ii. 140 ABSTRACTS OF CHEMICAL PAPERS. Inorganic Chemistry. Separation of the Element Chlorine into Isotopes (Isotopic Elements). The Heavy Fraction from the Diffusion. WILLIAM D. HARKINS and ANSON HAYES (J. Amer. Chem. Soc. 1921 43 1803-1825).-By means of elaborate diffusion a definite increase in thc atomic weight of chlorine has been obtained. The increase in atomic weight amounts in different experiments to from slightly less than to considerably more than one part in a thousand (1 in 645). The diffusion experiments were effected with hydrogen chloride and a considerable amount of the isotopic acid has been obtained. The separation of isotopes by diffusion is discussed in terms of the Rayleigh diffusion equation which applies to the diffusion into a vacuum. It is shown that for a high efficiency the pressure of the gas on both sides of the diffusion wall should be low first to secure good mixing and secondly to insure that the passage through the porous partition shall be entirely molecular.A rapid and precise method for the determination of the atomic weight of isotopic chlorine has been devised. It is found that in the separation of isotopcs the percentages of the different isotopes present are as important factors as the atomic weight differences. Thus contrary to what has been supposed it is shown even aside from the greater difficulties involved in obtaining and handling neon that it is easier to produce a small increase in the atomic weight of chlorine (of the magnitude of 0.05 unit) than to produce the same increase in the atomic weight of neon.Viscosities of the Hydrogen Haloids. H. HARLE (Proc. Roy. Xoc. 1929 [A] 100 429-440).-The viscosity of hydrogen chloride bromide and iodide has been determined by the method of transpiration through a capillary tube. Measurements were made a t temperatures in the neighbourhood of 20" and lOO" and from the results the viscosity is calculated for 0" and 100" rc- spectively. The following values in C.G.X. units are recorded hydrogen chloride yo= 1.332 x ; vlo0= 1.837 x lo4 ; hydrogen bromide yo= 1.710 x ; 17roo=2*365 x lo-* ; hydrogen iodide qo= 1.731 x The values of C the Suther- land const.ant are 357 375 and 390 for the three gases respectively. J. I?. S. q,00=2*403 x low4. J. F. S. Numerical Revision of the Data referring t o the Density of Gaseous Hydrogen Bromide Atomic Weight of Bromine.E. MOLES ( J . Chim. Physique 1921 19 135-138).-The molecular weight of gaseous hydrogen bromide with reference to oxygen has been calculated to be 80.944 and from this the atomic weight of bromine is 79.936. If however the value for the weight of the normal litre of oxygen recently published (cf. following abstract) is used the atomic weight of bromine becomes 79.927. J. F. S.INORGANIC CRER.IISTRY. ii. 141 Critical Study of the Modern Value of the Density of Gaseous Oxygen. E. MOLES ( J . Chim. Physique 1921 19 100-120).-A discussion of the modern results for the density of oxygen leads t o the value L,=1~42891&0~00003 as the most prob- able value for the weight of the normal litre of oxygen.The value is deduced from 162 measurements made by nine workers using oxygen obtained from three different sources and by three essentially different methods. If only the most recent measurements by weighing a globe of the gas are considered the mean is Lo=1.42894. Up to this the value accepted was 1.42905 which can no longer be maintained and although the difference is only 1 in 10,000 yet it cannot be neglected in the calculation of molecular weights. It is proposed that the value 1.42891 be accepted as the standard and this in some cases may be rounded to 1.4289. If this value is calculated to the conventional value of 9 y=980%65 C.G.S. units then the value of LON becomes 1.42897 which may be rounded to 1.4290. J. I?. S. Solubility. VII. Solubility Relations of Rhombic Sulphur.JOEL H. HILDEBRAND and CLARENCE A. JENKS ( J . Amer. Chem. SOC. 1921 43 2172-2177).-The solubility of rhombic sulphur has been determined in carbon tetrachloride benzene toluene nz-xylene heptane and ethylene dichloride a t 0" 25" 35" 45" and 54". The results are expressed in molecules of S per 100 molecules of solvent and in grams per 100 grams of solvent. The following values in grams per 100 grams of solvent are recorded Carbon tetrachloride 0" 0.339 ; 25" 0.831 ; 35" 1.155; 45" 1.564; and 54" 2.005; heptane 0" 0.124; 25" 0.362; 35" 0.512; 45" 0.698; and 54" 0.926; toluene 0" 0.897; 25" 2.018; 35" 2.722; 45" 3.620; and 54" 4.85; m-xylene 25" 1.969; 45" 3.604; benzene 25" 2.074; 54" 5.P65; ethylene di- chloride 25" 0.826; 40" 1,380; 79" 5.43; and 97*5" 9.97.It is shown that with the exception of certain minor discrepancies the solubilities of sulphur accord well with the internal pressure relations of the substances involved. J. P. S. The Constitution of Selenium. H. PBLABON (Compt. rend. 1921 173 1466-1468).-The grey selenium previously described (cf. A. 1921 ii 533) the specific resistance of which may vary from a few ohms to several million ohms is now shown to consist of two modifications. The a-modification has a very high resistance and is prepared by heating selenium a t a temperature only just above its melting point and then allowing it to cool slowly. The specific resistance of this modification diminishes a t first very rapidly with rise in temperature and then more slowly. At 200" the resistance is about 70,000 ohms and at 218" the substance melts the melting being accompanied by a sudden rise in resist'ance. The p-modification has a very small specific resistance.It is obtained by maintaining the molten selenium at a temperature near its boiling point for some time and then allowing it to cool. The p-modification is not stable a t low temperatures and it isii. 142 ABSTRACTS OF CHEMICAL PAPERS. readily changed into the a-form by oscillations of its temperature between 15" and 200". Influence of Freezing on Colloidal Selenium. 