIst zinc oxide brings about slow decomposition between 200° and 370° and at the same time forms barium zincate. Purely catalytic adio is shown by cadmium oxide lanthanum oxide and cerium peroxide whilst zirconium oxide and stannous and stannic oxide are without action.Alumina acts catalytically and forms an aluminate. Titanium oxide and barium peroxide in molecular proportions when heated below 300O. evolve oxygen slowly; above 300° a titanate is formed probably BaTiO,. When the proportions 2Ba02 to TiO are used a basic titanate completely soluble in acetic acid is formed. Litharge and barium peroxide between 300° and 400" 2*ii. 36 ABSTRACTS Oh’ CHEMICAL PAPERS. give no oxygen but form a brown product the nature of whicli has not been determined. Above 500° much oxygen is evolved and a product formed probably Ba,PbO which leaves a residue of lead peroxide when treated with nitric acid. Vanadium pent- oxide reacts vigorously with barium peroxide. When equimolecular proportions are used reaction begins a t 215O and is ended at 53OC barium metavanadate Ba(VO,) being formed.With 2Ba0 the metavanadate is first formed but a t 375O a second very vigorous reaction starts and the colour changes from brown to white the product formed being apparently Ba,V,O,. Tantalum pentoxide also reacts vigorously with formation of a tantalate. When arsenic trioxide is heated with barium peroxide (3 mols.) arsenic pentoxide is first formed at 310° t o 410° and above 4 6 5 O evolution of oxygen begins and barium arsenate is formed. Antimony tri- oxide behaves differently as a t 200° oxygen is evolved with almost explosive violence. Bismuth trioxide starts a gradual evolution of oxygen a t about 25OC and higher oxides of bismuth or compounds of theae with barium oxide appear to be formed.Chromium sesquioxide is rapidly oxidised above about 225O without evolution of oxygen barium chromate being formed. The oxides MOO WO U,O and UO all cause evolution of oxygen and form molybdates tungstates and uranates respectively. The lower oxides of manganese are all oxidised barium manganate being formed. Ferric oxide acts catalytically and a t the same time ferrate is formed. Nickel and cobalt oxides both act catalytically and are at the same time changed into higher oxides which how- ever do not agree in their properties with the known peroxides of these metals. E. H. R. Basic Exchange in Permutite. V. ROTHMUXD and G. KORNFELD (Zeitsch. anorg. Chem. 1919 108 215-225).-1n a previous paper (A. 1918 ii 315) basic exchange between univalent metals in permutite was studied. The inquiry has now been extended to bivalent metals with the study of the exchange between copper permutite and the nitrates of magnesium calcium strontium and barium and it has been confirmed that with equivalent ions the reaction is independent of the dilution.The tendency t o permutite formation is least with magnesium and increases steadily from calcium to barium. For the equilibrium between uni- and bi-valent ions the equation takes the form ( c 1 ’ ) 2 / c z l . (cp/c12)fl=K and this formula has been verified for the exchange between silver permutite and barium nitrate and for t h a t between calcium and sodium in sodium and calcium permutites. I n the latter case a gap was found in the series showing that calcium and sodium permutites are not completely miscible. The equilibrium for these cases is not independent of the dilution.A few experiments on the exchange between lanthanum and silver permutites showed that in this case the dilution has a marked influence on the equilibrium. Preparation of Colloidal Mercury by Cathodic Dis- integration. A. GUTBIER and G. L. WEISE (Kollozd Zeztsch. 1919 25 97-130).