INOR8GANIC CHEMISTRY.1. THE PERIODIC TABLE.IN the Mendelkef Centenary Lecture delivered before the ChemicalSociety by Lord Rutherford a striking tribute was paid to thogenius of Dimitri Ivanovitch Mendeldef, who was born on February7th, 1534. It was very fitting that the lecturer himself should havebeen intimately associated with those outst anding developmentswhich have given a new meaning to Mendeldef’s famous law of theelements.Since the Periodic Law was first enunciated, the periodicity ofthe elements has been fully confirmed, and although new discoverieshave from time to time demanded modification of the originalstatement of the law, yet the fundamental ideas still remainessentially as Mendeleef set them forth. It is well known thatinterest in MendelBef’s generalisation was not fully aroused untilthe discovery of gallium in 1875 and scandium in 1579.The factthat these elements possessed the properties predicted by Mendelhffor eka-aluminium and elra-boron was very impressive, and furtherdiscoveries of new elements made it clearer still that the PeriodicLaw was a fundamental truth. Some twelve years ago six elementsout of the possible ninety-two in the Periodic Table were stillunknown. Bohr’s theory of atomic structure led in 1923 t o theisolation from zirconium minerals of element 72, which was namedhafnium, after Hafnia, the ancient name of Copenhagen.2 I n1925 the elements 43 and 75 were found and named masurium andrhenium re~pectively,~ and in 1926 the discovery of the rare earth61 was announced.4 The last of the missing elements, 85 and 87,provisionally called eka-iodine and eka-czesium, were stated tohave been detected in 1931-32. Three years ago, therefore, itseemed that the phase of chemistry dealing with.the discoveryof new elements had come to an end. It cannot be said, bowever,that in every case these more recent discoveries have met withgeneral acceptance. I n particular, a number of writers havequestioned the significance of the evidence upon which the existenceof elements 61, 85, and 87 is based. In this Report an attemptis made to state the evidence for and against these discoveries SOAnn. Reports, 1923, 20, 45. J., 1934, 635; A., 1934, 713.Ibid., 1925, 22, 63. Ibid., 1926, 23, 61. ti Ibid., 1932, 29, 300CARTER AND WARDLAW : THE PERIODIC TABLE.139that the reader may judge whether criticism is justified or not.Reference is also made in the following pages to elements 43 and91, and some recent work on elements of atomic number higherthan 92 is discussed.Since the time of Moseley’s work, which showed dehitely thatan element should exist between neodymium (at. no. 60) andsamarium (at. no. 62), numerous papers have been published dealingwith element 61. In 1917 J. M. Eder deduced from an examinationof the arc spectrum of samarium that his material might containtraces of a new element, but when W. Prandtl and A. Grimm 7 in1924 carried out an exhaustive fractionation of the cerium earthsthey failed to obtain any indication of an element 61.Two yearslater, however, the search for the iiew element appeared to havemet with success, for J. A. Harris, L. P. Yntema, and B. S. Hopkiiisannounced that they had found weak optical absorption bandsfrom a material obtained by the fractionation of large quantitiesof neodymium and samarium salts, and that these bands must beattributed to element 61. Further, they stated that they hadobtained in the X-ray emission spectrum of their preparation aweak but perceptible line which they claimed was the La line ofelement 61. The mean value for this line agreed within 0-0004 8.with the value calculated from Siegbahn’s precision values, althoughthe differences ranged from + 0.0026 to - 0.0047 A. It was held,therefore, that element 61 had been discovered, and the name“illinium ” was proposed for it.Shortly afterwards, L. Rollaand L. Fernandes claimed priority of discovery, as they maintainedthat they had observed the absorption bands, ascribed to element61, before Hopkins