11. A. GUTBIER FR. HEINRICH and J. HUBER (KoZEoid Z. 1921 29 287-293; cf. A 1921 ii 693).-A continuation of work previously described (Eoc. cit.) on the effect of freezing on selenium sols. Three new series of experiments are now described in which it is shown that the sensitiveness of undialysed selenium sols prepared by means of sulphur dioxide towards cold increa'ses with the time during which they are frozen. The system may be frozen for short periods and will then melt to form a typical colloid but more pro- longed freezing increases the tendency to coagulate to irreversible gels on melting and very prolonged freezing will even cause coagul- ation to take placc in the solid.I n this respect there is no funda- mental difference between dialysed and undialysed sols but the dialysed system is somewhat more sensitive. The reducing action of sulphur dioxide on solutions of selenious acid is retarded by freezing the mixture. If the mass is kept frozen for a long time the colloid is precipitated in the ice and thereby the reduction process is accelerated.The precipitation appears to exercise it nucleus action by which the reduction in the solid mass is acceler- ated so that eventually more selenium is precipitated in the frozen mass than would be produced in a solution in the same time. In general the coagulation in the frozen sol commences where the sol is in contact with the air as is shown by the formation of a coloured ring on the upper surface of the ice. When test-tubes are used for the experiments the curved lower portion of the tube always shows an increased concentration of selenium. The form of the vessel in which the freezing takes place is shown to exert an influence on the coagulation. When a sol is poured on ice and frozen it is found that the disperse phase diffuses into the ice.It has been W. G. found scarcely possible to" freeze selenium sols in capillary tubes. J. I?. S. Metallurgy of Tellurium by the Wet Way. PIERRE HULOT (BUZZ. SOC. chim. 1921 [ivl 29 107Q-l071).-The method of reduction of potassium anhydrotellurate by zinc and hydrochloric acid with the precipitation of metallic tellurium (cf. A. 1920 ii 174) is modified by using aluminium and sodium or potassium hydroxide for the production of nascent hydrogen with the result that the reduction is completed in one hour instead of ten. The aluminium must be pure and free from copper. Manganese in the Catalytic Oxidation of Ammonia. CHARLES SNOWDEN PIGGOT (J. Amer. Chern. Soc. 1921,43 2 0 3 G 2045).-The action of manganese dioxide alone and when mixed with copper oxide or silver oxide and various alloys of manganese copper silver iron and silica as catalysts in the oxidation of ammonia have been investigated.A mixture of manganese dioxide with 40% of copper oxide a t 800" is shown to have an efficiency of more than 90%. All the substances mentioned catalyae the W. G.INORGANIC CHEMISTRY. ii. 143 oxidation of ammonia with various efficiencies which with suitable conditions are about 50%. The physical structure of the oxide catalyst is of great importance extremely fine subdivision giving a large adsorbing surface with very small pores being conducive to greater activity. J. F. S. Attempt to Prepare Nitro-nitrogen Trichloride. 11. Behaviour of Mixtures of Nitrogen and Chlorine in a Flaming Arc.WILLIAM ALBERT NOYES (J. Arner. Chem. SOC. 1921 43 1774-1782; cf. A. 1913 ii 584).-Attempts to prepare nitro- nitrogen trichloride or nitrogen trichloride by passing mixtures of nitrogen and chlorine through a flaming arc were entirely un- successful. Passing the same gases through a powerful ozoniser had the same result and no action at all occurred with Strutt's active nitrogen. The results although negative confirm Lewis's hypothesis that electrons are held jointly by two atoms rather than the view that electrons are transferred from one atom to another when atoms combine. J. F. S. Vapour Pressures of Aqueous Solutions of Nitric Acid. WILLIAM C. SPROESSER and GUY B. TAYLOR (J. Amer. Chern. SOC. 1921 43 1782-1787).-The total and partial vapour pressures of nitric acid solutions of the concentrations 20% 40% 56% 68% and SOY0 by weight have been measured a t temperatures 0" 35" 50" 65" and 80".From the data the vapour pressure for each 10% increase in concentration and each 10" rise in temperature has been calculated and tables drawn up. Attempts to measure the vapour pressure of 90% acid were unsuccessful owing to the decomposition of the acid at all temperatures except 0". J. F. S. Reducing Actions of Arsenious Acid. MORITZ KOHN (Monatsh. 1921 42 221-226) .