-Colloidal solutions of mercury may be prepared by E. H. R.INORGAKIC CHEMISTRP. ii. 37 striking an arc between a clean mercury surface and a thin platinum foil catlode under water. The most suitable current to employ for the purpose is 3 4 amperes at either 110 or 220 volts. By this means at. ordinary temperatures the sols are obtained which have a grey colour and change from light grey to bluish- grey and then to grey as the concentration increases.The sols are not very stable; on keeping they coagulate in about twenty-four hours and are coagulated by very small quantities of electrolytes. I n all cases the coagulation is irreversible. Much more stable sols are produced when a protecting colloid is present; thus using 1 10 or 1 50 gum arabic solution in water grey sols are produced which may be preserved for ten to fourteen days. These sols are also very sensitive to electrolytes and are irreversibly coagulated by the addition of alcohol. J. F. S. Scandium. R. J. MEYER and B. SCHWEIG (Zeztsch. anorg. Chem. 1919 108 303-317).-Jk was shown in a previous paper (A. 1914 ii 369) that the atomic weight of scandium given in the Inter- national tables cannot be accepted as a final figure. The impuri- ties likely to be present in scandium preparations separated by the usual methods from the other rare earths are traces of thorium and of the elements of the yttrium earths.It has now been found that complete separation of scandium from these impurities can be effected by fractional crystallisation of scandium formate or of ammonium scandium fluoride (NH,),ScF,. A sample of scandium oxide the spectrum of which showed no trace of yttrium lines and only faint traces of those of thorium and ytterbium was frac- tionated by the formate method and when the most soluble frac- tion was examined spectroscopically the lineis of all these elements showed up very strongly showing that the spectroscopic test for them is not so sensitive as is usually supposed.The least soluble fraction appeared to be pure scandium formate and the atomic weight of the metal determined by the sulphate method was 45.33 to 45.35. Ammonium scandium fluoride was crystallised from hot weakly ammoniacal aqueous solution. The impurities separated first from the hot solution and after these had been filtered off the solution was allowed to cool. After two or three such crystallisa- tions the spectrum of the double salt showed no signs of the presence of foreign elements. Atomic weight determinations on the pure materials thus prepared were made by Honigschmidt by analysis of scandium bromide the mean result of a series of experi- ments pointing to 45.1 as the true atomic weight of scandium (A.1919 ii 285-286). This is one whole unit above the accepted value. The present authors confirm Honigschmidt’s conclusion that an exact determination of the atomic weight cannot be made by the sulphate method on account of the impossibility of preparing a pure scandium sulphate free from both basic and acid salts. E. H. R. The Changes Undergone by certain Alloys of Aluminium. LBON GUILLET (Compt. rerid. 1919 169 1042-1043).-1t has 2*-2ii. 38 ABSTRACTS OF CHEMICAL PAPERS been shown previously (A. 1902 ii 364) that certain alloys 01 aluminium with either iron manganese or nickel rapidly crumble to powder in the air. These experiments have been repeated and this time the alloys of aluminium with iron o r nickel remained stable. The aluminium-manganese alloy containing 86.4% of man- ganeee crumbled very rapidly in the air but more slowly in oxygen nitrogen or hydrogen without gaining in weight.The change is to an allotropic modification. Similarly an alloy of aluminium and antimony fell to powder in moist air but not in dry air but in this case the change was chemical oxidation taking place. W. G. The Ternary System Aluminium-Copper-Zinc with Reference to Zinc in Particular. V. JARE; ( h i . Zeitsch. Metdlog. 10 1-44; from Chenz. Zenty. 