and his collaborators. R. J. Meyer lo and R.Glocker l1 and their respective co-workers also believed that theyhad detected element 61 in the difficultly soluble bromate fractionsof rare earths, and proved its presence by three lines in the Kseries of X-ray lines. At this period one gains the impression thatseveral groups of investigators had simultaneously discovered themissing element. J. M. Cork, C. James, and H. C. FoggX2 evenstated that they had produced a sample of material containing1% of the new element which gave seven lines of the L series ofillinium.These findings, however, did not pass without criticism,6 Sitzungsber. Akad. Wiss. Wien, 1917, IIa, 125; A., 1917, ii, 185.7 2. Ccn.org. Chem., 1924,138, 283; A., 1924, ii, 615.8 J . Amer. Chem. SOC., 1926, 48, 1585, 1594; A., 1926, 810, 780.Q Gazzetta, 1926, 58, 435; A., 1926,1083; Atti R. Acmd. Lincei, 1926, [vi],10 R. J. Meyer, G. Schumacher, and A. Kotowski, Naturwiss., 1926,14, 771.11 U. Dehlinger, R. Glocker, and E. Kaupp, ibid., p. 772,12 Proc. Nat. Acad. Sci., 1926, 12, 696; A., 1927, 190.4,498; A., 1927, 190140 INORGANIC CHEMTSTRY.Especially noteworthy is a paper by W. Prandtl l3 in which adetailed examination i8 made of the evidence upon which Hopkinsand his co-workers based their discovery of element 61.Prandtlshows that the absorption bands mentioned by Hopkins as occurr-ing with neodymium salts in dilute solutions were intensified whena small quantity of neodymium nitrate was added to a large amountof samarium nitrate. Consequently, these bands did not arise froma new element lying between neodymium and samarium. More-over, he maintained that the X-ray investigation of the rare-earthmixture was carried out in an apparatus unsuitable for the purpose,and that the so-called La, and Lpl lines of element 61 were reallydue to small amounts of platinum, barium, and bromine. It isinteresting that later investigators have failed to obtain any evidencefor element 61. In 1931 S. Takvorian14 reported that no linesattributable to element 61 could be observed from an examinationof the rare earths from Indian monazite.(Frau) I. Noddack in1934 in an important communication l5 recorded the results of anextensive investigation a t the Phy si kalis c h Te chnischen- Reichs -anstalt on the problem of element 61. At the request of variousworkers, over a period of eight years, she and W. Noddack haveexamined fifteen preparations by X-ray spectroscopy for thepresence of element 61. I n none of these preparations was anytrace of this element found. Further, 100 kg. of rare earths fromvarious sources were worked up for fractions between neodymiumand samarium without giving any indication of the required element.Again, an investigation by X-ray methods of the rare-earth fractionsfrom the Auergesellschaft which had been obtained from very largequantities of monazite did not afford, after very exhaustive fraction-ation, any evidence of element 61.(Frau) I. Noddack expressesthe opinion that, i f one surveys the negative results from so manyinvestigations, one must conclude that there is not sufficient evidencein the literature since 1926 to say that this element has been dis-covered, and it must be admitted that this conclusion is a reasonableone.Apart, however, from the controversial question whether element61 has ever been detected, it is interesting to speculate why thiselement does not occur in quantity in rare-earth mixtures. (Frau) I.Noddack suggests that its absence may be due to its instability.Samarium (at.no. 62) is radioactive, according t o G. von Hevesy 1613 I;. angew. Chem., 1926, 39, 897; W. Prandtl and A. Grimm, ibdd., p.14 Cornpt. rend., 1931, lga, 1320; A,, 1931, 783.l5 2. angew. Chem., 1934, 47, 301; A., 1934, 853.16 G. von Hevesy and M. Pahl, Nature, 1932,130,846; A., 1933,4.1333; A., 1927, 9CAltTER AND WARDLAW THE PESIODIC TABLE. 