-When copper sulphate is heated with ammonia solution and arsenious anhydride in a sealed tube in a boiling water-bath it undergoes reduction to cuprous salt with formation of arsenic acid. After the resulting cuprous solution has been oxidised to the cupric condition by atmospheric oxygen estimation of the arsenic acid reveals more of the latter than corre- sponds with the equation 2Cu"+As0,"'+20H'=H,O+ 2Cu'+ As04"'.Evidently oxidation of the ammoniacal cuprous solution to the cupric stage activates the atmospheric oxygen for the oxidation of the residual unchanged arsenious acid. Depression of the con- centration of hydroxyl ions by addition of ammonium salts retards the reduction of cupric to cuprous salt. Catalysis in the Interaction of Carbon with Steam and with Carbon Dioxide. HUGH STOTT TAYLOR and HARVEY A. NEVILLE ( J . Amer. Chem. SOC. 1921 43 205&2071).-The effect of potassium carbonate sodium carbonate lithium carbonate barium carbonate calcium carbonate sodium chloride ferric oxide copper sodium silicate borax and nickel as catalysts on the interrtc- tion of steam on carbon has been investigated using various forms of T.H. 9.ii. 144 ABSTRACTS OF CREMICAL PAPERS. carbon. It has been discovered that the interaction of carbon and carbon dioxide is likewise accelerated by the presence of the same materials and a striking parallelism between the catalysts for the two reactions has been shown. The probable mechanism of the steam-carbon reaction has been investigated. It has been shown that good catalysts for the water-gas reaction for example iron oxide are ineffective in the steam-carbon and carbon dioxide- carbon reactions. Catalysis of the water-gas reaction thus being excluded from consideration it has been shown that the acceleration of the reaction C+C02=2C0 and therefore the acceleration of the reaction C+ 2H,O = C0,j- 2€I may be ascribed to increased ad- sorption of carbon dioxide by the carbon surfaces in presence of active catalytic agents.Adsorption measurements confirm this view. The results have been considered in connexion with the suggested existence of a surface complex C,O and have been found to agree witch this conception. J. F. S. A Modification of Silicon Soluble in Hydrofluoric Acid. WILHELM MANCHOT (Bey. 1921 54 [B] 3107-3111).-Moissan and Siemens (A. 1904 ii 560) have isolated specimens of silicon from a silver regulus which suffer loss of weight (up to 99%) when treated with hydrofluoric acid and subsequently Lebeau (A 1906 ii 168) has brought forward evidence to show that a similar modification is present in copper-silicon alloys rich in the latter.The substance however does not appear to have been investigated Considerable difficulty is experienced in preparing silicon which is completely resistant towards hydrofluoric acid ; the product obtained by Wohler's method requires frequently repeated treat- ment with the acid before a stable specimen is secured. It is shown that an appreciable gain in weight occurs when the latter is melted with silver in an electric furnace in an atmosphere of air oxygen or carbon monoxide but this is in itself insufficient to account for the subsequent solubility of the silicon. The latter however is only produced in the soluble condition when the regulus is suddenly cooled. The silver may be replaced by aluminium.It dissolves in hydrofluoric acid with evolution of hydrogen. It is remarkable that Moissan and Siemens describe their product as exactly resembling crystalline silicon whereas the author's specimens are dark to pale brown and amorphous; they leave a pale brown powder after being treated with hydrofluoric acid. fully . H. W. Organogels of Silicic Acid. B. S. NEUHAUSEN and W. A. PATRICK ( J . Amer. Chem. Soc. 1921 43 1844-1846).-Hydrogels of silicic acid have been repeatedly soaked in alcohol acetone and benzene for long periods dried in a vacuum of 5 mm. a t SO" 120° and 270" successively for periods of one hour and analysed. The dried products contain alcogel 4*23y0 water 3'90y0 alcohol ; acetone gel 4.77% water 3'75y0 acetone ; benzene gel 4.31y0 water 3.82% benzene.Hence it follows in opposition to Graham'sINORGANIC (3HEMISTRY. ii. 145 statement that the water of a hydrogel of silicic acid cannot be entirely replasced by a second solvent. It has been previously shown that heating silicic acid gels in a vacuum at 300" for six hours does not reduce the water content below 4%% a quantity which is very near the amount of water left in the organo-gel above and indicates that this water is very firmly combined in the gel. J. F. S. Relative Densities of Alkali Metal Amalgams and Mercury. 11. EDGAR C. BAIN and JAMES R. WITHROW ( J . Physical Chem. 1921 25 535-544; cf. A. 1916 ii 431).-The methods of for- mation and the relative densities of liquid and solid amalgams of sodium potassium ammonium and calcium have been inves- tigated.It is shown that solid amalgams float on the mother- liquor and do not sink as stated by Kerp (A. 1898 ii 516). The method used by Merp (Zoc. cit.) produces sodium and potassium amalgams which are like the amalgams produced by other methods inasmuch as the solid is lighter than the liquid. Oily and pasty amalgams are due to fine crystals incorporated in the liquid. By electrolysis of solutions of potassium chloride amalgams with concentrations up to 2.31:/ of potassium were obtained. No difficulties except those of collection and preservation were met with in the production of ammonium amalgams by the electrolysis of solutions of ammonium chloride but crystalline amalgams cannot be obtained in this way. Electrolysis of an acidified solution of calcium acetate is not a satisfactory method for the production of calcium amalgams but it does produce a dilute calcium amalgam. J.F. 8. The Production of Potassium Hydrogen Sulphate from Ammonium Hydrogen Sulphate and Potassium Sulphate. W. DOMINIK (Przemysl Chem. 1921 5 10-15 3 7 4 0 63-67).