1919 iii 512-514). -The literature of the binary systems Al-Zn A1-Cu and Cu-Zn is critically reviewed in addition to that of the ternary system with reference to copper. The author has chiefly investigated the portion of the system Al-Zn-CuAl-CuZn,. This can be resolved into four equilibria triangles to each of which a ternary non- variant pseudo-eutectic point belongs which (except in one doubtful case) lies outside the equilibrium triangle. A t 424O with 68% zinc 12% copper and 20% aluminium reaction is expressed by the scheme aluminium mixed crystals + fused mass A1,Zn3 + CuA1,; at 418O with 72% zinc 11.5% copper 16.5% aluminium the equation is CuA1 + fused mass Al,Zn3 + CuAl ; a t 406O with 79.5% zinc 9% copper and 11.5% aluminium equilibrium occurs as follows Al,Zn,+CuZn f fused mass; a t 386O with 94% zinc 1% copper 5% aluminium the reaction is CuZn,+fused mass S Al,Zn,+Zn mixed crystals; a t 590° with 76% zinc 23% copper 11% aluminium equilibrium occurs thus 6+fused mass CuAl+ CuZn,.Ternary compounds were not observed. The alloys were prepared by melting the requisite amounts of the pure metals in unglazed crucibles in an electric furnace with alloys containing 20 50 and 70% of copper.The temperatures were measured with platinum-platinum rhodium or with silver-nickel thermo-elements (calibrated by the m. p.’a bismuth 269O lead 327O zinc 419O 3b 630‘5O). The changes in the solid alloy resulting from the decomposition of the compound Al,Zn a t 256O were not investigated. The microscopic observations were made with sections etched with aqueous sodium hydroxide solution (2%) or alcoholic picric acid (2%). The mechanical properties of alloys of aluminium copper and zinc rich in the latter have been investigated; t o secure a near approach to the technical alloys commercial zinc containing about 1% of lead was used. The hardness of slowly cooled specimens in the system Zn-A1 increases rapidly with the aluminium content to about 70 a t 10% Al remains constant to about 20% Al and then increases very slowly; it reaches a maximum in the system Zn-Cu at about 5% Cu sinks to a minimum a t 12-14% Cu and subse- quently increases very rapidly.The minimum of the Zn-Cu alloyINORGANIC CHEMISTRY. ii. 39 is also observed in the ternary system Al-Cu-Zn in slowly cooled specimens but is absent from quickly cooled specimens ; the hard- ness increases considerably more uniformly. The tenacity of Cu-Zn alloys shows a maximum a t about 10% Cu whilst that of Zn-A1 alloys increases fairly rapidly with addition of copper to about S-lO% Cu and then sinks. Fractured surfaces of zinc become finer by addition of aluminium without altering the usual structure; addition of copper causes little alteration in the region of the zinc mixed crystals but subsequently the fracture is more finely granular.The compressibility of alloys of zinc 90%. Cu 6% and A1 4% has also been investigated. The tenacity and hardness of brass or ordinary bronze is readily attained in Al-Cu-Zn alloys but the brittleness of the latter make them an imperfect substitute for the former. The technical possibilities of A1-Cu-Zii alloys rich in the latter are discussed. The maximum useful copper content is estimated at about lo% the minimum a t about 4"/. Aluminium should not exceed 5% and should be diminished with increasing amounts of copper. The alloys are not suitable for steam; they oxidise too readily only withstand moderate pressure and are corroded by many salt solutions alkalis and natural waters.They can only be used with caution in making parts subject to considerable mechanical strain. They are frequently useful f o r fine mechanical The Ternary Systems MgO-A1,Oa-SiO and CaO-A120a- MgO. A. MEISSNBR (Zenzem! 8 296-298 308-310; from Chem. Zentr. 1919 iii 511) .-The temperahre-concentration relationships in the first system of the different crystalline phases in equilibrium with the liquid phases have been investigated and the results are expressed in diagrams and by a model. A ternary copzpound 2Mg0,A1203,Si0 unstable a t its m. p. and showing considerable tendency to the formation of mixed crystals has been observed in two forms ; the unstable p-variety crystallises from glasses a t a temperature of about 950° and passes a t a somewhat higher temperature into the stable a-variety.The properties of both forms more particularly of the a-form are similar to th3sn of the mineral cordierite. The system CaO-Al&MgO is QXK- paratively simple since it does not yield a ternary compound which is stable in the presence of the liquid phase. Its investigation therefore is concerned with the equilibrium of the components CaO A1,0 and MgO and of their binary compounds 3CaO,Al,O 5Ca0,3AI20 CaQ,A1,03 3Ca0,5A1,03 MgO,A1,0 in ternary solution. The results are expressed in a series of temperature- concentration diagrams. A new form of aluminium oxide. is designated P-AI,O,. The relationship of each of the forms in the binary system MgO-A1,0 demands the presence of solid solutions the amount of which has been estimated.Manganese-Bismuth Manganese-Zinc and Manganese- Silver Alloys. PAUL SIEBE (Zeitsch. anorg. Chem. 1919 108 161-183).