141and others, and loses an a-particle. May it not be that element 61also is radioactive, losing a p-particle, whereby its concentrationin the lithosphere may become so small that it would no longer bedetectable by the methods so far employed? Or again, may notelement 61 have a marked tendency to exhibit bivalency, for bothsamarium (at. no. 62) and europium (at. no. 63) are bivalent insome of their compounds? To test this theory (Frau) I.Noddackhas explored the possibility that element 61 may be presentin minerals containing the alkaline earths, but so far, withoutsuccess.In the search for element 85, the idea that it may result from aradioactive decomposition has not been overlooked. It is obviousthat it may be produced by the loss of two a-particles from actinium(at. no. 89) or by the expulsion of a p-particle from polonium (at.no. 84), and with these ideas in mind G. von Hevesy and R. Hobbie 1'have worked up a kilogram of Katanga pitchblende. The combinedprocess of concentration and examination by X-ray spectroscopymakes it possible to detect less than leg g. of the element, but eventhis could not be found. It will be mentioned later that F.Allisonand his colleagues claim to have found elements 85 and 87.Numerous attempts have been made to discover the elusiveelement 87, and a wide range of methods has been employed inthe search. P. R. Gennet6 l8 has prepared an excellent summaryof the results obtained up to 1933. The positive results are fewand include the observations made by 5. Papish and E. Wainer l9in 1931. It will be recalled that these investigators preparedfrom 10 kg. of samarskite, rich in uranium, a caesium preparationwhich they stated showed the X-ray lines of element 87. Theyheated the mineral to 1000" in a stream of hydrogen chloride,collected the sublimate, and converted it into sulphates, whichwere purified and fractionally crystallised as alums.The leastsbluble fractions were examined in a Siegbahn apparatus of highdispersion. As an example of the results obtained, the Lal linefor element 87 calculated to be 0.8524 was found as 0.853 A. L. L.Barnes and R. C. Gibbs 2o then examined by positive-ray methodsa cssium alum prepared from the samarskite used by Papish andWainer. They obtained a result which confirmed the existence ofelement 87, but the iEtensity of the important lines was very feeble.So far, no other investigator has been able to obtain similar results.(Frau) I. and W. Noddack,15 using various czsium-containing17 2. anorg. Chem., 1932, 208, 107; A., 1932, 1073.18 Bull. SOC. chim., 1933, [ivj, 53, 140.19 J . Amer. C h m . SOC., 1931, 53, 3818; A., 1931, 1348.20 Phguicd Rev., 1932, [El, 40, 318; A*, 1933, 1223142 INORGANIC CHEMISTRY.minerals, have failed to isolate any preparation that gives theX-ray lines required for element 87, and K.T. BainbridgeFl employ-ing a Dempster spectrograph, has examined czesium chloride, madefrom pollucite and from lepidolite mica, with negative results.Element 87 has also been looked for as a secondary product fromthe disintegration of radioactive substances, but again withoutsuccess. Theoretically, it may arise from element 89 (meso-thorium-11) by loss of an a-particle, or from element 86 (radon)by emission of a p-ray. Experiments by G. von Hevesy22 and 0.Hahn and 0. Erbacher33 to test these possibilities have yieldednegative results. The only other positive result comes from theuse of a magneto-optic method, which is based on the researchesof 3.W. Beams and F. Allison on the Faraday effect.24 This effectis the well-known phenomenon that the plane of polarisation oflight is rotated by a liquid when this is placed in a magnetic field.It has generally been assumed that the Faraday effect does notlag behind the application of the magnetic field. J. W. Beamsand I?. Allison maintain, on the contrary, that there is a retardation,and their conclusions are based on this assumption. They haveutilised their method for the detection of cations in analyticalchemistry and also claim 25 to have found evidence of elements85 and 87 in minerals such as lepidolite and pollucite. P. R.Gennet6,18 