-- Reaction in the presence of steam superheated a t 200-400" takes place according to the equation M,S0,+NH4HS0,=2MHS04+ NH,; lack of steam results in the production of pyrosulphates. In the above reaction K=[MHS04]2/[M,S0,] . [NH,HSO,] increases with the temperature according to a logarithmic function. Theor- etical consideration is given to the use of a mixture of sodium and potassium sulphat,es in the reaction. CHEMICAL ABSTRACTS. Crystal Structures of Sodium Chlorate and Sodium Bromate.ROSCOE G. DICKENSON and ELBRIDGE A. GOODHUE ( J . Amer. Chem. Xoc. 1921 43 2045-2055).-A large amount of X-ray spectrometer data for sodium chlorate and bromate has been obtained and tabulated. In agreement with other observers it has been found that the nature of the crystal surface has a considerable effect on both the absolute and relative inten- sities of reflection. The ease with which reflections from one face may be mistaken for those from another and the necessary precautions to avoid this have been pointed out. It is shown that the atoms in sodium chlorate and sodium bromate are veryii. 146 ABSTRACTS OF CHEMICAL PAPERS. probably arranged with the symmetry of the Schonflies space group T4 all oxygen atoms being equivalent. Sets of positions of the atoms in the unit structure only slightly different for the chlorate and bromate have been suggested.J. F. S. Preparation and Study of the Rarer Alkali Bromates. Rubidium Bromata. HAROLD D. BUELL and C. R. MCCROSICY ( J . Amer. Chern. Xoc. 1921 43 2031-2034).-Rubidium bromate was prepared by treating the pure carbonate with a n excess of bromic acid and recrystallising the product (cf. -4. 1920 ii 688). The following figures are the solubilities in 100 grams of water a t various temperatures 25" 2.93; 30" 3.55; 35" 4.28 and 40" 5.08. The corresponding figures for cmium bromate are 25" 3.66; 30" 4.53; 35" 5.32. The following melting points are recorded potassium bromate 405" ; czesium bromate 420" ; rubidium bromate 430". Both rubidium and czesium bromates form small cube-like crystals which however belong to the hexagonal system ; they have a refractive index between 2.144 and 2.22.J. F. S. Vapour Pressure of some Salts. 11. H. VON WARTENBERC and H. SCHULZ (2. Elektrochem. 1921 27 568-573; cf. Albrecht and Wartenberg ibid. 162).-Using the method previously employed the authors have determined the vapour pressures of lithium chloride czesium chloride rubidium chloride lithium bromide cmium bromide rubidium bromide sodium fluoride potassium fluoride lithium fluoride czsium fluoride rubidium fluoride sodium iodide cmium iodide and rubidium iodide. The measurements were carried in most cases up to the boiling point of the salt in question. The following data are recorded pressures being in atmospheres lithium chloride b.p. 1382" m. p. 606" log p= -37200/4*57T+4*923 ; czsium chloride b. p. 1303" m. p. 626" logp= -37400/4.57T+5.190; rubidium chloride b. p. 1383" m. p. 717" logp= -37800/4*57T+4*998; lithium bromide b. p. 1310" m. p. 549" logp= -35600/4.57T+ 5.109; cssium bromide b. p. 1300° m. p. 627" logp= -36750/4*5717+5*113 ; rubidium bromide b. p. 1350" m. p. 681" log p= -36980/4.5711+4*964 ; sodium fluoride b. p. 1695" m. p. 988" log p= -56600/4*57T+6.299 ; potassium fluoride b. p. 1505" m. p. 846" logp= -41900/4*57T+5*138 ; lithium fluoride b. p. 1676" m. p. 842" logp= -55100/4-5713+6.190; czsium fluoride b. p. 1251" m. p. 684" logp= -34700/4.57T+ 4.982; rubidium fluoride b. p. 1410" m. p. 775" logp= -40000/4.57T+5*243 ; sodium iodide b. p. 1300" log p= -37000/4*57T+5.130 ; lithium iodide b.p. 1170" log p= -40300/4-57T+6*105 ; caesium iodide b. p. 1280" log p= -36600/4*575!'+5*165 ; and rubidium iodide b. p. 1305" log p= -37000/4*5711+5*148. The results show that the heats of for- mation of the haloids of potassium rubidium and czesium lie very close together whilst that of the lithium salts is much higher and that of the sodium salts much lower. J. F. S.INORGANIC CHEMISTRY. ii. 147 Catalytic Influence of Foreign Oxides on the Decomposition of Silver Oxide Mercuric Oxide and Barium Peroxide. JAMES KENDALL and FRANCIS J. FUCHS ( J . Amer. Chem. Xoc. 1921 43 2017-2031).-The effect of foreign oxides on the temperature and rate of decomposition of silver oxide mercuric oxide and barium peroxide under an oxygen pressure of one atmosphere has been experimentally investigated.In almost all the systems examined the added oxides (CuO IVTiiO Fe203 CeO SiO CrO,) induce a considerable change in the decomposition temperature. Most commonly there is a marked lowering in this point; thus quoting an extreme case an equirnolecular mixture of barium dioxide and cupric oxide has an oxygen equilib- rium pressure of 1 atm. at 322" a temperature which is approxi- mately 500" below the decomposition temperature of pure barium dioxide. I n a few systems a comparatively small rise in the decomposition temperature is indicated. In all cases however the rate of oxygen evolution is markedly increased. The decompo- sition temperatures recorded refer to true equilibrium conditions concordant values being obtained with rising and falling tempera- ture.The results are tentatively ascribed to the formation of unstable intermediate compounds between the two oxides present. The increased rates of decomposition may be referred to adsorption effects but the large temperature changes point strongly to the actual participation of the added oxide in the reaction. Direct evidence was given in certain mixtures of the production of stable complexes. J. F. S. Physical Chemistry of the Oxides of Lead. 111. Hydrated Lead Monoxide. SAMUEL GLASSTONE (T. 1922 121 58-66). Reactions in Fused Salt Media. I. Basic Lead Chromates. J. F. G. HICKS ( J . Physical Chem. 1921 25 545-560).