-The condition diagram for manganese-bismuth alloys work such as physical instruments. H. w. H. W.ii. 40 ABSTRACTS OF CREWICAL PAPERS. has been determined by the cooling curve method. The two metals are not completely miscible. When more than 23% of manganese is present two layers are formed t h e upper consisting of practic- ally pure manganese whilst the lower is a 23% solution of man- ganese in bismuth. Five arrest points were found corresponding with five crystalline phases. The eutectic horizontal is at 259O and the other four horizontals corresponding with the arrest points are at 442O 59'i0 1043O and 1252O.The last represents a eutectic containing a t m.ost 0'5% bismuth which lowers the melting point of manganese 13O. From alloys containing up t o 0.5% of manganese pure bismuth first separates; with 0-5 t o 9% of manganese the first compound separates; from 9% t o 11% a second compound and from 11% t o 20% a third. The nature of the compounds has not been determined. Bet%ween 23% and 99.5% of manganese there is a complete gap. The microscopic appearance of the different crystalline phases is described. The hardness of the alloys increases with the manganese content. They are ferromagnetic and this property has been shown to be peculiar t o the first compound crystallising from alloys containing between 0.5% and 9% of manganese. Alloys of zinc and manganese could be prepared containing only up to 50% of manganese.Those containing from 0% t o 11% of manganese show an arrest point at 416O the eutectic point. The eutectic compound contains a very small undetermined quantity of manganese which lowers the melting point of zinc 3O. The mixed crystal constituent of the eutectic contains 11% of man- ganese and a series of mixed crystals are formed containing up to 50% of manganese. The alloys are brittle and increase in hard- ness as the manganese content increases. The behaviour of a series of manganese-silver alloys with a number of chemical reagents has been studied. A sharp distinc- tion was found between those containing 0.22 and 0.24 mol.of manganese and those containinc 0.26 and 0.28 mol. respectively. The latter pair were readily attacked by silver sulphate solution. with deposition of silver; by copper acetate and copper sulphate with deposition of copper ; and by sulphuric and hydrochloric acids with liheration of hydrogen whilst the former pair containing less manganese were unattacked. There was found a similar shai-1) difference when the electric potentials of the two pairs were com- pared. Those alloys containing more than 0.25 mol. of manganese appear t o behave both chemically and electrically practicallv as pure manganese. E H. R. I. BELLUCCI (Gazzetta 1919 49 ii 180-186).-The author has isolated the tervalent manganese fluoro-salt which causes the anomalous results obtained when Lunge's method of estimating nitrous acid by means of per- manganate is applied t o solutions containing fluorine ions in high (,oncentration (compare A 1919 ii 476).This salt KIMnFj,H,O. which belongs to the only known type of fluc-salts of tervaleiit :i:qnganese may be obtained by the action of nitrous acid in Fluoro-salts of Tervalent Manganese.INORGANIC CHEMISTRY. ii. 41 presence of hydroRuoric acid either on potassium permanganate or on a manganous salt the nitrous acid acting in the former case as a reducing agent and in the latter as an oxidising agent. T. H. P. The Atomic Weight of Iron. ALOIS BILECKI (Zeztsch. angig. Chent. 1919 108 318-320).