in a critical examination of the value of this method,concludes that the results of Allison and his collaborators must beaccepted with the greatest reserve.It is not without interest tofind that last year J. Papish and A. C. Shuman 26 tested the apparatusused by Allison and reported unfavourably on it. Finally, it may bementioned that (Frau) I. Noddack l5 expresses the opinion that theresults obtained by the magneto-optic method are dehitely specula-tive and considers that the discovery of element 87 by Papish isnot yet substantiated. It must be concluded that the evidencefor the discovery of elements 85 and 87 is not very substantial,and that elements 61, 85, and 87 still afford ample scope for theinvestigator.No account of the Periodic Table would be complete withoutsome reference to element 43, named masurium.Althoughmasurium was discovered at the same time as rhenium, there is aremarkable difference in their subsequent history. Very littleis known even now about masurium, but the chemistry of rhenium31 Physical Rev., 1929, [ii], 34, 752; A., 1929, 1210.22 Ann. Reports, 1928, 25, 317.23 Physihl. Z., 1926, 27, 531; A . , 1926, 990.2* Phil. Mag., 1927, [vii], 3, 1199; A., 1927, 610.25 Ann. Reprta, 1932, 29, 300. 26 Science, 1934, 79, 297; A., 1934, 626CARTER AND WARDLAW : THE PERIODIC TABLE. 143is considerable and rapidly developing. W. Noddack, (Frl.) I.Tacke, and 0. Berg2’ estimated that the lithosphere containedof masurium and 10-12 of rhenium compared with 7 xof manganese and 10-2 of iron.Their later revised estimate sug-gested that the two elements masurium and rhenium were presentin equal amounts as 10-9. This assumption makes it still moreextraordinary that so little progress has been made with the studyof masurium. The German investigators detected masurium byX-ray spectroscopic analysis in platinum ores, columbite, sperrylith,gadolinite, and fergusonite, and later 28 in tantalite and possiblyin chrome iron ore, olivine, and pitchblende. They examined1800 minerals and 21 meteorites for the elements 43 and 75. Theevidence for the discovery depends on the identification of theK,,, K,%, and Kp, lines in the X-ray spectra. The wave-lengthsfound were 0.672, 0.675, 0.601, and the calculated values were0.6734, 0.6779, and 0.600 8.respectively. Although the observedvalues are not identical with the calculated, they may be regardedas strong evidence of the presence of masurium, since there are noother elements which could give these results. The most successfulattempt to obtain masurium has been made by W. and (Frau) I.Nodda~k,~~ who isolated a sulphide product containing 0.2-1 yoof masurium during an investigation of columbite. Although thiswas insufficient for an examination of the chemical properties ofmasurium, it served for the identification of the optical arc and sparkspectral lines. Certain observers have had evidence of the presenceof masurium in rhenium concentrates. For example, 5. Hey-rovsky and V. DolejrSelr 30 stated that, in their polarographic studiesof manganese salt solutions with the dropping-mercury cathode,they noted a wave at - 1.15 volts in the deposition-potentialcurves, which they ascribed to masurium.There seems no reasonto doubt the existence of element 43. J. G. F. Druce 31 has madethe interesting suggestion that some rhenium preparations may becontaminated with masurium, and that this may account for someof the discrepancies in the observations and deductions of differentinvestigators.The important additions to our knowledge of protoactiniumwhich have been made in recent years are the justification for thisspecial reference to element 91. It will be recalled that this element27 Naturwiss., 1925, 13, 567; Sitzungsber. preuss. Akad. Wiss. Berlin, 1926,28 Z. angew.Chem., 1925, 38, 1157; A., 1926, 112.29 Metallbiirse, 1926, 16, 2129.30 Chem. Listy, 1926, 20, 4; A., 1926, 258; Rec. trau. chim., 1927, 46, 248;al Science Prog~ea8, 1933, 27, 687.400 ; A., 1925, ii, 939.A , , 1927, 636144 INORGANIC CHEMISTRY.