-The reaction between lead monoxide and sodium chromate in fused sodium chloride and a 50% mixture of sodium and potassium nitrates has been investigated and the equilibrium diagram of the system PbO-PbCrO constructed.It is shown that reactions in fused salts can be brought about in such a way as to yield products analogous to those prepared from the same initial sub- stances in aqueous solution. The chief difference between these reactions and their analogues in aqueous solution lies in the smaller velocity of the former probably due to the relative insolubility of the reacting substances in the fluxes as compared with water. There is in several cases a reaction between the flux and the dis- solved substance similar to hydrolysis. This analogy would appear to indicate that these solvolytic reactions are ionic but the second phase of the reaction forming basic lead chromates points to a non-ionic reaction.It may be a purely molecular (additive) reaction resulting in the formation of compounds of higher orders. Such a conclusion explains satisfactorily the slow- ness of the reactions aside from the low solubility of the solutes in the fused salt medium. Whilst at least one basic lead chromate,ii. 148 ABSTRACTS OF CHEMICAL PAPERS. PbO,PbCrO may be prepared by alkaline hydrolysis it does not follow that the basic lead chromates formed in fused salt media are formed by an entirely analogous process. Two new basic chromates have been prepared by the present method which have not been obtained by the wet process; these have the formulze 2PbO,PbCrO and SPbO,PbCrO respectively. A fourth compound Pb0,2PbCr04 is readily prepared by the fusion process and has been stated to be formed by the wet process but the author has been unable to prepare it by this method.The existence of the four basic lead chromates named is confirmed by the equilibrium diagram of the system PbO-PbCrO,. Other basic lead chromates described in the literature are shown to be mixtures or solutions of the compounds named above in one another. Salts of ortho- chromic acid are shown not to exist; whilst the salt Pb,CrO may be the salt of the monohydrate of ordinary chromic acid it could equally well be a true basic salt or a compound of a higher order so far as the present work is concerned. At the temperature (225-800") the stable form of lead monoxide is yellow in colour. This combining with yellow lead chromate yields red compounds all of which point to compounds of a higher order.The red com- pound formed by alkaline hydrolysis of normal lead chromate could well be considered. a basic salt but the addition of lead oxide and lead chromate molecules to form a compound of higher order might as easily take place in aqueous solution as in the nitrate flux used to form a compound of very closely the same composition. The reddening of lead monoxide in the nitrate flux has been shown to be due to change in crystalline form and not as might be supposed to the formation of red lead. Reduction of Copper Oxide by Hydrogen. ROBERT N. PEASE and HUGH STOTT TAYLOR ( J . Amer. Chem. Soc. 1921 43 2179-2188).-An investigation of the characteristics of the reduction of cupric oxide by hydrogen and the effect on the reaction of adding metallic copper to the oxide and water vapour and oxygen to the hydrogen has been carried out.It is pointed out that the reaction is auto-catalytic copper being the auto-catalyst. The reaction appears to take place a t the copper-copper oxide interface. This is shown by the character of the reduction curve and the fact that the addition of metallic copper accelerates the reaction. It is shown that the presence of water vapour in the hydrogen markedly interferes with the formation of the original copper nuclei from which the reaction zone that is the copper- copper oxide interface spreads out ; it does not markedly affect the subsequent reaction a t the interface however. The presence of oxygen in the hydrogen strongly inhibits the reaction a t the interface but in all probability has no marked effect on the primary reaction that is the formation of the original copper nuclei.J. F. S. J. F. S. The Oxidising Properties of Sulphur Dioxide. 111. WILLIAM WARDLAW and FREDERICK WILLIAM Copper Chlorides. PINKARD (T. 1922,121 210-221).INORGANIC CHEMISTRY. ii. 149 Behaviour of Ammoniacal and Alkaline Copper Solutions. MORITZ KOHN (Monatsh. 1921 42 83-87).