-In a previous paper the atomic weights of silver and oxygen were compared through t h a t of man- ganese (this vol.ii 26). A similar comparison can be made through the atomic weight of iron. Baxter by analysis of ferrous bromide found Fe=55.8378 when Ag= 107.88 o r Fe=55.90 when Ag=108'00. By analysis of ferric oxide Richards and Baxter obtained in two series of experiments 55.90 and 55.883 taking 0 - 1 6 . A careful comparison of the data of the individual ex- periments establishes the fact t.hat the ratio of silver t o oxygen must be 108 16 or 107.88 15.9822 thus confirming the conclusion which was arrived a t from a consideration of the atomic weight of manganese. E. H. R. The Penetration of Iron by Hydrogen. T. S. FULLER (Y'tuu\. &lnter. Electrochem. Soc. 36 16 pp.).-Penetration of iron by molecular hydrogen begins a t 325O and increases rapidly Gith the temperature but nascent or atomic hydrogen penetrates iron at the ordinary temperature as shown by these experiments.An iron tube with 1/16-in. walls connected a t the ton to a U-tube device f o r measuring the volume of gas entering the tube through its walls. was immersed in a 1% solution of sulphuric acid. Hydrogen generated on the outside of the tube passed through the walls the rate varying with the conditions. The rate was greater for a unit immersed without electrical connections than when the unit was used as a cathode but in the latter case the greater the current the greater is the penetration. The rate also increases with temperature. Copper is not penetrated by nascent hydrogen but a coating of tin on the iron increases the rate.Results are given for other conditions and electrolytes. It was proved t h a t there was no peiietratiou by acid but by gas alone. CHEMICAL ABSTRACTS. Tin Hydride. I. FRITZ PANETH arid KARL FijRTH (Bey. 1919 52 [HI 2020-2029).-The authors have applied the method which was successfully used in the study of bismuth hydride (Paneth and Winternitz A. 1919 ii 68) to the problem of the isolation of a hydride of tin. An alloy of tin and magnesium corresponding with the formula Mg,Sn is dissolved in 4A7-hydrochloric or sulphuric acid ; the gas which is evolved is filt,ered through cotton- wool dried by calcium chloride and phosphoric anhydride and led through a heated hard-glass t.ube; a mirror of metallic tin is formed in close proximity to the hot part of the tube which when well developed consists of three zones a whitish-grey blackish-grey to black and finally brown in colour. The reactions of the mirrorii.42 ABSTRACTS OF CHEMICAL PAPERS. are very fully described the most distinctive of them being the insolubility in cold concentrated nitric acid and the formation of purple of Cassius and calomel by gold and mercuric chlorides respectively after preliminary treatment of the ring with dry gaseous hydrogen chloride. Preliminary experiments also show t h a t the gas can be condensed by liquid air and re-evaporated with- out decomposition so t h a t its collection in some quantity appears possible. The yields however are as poor as in the case of bismuth hydride oiily a few thousandths per cent. of the tin of the alloy being converted into the hydride.[With A. M ~ R s C ~ ~ ~ ~ . ] - P r e l i m i n a r y experiment appears t o show that lead hydride can exist in the gaseous state. It is formed from the magnesium-lead alloy in yield which is worse than in the cases of bismuth or tin but better results are obtained by an electrolytic process of reduction. H. W. The Question of the Existence of Zirconium Monoxide. ROBERT SCHITVARZ and HCGO DEISLER (Ber. 1919 52 [n]. 1896-1903).-Winlrler ( A 1890. 1375) Dennis and Spencer ( A 1.896 ii 558) and Wedekind (A. 1905 ii 596) have studied the reduction of zirconium dioxide by magnesium and obtained black powders which recombined with oxygen on heating in the air. the increase in weight approximating to t h a t required by the formula ZrO.The experiments have been repeated under various condi- tions (in a Rose crucible in a combustion tube in an atmosphere of carbon dioxide in porcelain tubes a t 1000O filled with carbon dioxide or hydrogen or evacuated and in a crucible fired by thermite) but i t is found that it is quite accidental t h a t the black powder left after removing the excess of magnesium and its oxide should increase by the theoretical