-the eka-tantalum of Mendel6ef-was discovered independentlyand almost simultaneously by 0. Hahn and L. Meitner in 1917 andby F. Soddy and J. A. Cranston in 1918. Not until 1927, however,was the element available in a weighable amount. In that yearA. von Grosse,32 in the Kaiser Wilhelm Institut fur Chemie, isolated2 mg. of pure Pa205, and in 1928,33 carrying out a large-scale pre-paration, he obtained 9 mg. of pure oxide. As protoactinium has ahalf-value period of about 22,000 years, it is obvious that the elementshould be available in much larger quantity if sufficient startingmaterial is used.In 1934 G. Graue and PI. Kading,3* working in0. Hahn’s laboratory, reported that by working up 5.5 tons ofJoachimsthal radium residues they had prepared pure K,PaF,containing 0.5 g. of the element, and A. von Grosse and M. S.A g r ~ s s , ~ ~ a t the same period, described their process whereby theyhad obtained 0.1 g. of oxide. It will be realised that the develop-ment of the process for the isolation of these compounds has furnishedvaluable information concerning the chemical behaviour of thiselement. The details of the method of extraction used by theseinvestigators have been published, but a working description forthe preparation of pure protoactinium compounds cannot begiven in a few sentences.It may, however, be of some interestt o outline, briefly, the principal stages in the process used by vonGrosse .36Natural uranium minerals contain 8 g. of protoactinium for every10 g. of radium, and during the process for radium extractionthe protoactinium accumulates in the residue, which provides abetter starting material than that found in nature. This finalresidue, the so-called Riickriickstande from Joachimsthal, has theaverage composition : SiO,, 60; Pez03, 22 ; PbO, 8 ; Al,O,, 5 ;MnO, 1 ; CaO, 0.6 ; MgO, 0.5%. It also contains small quantitiesof titanium (0~3%)~ zirconium (0.1 %), and hafnium, together withmany other elements. The average protoactinium content is300 mg. of Pa,05 per metric ton (a concentration of 1 : 3,000,000),whereas the richest pitchblendes contain only 200 mg.per ton.The plant process is carried out in three main stages. I n the firststage, treatment with hot hydrochloric acid (25%) extracts fromthe Riickruchtunde the iron, the more basic oxides, and most ofthe lead, leaving a protoactinium concentrate consisting chiefly82 Nature, 1927, 120, 621; A., 1927, 1120; Naturwies., 1927, 15, 766; A.,1928, 495; Ber., 1928, 61, [B], 233; A., 1928, 259.33 See 0. Hahn, Ber., 1935,68, [B], 478; A., 593.34 Angew. Chem., 1934, 4’7, 650; A., 1934, 1186.35 J. Amer. Chem. SOC., 1934, 56, 2200; A , , 1934, 1319.36 Ind, Eng. Chern., 1935, 27, 422CARTER AND WARDLAW : THE PERTODIC TABLE. 145of silica and small amounts of zirconium, titanium, and other lessbasic oxides.This concentrate is fused with sodium hydroxide,and the silica is converted into sodium silicate, which is extractedwith water. The residue, containing most of the protoactinium,is dissolved in acid, and silica (from the insoluble silicates) is pre-cipitated. Soluble zirconium salts and phosphoric acid are nowadded to this acid filtrate, and the protoactinium is precipitatedtogether with zirconium phosphate. The precipitated silica, whichcontains 70% of the protoactinium, is extracted with sodiumhydroxide solution ( 20y0), the residue dissolved in hydrochloricacid, and the protoactinium precipitated from acid solution inthe usual way with zirconium phosphate, ZrP,O,. A characteristicreaction of protoactinium is its complete coprecipitation withzirconium phosphate. This process of A.von Grosse was modifiedby G. Graue and H. K a d h ~ g , ~ ~ whereby important changes weremade in the sequence of the operations-fusion of the Ruckruck-stande with sodium hydroxide preceded the extraction with hydro-chloric acid. Graue and Kading discuss in detail the furthertreatment of the precipitated zirconium phosphate. It is reallya complicated