-When arsenic is heated in a sealed tube with ammoniacal copper solution the cupric compound undergoes reduction first to cuprous salt and subsequently to copper 3Cu**+As+30€I'=AsO,'''+3Cu'+3H' and 3Cu'+ As+3QH'= AsQ,"' + 3Cu+ 3H' ; the copper liberated unites with the excess of arsenic to form greyish-black copper arsenide.Exactly similar changes occur when antimony is heated with ammoniacal copper solutions containing tartaric acid the solid deposited being then reddish-black and containing the copper and the excess of antimony. With bismuth and ammoniacal copper solutions containing tartaric acid rapid action occurs but the reduction proceeds only as far as the cuprous compound. Alkaline solutions of copper salts such as Fehling's solution or an aqueous solution containing copper sulphate glycerol and potassium hydroxide are rapidly reduced by arsenic antimony or bismuth with separation of copper 2As( 2s b) + 3Cu" + 60H'== 2As0,( Sb 0,)"' + 6H'+3Cu or 2Bi+3Cu"= 2Bi'" +3Cu. T. H. P. Separation of the Isotopes of Mercury. J. N. BXONSTED and G. VON HEVESY (2. physikal. Chem.1921 99 189-206 and Phil. il!Iag. 1922 43 [vi] 31-49).-A partial separation of the isotopes of mercury has been achieved by two processes. (1) Evaporation method (ideal distillation) based on the difference in the velocities of evaporation OP the isotopes. The distillate was found to be richer and the remainder poorer in the lighter isotope than the original substance. (2) Effusion method. A fraction of the mercury vapour pcnetrates through narrow openings into a condensation chamber where the lighter isotope is found in a relatively larger amount than in ordinary mercury. The results of the experiments agree with the hypothesis according to which the evaporation as well as the effusion velocity of the isotopes is inversely proportional to the square root of their molecular weights ; they are further in conformity with Aston's results obtained by means of mass spectrographic observations.The partial separation achieved was proved by measurements of density. The density difference found between the heaviest and lightest mercury amounts to O-49q(3 corresponding wit,h a difference of 0.1 unit in the atomic weight of mercury. Physico-chemical Analysis of Aluminium Oxy-salts and Aluminium Oxide Sols. MONA ADOLF WOLFGANG PAULI [with FRANZ JANDRASCHITSCH] (Kolloid z. 1921 29 281-287; cf. A. 1017 ii 563; 1921 ii 700).-The composition and nature of aluminium oxy-chloride sols have been investigated by means of measurements of the concentration of hydrogen and chlorine ions the total chlorine concentration and the electrical conductivity. A number of transport determinations have also been made.It is shown that it is impossible to remove all the chlorine from the products of hydrolysis of aluminium chloride by washing. A quantity of chlorine which is greatly in excess of that contained J. F. S.5. 150 ABSTRACTS OF CHEMICAL PAPERS. in the hydrochloric acid required for peptisation always remains. On boiling well washed aluminium hydroxide with dilute hydro- chloric acid sols of various compositions are obtained up to a limiting composition represented by the formula [Al(OH),],,AlOCl. Of these two have been investigated. Aluminium oxy-dichloride AlCl,*OH behaves as a ternary electrolyte and on dilution is strongly dissociated the brocess being complete a t a dilution 0.00106N.This compound shows a remarkably small hydrolysis which amounts to 0.1% at 0*068N and in a 0.00106N solution is only 0.25%. The compound alumiiiyl monochloride Al( OH),Cl or AlOC1 behaves as a binary electrolyte. The conductivity data indicate that a complex ionisation OCCUI"~ in this case of the type Al( OH),ClIAlO indicating a compound in which one aluminium atom of the complex acts as the central atom of a negative complex whilst the other furnishes a stable univalent positive ion. The peptisation of aluminium hydroxide leads to a sol of the composi- tion 2[A1( OH),]Al( OM),Cl and this on dilution undergoes complex ionisation represented by the formula? (1) 10A1( OH),,4A1OC1,A1O~Cl; (2) 12Al(QH),,5AlQCl,AlO~Cl ; (3) 16A1(OH),,7Al0C1,A10~Cl. On the other hand no complex ionisation of the form Al(OH),IAlO has been observed in the case of aluminium hydroxide.J. F. S. Germmiem. I. Extraction from Germanium-bearing Zinc Oxide. Non-occurrence in Samarskite. L. M. DENNIS and JACOB PAPISH ( J . Amer. Chem. Xoc. 1921 43 2131-2144).- A method of extracting germanium residues obtained in the smelting of certain American zinc ores has been investigated and is described. The residue contains zinc oxide considerable quanti- ties of lead arsenic and cadmium and small quantities of indium tin and antimony in addition to the germanium. A kilogram of the crude oxide is placed in a 5-litre Pyrex flask which is fitted with a two-holed rubber stopper carrying a bent glass tube for connexion to a Liebig condenser and a second short glass tube for introduction of the acid. The condenser leads under the surface of water contained in a 4-litre bottle which acts as receiver.Two and a half litres of hydrochloric acid (d 1.18) are added to the ore and the flask is heated until 2 litres of distillate have been collected. The distillate at this point contains all the germanium and a great deal of the arsenic from the ore. The distillate is poured into a 15-litre bottle until 10 litres have been accumulated carefully acidified with sulphuric acid until it is 6N (this must be carefully done to prevent loss of germanium chloride) and treated with washed hydrogen sulphide. The precipitate is at first yellow due to arsenic but later becomes whiter due to the germanium. When precipitation is complete the bottle is stoppered and kept for twenty-four hours.The solution is filtered by suction and washed with 3N-sulphuric acid which has previously been saturated with hydrogen sulphide. The filtrate is kept for