quantity (nearly 155L) on heating in oxygen. I n fact the more quickly the reduction is performed the greater is this subsequent absorption of oxygen rising in one case t o 28%. It appears therefore that the black powder is a mixture of metallic zirconium and its dioxide To test this assumption varioiiq specimens of the powder were heated at about 250° in a stream of chlorine or a t high temperature. in hydrogen chloride gas the sublimate being analpsed.It was expected that the free inetal would give the tetrachloride the diovide remain unattacked and the monoxide. if any produce water in the case of hydrogen chloride. No indication whatever could be obtained of the presence of the monoxide. The sublimate consisted of the tetrachloride. and its weight aqreed exactly with the increase of weight suffered by the powder when heated in oxygen. The dioxide iq not attacked at all. even at 500O. J. C. W. F. P. VENABLE and I. W. SMITHEY ( J . .4mer. Chem. S o r . 1919 41. 1722-1 7 2 7 ) .-Zirconium hydroxide does not dicsolve in iodic acid. and therefore zirconium iodate waq prepared by adding iodic acid solution to a solution of zirconyl chloride.when a white precipitate Zirconyl Compounds with the Oxy-halogen Acids.INORGANIC CHEMISTRY. ii. 43 insoluble in water alcohol and ether was obtained. It was ileconi- posed by hydrochloric acid with evolution of chlorine and a t 125' iodine was evolved. The precipitate was filtered without any wash- ing and other samples were filtered and washed with measured volumes of water with the object of determining the amount of hydrolysis. The unwashed precipitate had the composition ZrO(OH)2,2ZrO(I03) when obtained from concentrated solutions but from dilute solutions the salt is more basic and has the com- position 5Zr0(OH),,8ZrO(I03)~. When the second compound was washed with 6 litres of water a t the ordinary temperature the compound had the composition 3ZrO(OH),,4ZrO(I03),.On wash- ing with 20 litres of boiling water the compound 2ZrO(OH),,ZrO(TO,) was obtained and with 30 litres of boiling water the compound 3ZrO(OH),,ZrO(I0,)2 was produced. Zirconium hydroxide is readily soluble in 30% perchloric acid. When the action was allowed t o take place a t the ordinary temperatures the compound ZrO(C10,)2,HC10 crystallised out in large triclinic crystals but when perchloric acid was heated with an excess of zirconium hydr- oxide the compound Zr0(OH),,9ZrO(C104) was obtained in crystal clusters which were soluble in alcohol ether benzene chloroform or carbon tetrachloride but could not be recrystallised from these solvents. On adding potassium chlorate to a cold con- centrated solution of zircoiiyl perchlorate crystals of basic zirconyl chlorate Zr0(OH),,3ZrO(C103) were obtained.The crystals were extremely deliquescent very soluble in alcohol but insoluble in ether. They were faintly yellow in colour had an odour of chlorine dioxide and oxidised organic matter very readily J. F S. Elimination of Vanadium from the Arsenical Waters of the Bellville District Province of Cbrdoba Argentine. FREDEKICO REICHERT and RAUL WERNICEE (Anal. SOC. Quim. Aiyentina 1919 7 ll0-113).-Solutions of ferric salts and of ferrous sulphate added in small quantit,ies to alkaline solutions of sodium vanadate cause the precipitation of the vanadium as i n - soluble ferric vanadate and also by adsorption on the ferric hydr- oxide formed. It is proposed t o apply the method for the simultaneous removal of arsenic and vanadium from natural waters.w. s. M. [Preparation of Potassium Pyroantimonate for the Esti- mation of Sodium.] J. D. VAN LEEUWEN (Chem. Weekblad 1919 16 1426) .-Potassium pyroantimonate which will remain un- affected for a prolonged period is made as follows 20 grams of potassium antimony1 tartrate are mixed with an equal quantity of potassium nitrate and the mixture is heated to redness in a crucible. When reaction has ceased the lid is placed on the crucible and heating continued for fifteen minutes. After cool- ing 50 C . C . of warm water are added and the mass stirred until it 2" *ii. 44 d\RSTRACTS OF CHEMICAL PAPERS. liecomes powdery. It is then filtered through a porcelain vdciiuiii filter the residue is treated with 100 C.C.of cold water and trans- ferred with the filter paper to a flask containing 500 C.C. of boil- ing water. After boiling for one minute and rapidly cooling some aluminium hydroxide is added the mixture is shaken and finally passed through an ordinary filter paper. W. J. W. Bismuth Thiosulphate Compounds. L. VANINO and F. MUSSGNCG ( A ~ c h . Pharm. 1919 257 264-266).-The authors have prepared various double thiosulphates of bismuth by making use of bismuth-mannit,ol solution. Hauser ( A . 1903 ii 48i) describes an unstable bismuth sodium thiosulphate but a stable compound BiNa,(S,O,) which crystallises in small octahedra is obtainable by the interaction of bismuth-mannitol solution and sodium thiosulphate in presence of manganese chloride. Bism21th ammoti,ium thiosuzphate Bi(NH,),(S20,),,2H,O forms a yellow precipit,ate which soon changes colour and 011 solution in water undeqoes decomposition.Rifirn~ith .strot~.tiirnz thiosii7pIiate prepared from bismuth-mannitol solution and strontium thio- sulphate forms a yellow indistinctly crystalline mass and under- goes hydrolysis when dissolved in water. The bismuth-silver com- pound similarly obtained forms a yellow precipitate becoming black in a few seconds; thO bismuth-copper compound is pre- cipit.ated only on addition of alcohol. Unsuccessful attempts were made to prepare bismuth dithionate and trit.hionate. On addition of sodium trithionate (compare Willstatter A. 1903 ii 543) to bismuth-mannit.01 solution. the liquid soon blackens; as this behaviour is not shown by sodium dithionate the presence of bivalent sulphur in trithionic acid is proved.T. H. P. s r 3 p (S203)312! Platotriammine Sulphite. D. STRdarHoLtr (Zeitsch. cmorq. Chem. 1919 108 211-214).-When the ammonium salt. of tri- chloroplatosulphonic acid is warmed in aqueous solution with excess of ammonia platotriammine sulphite Pt(NH,),S03 is formed as a sparingly soluble white microcrystalline powder. It. seems t o he identical with the substance which was described by Birnbaum (dnnalen 1869 76 142) as platotetrammine sulphit'e. When its hydrochloric acid solution is oxidised with chromic acid the diammine Pt(NH,),Cl is formed. Solubility determinatioiis showed t h a t only t.he tmns-isomeride was formed. This conclusion was confirmed by the preparation of the osalate which is quite distinct from that given by the' cis-compound .E. E. R . Platinitetrammine Disulphite. D. STR~MHOLM (Zeitsch. anorg. Ch,em. 1919 108 184-190).-To determine d.eSnitely whether the compound prepared by Cleve by the action of sulphurous acid on Gros's nitrate PtC1,(NH3)4(N03) contains bi- or quadri-valent platinum a study of its formation reduction,MINERALOGICAL CHEMISTRY. ii. -15 and oxidation lias lieen made with the result that the compound has been proved to have the composition Pt(NH,),(SO,),,SH,O. The yield of the disulphite obtained from the nitrate is only about 6676 of the theoretical and a certain amount of sulphuric acid is formed during the reaction but not so much as would be required t o be formed if the platinum were reduced from the quadrivalent to the bivalent condition. When the compound is oxidised by means of chromic acid the oxygen used is all accounted for by the oxidation of the sulphurous acid t o sulphuric acid t h a t is t o say there is no oxidation of the platinum. The product of the chromic acid oxidation is a chromi-chromute of the formula By prolonged heating with dilute sulphuric acid this compound is transformed into the salt [(OH)(SO,)Pt(NH,),]Cr,0 previously prepared by Cleve. The same two chromates were prepared from Rajevski’s nitrate [(OH)ClPt(NH,),](NO,),. When the platinitetrammine disulphite is boiled with water or heated with acids platinotetrammine sulphate is formed the quadri- valent platinum being reduced by half the sulphurous acid which becomes oxidised t o sulphuric acid whilst the other half of the sulphurous acid escapes. The same reaction takes place incom- pletely when the salt itself is heated. [(OH)~S0,P~(NH,)s11pCr,07,Cr0,. E. H. R.