forty-eight hours when usually a small quantity of a white precipitate (corresponding with 2 mg. of germanium) separates the majority of the super- natant liquid is siphoned off and the small amount of liquid andINOBGTANIC CHEMTSTkEP. ii. 151 precipitate worked up with the filtrate from the next lot of pre- cipitate. At this point two different methods of procedure are possible. (1) The moist sulplzides are added to hot 50% sodium hydroxide until a small quantity remains undissolved; this is then just dis- solved by the addition of a little more sodium hydroxide.The solution is made strongly alkaline by the addition of 8 grams of solid sodium hydroxide and placed in a large Pyrex flask fitted with a rubber stopper carrying a delivery tube a separating funnel and a glass tube reaching almost to the bottom of the flask. The flask is connected to a condenser and receiver as before. Washed chlorine is passed in to oxidise the arsenic to the quinquevalent condition. When the solution is saturated with chlorine the rate of entry of the chlorine is reduced and concentrated hydrochloric acid is added in large excess from the funnel. The flask is heated until half the liquid has distilled. The germanium chloride passes over and is hydrolysed by the water in the receiver forming white hydrated germanium dioxide. Should oily drops form in the re- ceiver more water is added to reduce the acid concentration and so allow the hydrolysis to proceed.The receiver is now replaced by a second one the distilling flask is filled up again with con- centrated hydrochloric acid and the distillation continued as before. Most of the germanium chloride passes over in the first distillation but for a complete separation the distillation must be repeated several times. The hydrated oxide is filtered washed first with dilute sulphuric acid and then with water and dried a t 110". It is pure white and contains traces of sodium calcium and iron but no arsenic. The filtrates from the hydrated oxide are treated with hydrogen sulphide and the germanium sulphide is recovered. The impurities mentioned are removed by dissolving in a slight excess of sodium hydroxide saturating with chlorine and distilling with hydrochloric acid hydrolysing the distillate as before.(2) This process is generally superior to the former in its greater economy of reagents. The wet sulphides are washed with 3N- sulphuric acid until free from chlorine and dried a t 110". They are then roasted in shallow iron dishes a t temperatures not exceed- ing 500". The roasted material is dissolved in sodium hydroxide (50%) chlorinated and distilled with twice its weight. of concentrated hydrochloric acid. A repetition of the distillation removes the last trace of arsenic. Either process gives a very pure germanium dioxide ; the yield is better by the first but the second is more rapid and economical. Germanium in ores is estimated by grinding 20-100 grams of the finely ground dried and weighed ore into a paste with water and pouring into a solution of sodium hydroxide in a hard flask the proportions being 2 of ore 1 sodium hydroxide 5 of water.The flask is fitted with a delivery tube for leading in chlorine a fraction- ating column and a small tap funnel. It is connected with a Liebig's condenser which leads to two Erlenmeyer flasks in series containing water to the depths of 3 cm. and half full respectively. The receivers are cooled with ice. The distilling flask is surrounded This removes a great deal of the arsenic.ii. 152 ABSTRACTS OF CHEMICAL PAPER& by ice and the contents are saturated with chlorine; the solution is then neutralised by hydrochloric acid and an excess equal to twice the weight of the ore added.The ice is removed and a slow distillation in chlorine carried out until one-half the liquid in the flask has passed over. An equal volume of hydrochloric acid is added to the distilling flask and the distillation continued until its volume is again reduced by one-half. This is then repeated once more. The receivers are disconnected acidified with sulphuric acid to make the solutions 6N saturated with hydrogen sulphide and kept for twenty-four hours. In a successful experiment there will be no germanium in the second flask. The precipitate is brought on to an ashless paper and washed with 327-sulphuric acid saturated with hydrogen sulphide until free from chloride then washed with alcohol to remove the acid and dried.The bulk of the precipitate is placed in a porcelain crucible and dried moistened with 1 1 nitric acid and warmed to drive off all the liquid allowed to cool treated with concentrated nitric acid dried and ignited. The filter-paper is incinerated in a second crucible treated with concentrated nitric acid and ignited. The filtrate from the ger- manium sulphide is kept forty-eight hours and the small precipitate formed filtered and treated as above. The weight of the germanium dioxide in the crucibles represents the germanium in the ore. Using this method the amount of germanium in the material used for the extraction of germanium was found to be sample I 0-247% sample I1 0.19%. Samarskite has been analysed by this method and found not to contain germanium.ARNE WESTGREN and AXEL E. LINDH (2. p7ysikal. Chem. 1921 98 181-210).- Various types of iron and steel have been subjected to Rontgen ray spectrographic examination. It is shown that the results of Hull (Phys. Review 1917 9 84 10 661) that pure iron a t ordinary temperatures (a-iron) possesses a space-centred cubic lattice is in keeping with the present work. At 500-836" that is within the so-called p-iron region the atoms are groupcd in exactly the same way as in a-iron. Since in the authors' opinion allotropy and polymorpliy are synonymous @-iron can only be regarded as a particular modification of a-iron. I n austenite and in pure iron stable a t lOOO" the crystals possess a face-centred cubic lattice. This is also characteristic of y-iron and thereby a fundamental difference is established between a-iron and y-iron.In martensite the iron occurs in its a-modification. This is also the case with high speed tool steel which has been hardened a t 1275". Photo- micrographs of some of the preparations examined are included in E. P. POLUSHKIN (Iron and Steel Inst. Carnegie Xchol. Mem. 1920 10 129-150; cf. Rev. iietal. 1920,17 421).-Alloys of iron with uranium are pyrophoric this property varying directly with the content of uranium. Alloys containing carbon are decomposed by water but this effect is not J. F. S. Crystalline Structure of Iron and Steel. the paper. J. I?. x. Alloys of Iron and Uranium.INORGANIC CHEMISTRY. ii. 163 exhibited by silicon except when present in very large amount or by vanadium.A polished sample placed on the emulsion of a photographic plate leaves an impression which after development by the usual process represents the structure of the specimen. Iron-uranium alloys contain the uranium carbides UC U,C and U,C the compound Fe,C,U,C as well as the compounds V,C Pe6U and FeSi. Neither uranium nor any of its compounds already mentioned forms solid solutions with iron. Vanadium has a greater affinity for uranium than for carbon. Equilibria in the System Fe-C-8 the Equilibrium FeB-Martensite-~errQ~ Oxide-Gas. W. REINDERS and P. VAN GRONINGEN (Rec. trav. chirn. 1921 40 701-706).-A considerable number of determinations of points of univariant equilibrium for this system have been carried out. The conclusion is drawn that the transition temperature for Fes + Fe is 905"; this is in accord with previous results obtained by different methods. By extrapolation of the graphic results the quintuple point of the system is found a t 740" and 2300 mm.Researches on the Metallic Carbonyls. ROBERT LUDWIG MOND and ALBERT EDWARD WALLIS (T. 1922 121 29-32). The Action of Nitric Oxide on the Metallic Carboiiyls. ROBERT LUDWIG MOND and ALBERT EDWARD WALLIS (T. 1922 Expansion of Chromium and of Nickel-Chromium Alloys over a Wide Range of Temperatures. P. CHEVENARD ( C m p t . rend. 1992,174 109-112).-Between 0" and 100" the expansion of chromium is exactly reversible the curve showing no singular point. The true coefficient of expansion which is 6.8 x 10-6 a t O" increases rapidly with the temperature but the curve shows a slight concavity towards the increasing temperature.Nickel-chromium alloys containing up to 16% of chromium and from 0.5 to 2.5:/ of manganese were examined over the tempera- ture range 0" to 1000". The addition of chromium leads to a very rapid weakening of Che anomaly of dilatation of nickel and when the chromium content reaches 5yo the anomaly disappears. The addition of chromium to nickel affects the expansion of this metal very little a t the ordinary temperature but tends to increase i t a t higher temperatures ; this effect being probably due to the presence of the compound Ni,Cr3. J. W. MARDEN and M. N. RICH (Bur. Mines BUZZ. 1921 186 146 pp. ; cf. A. 1920 ii 547).-The work is divided into four parts dealing successively with an historical review of zir- conium minerals the salts of zirconium and the metal; experi- mental work on zirconium ; the furnaces used ; and a bibliography of zirconium and its compounds.Analytical methods are given for the estimation of zirconium in ferrozirconium steel alloys such as nickel-zirconium and a method of separation of titanium columbium tantalum and zirconium. The physical and chemical CHEMICAL ABSTRACTS. H. J. E. 121 32-33). W. G. Zirconium.ii. 154 ABSTRACTS O F CHEMICAL. PAPERS. properties of amorphous and coherent zirconium are described. The cupferron method is the only one effecting complete separation of zirconium from aluminium. Polymorphic Transformations of Anthony Trisulphide. SAMUEL WILSON and C. R. MCCROSKY ( J . Amer. Chem. SOC. 1921 43,2178-2179).-The rate of transformation of the red amorphous variety of antimony trisulphide into the black variety has been determined in the presence of N - 7N- 12N-hydrochloric acid 7N-phosphoric acid 7N-sulphuric acid glacial and N-acetic acids and a saturated ethereal solution of hydrochloric acid. Of these acid solutions only the aqueous solution of hydrochloric acid gave any appreciable transformation a t 18-22' even after two months. With the aqueous hydrochloric acid the time necessary for complete transformation was 0-5 day 1 day and 10.5 days for E N 7N and N respectively. The action is due to a solution of the red variety and a reprecipitation of the less soluble black variety. The effect of temperature was investigated with a 20% aqueous hydrochloric acid. Complete conversion a t 26.5" required forty- four hours ; a t 30" twenty-nine hours ; 35" sixteen hours ; 40' nine hours ; 68.5" sixty-two minutes and 75" thirty-two minutes. A 20% solution of hydrobromic acid gave no change after twenty hours a t 75". CHEMICAL ABSTRACTS. J. F. S.