problem of the separation of a tantalum-zirconium-protoactinium mixture. The separation of protoactinium fromthe carrier substance, zirconium phosphate, is effected by crystal-lisation of ZrOC1, from a hydrochloric acid solution, whereby theprotoactinium in the filtrate is enriched.The final purification iseffected by the alternate use of the zirconium and the tantalumreactions, and ultimately the protoactinium is isolated as K,PaF,which gives no X-ray spectrum of foreign elements. To obtain theoxide, Pa205, this complex fluoride is heated with concentratedsulphuric acid, ammonia is added to precipitate the hydroxide,and this is ignited to the oxide.The element has been isolated by A. von Grosse 37 by bombardingthe oxide on a copper target with a stream of electrons in a, highvacuum. The use of 35,000 volt-electrons for a few hours at acurrent strength of 5-10 milliamps. splits the oxide into oxygenand the metal, the latter remaining as a shiny, partly sintered,metallic mass, stable to air. Another method for the preparationof the element is to decompose the halide in a high vacuum on anelectrically heated tungsten filament according to the reaction :ZPaX, = 2Pa + 5X,The protoactinium forms a shiny, greyish-white, partly moltendeposit on the filament.The metal does not oxidise in air, inexpected contrast to metallic radium, and retains its lustre for an37 6. Amer. Chm. SOC., 1934,56, 2200; A., 1934, 1319146 INORGANIC CHEMISTRY.appreciable time.550" by reaction of the oxide with carbonyl chloride :The pentachloride can be prepared readily a tPa,05 + 5coc1, = 5C0, + 2PaC1,The pentachloride sublimes in transparent, nearly colourless, longneedles. The chloride melts to a pale yellow liquid at 301", butsublimes appreciably below its melting point.The complex fluoride,K,PaF,, crystallises in colourless long needles, very sparinglysoluble in water containing hydrofluoric acid (06y0). They arestable in air and can be dried to constant weight at 20" or 100".A survey of the analytical reactions of protoactinium provesdefinitely that the oxide Pa,O, is basic, in contrast to the pentoxidesof tantalum, niobium, and vanadium, which are acidic or amphoteric.This is illustrated by the fact that when a zirconium-protoactiniummixture containing tantalum is fused with potassium carbonate,the tantalum goes, at least partially, into solution, whilst theprotoactinium remains in the residue. Protoactinium is evidentlyless basic than zirconium, for when an acid solution of zirconiumchloride containing a low proportion of protoactinium reacted withan ice-cold solution of ammonium carbonate, some 12% of the zir-conium but only 0.8 yo of protoactinium dissolved.The importance of this work on protoactinium is not restrictedto the fact that it extends in a most interesting way our knowledgeof the chemistry of the elements of Group V.It also gives us newknowledge about the element which is the direct mother-substanceof actinium, and enables us, for the first time, to determine thisimportant atomic weight. The simplest and most accurate wayof obtaining this atomic weight would be by a mass-spectrographanalysis. I n view of F. W. Aston's success with the fluoridesof uranium, tantalum, and niobium, the fluoride PaF, would nodoubt be very suitable for the purpose, but unfortunately its pre-paration is excluded by the relatively large quantity of startingmaterial required.However, using the chemical method, wherebythe ratio K,PaF, : Pa,O, was determined, A. von Grosse 38 hasfound that the atomic weight is of the order 230.6, with an accuracyof This value is in complete agreement with F. W.Aston's results 39 on actinium-lead (Ac-D = 207) obtained by themass spectrograph. For the chemist, this atomic-weight determin-ation has a peculiar interest, as it adds a fourth to the three well-known anomalies of atomic weights, cobalt-nickel, argon-potassium,tellurium-iodine. Thorium (at. no. go), which precedes proto-actinium, has an atomic weight of 232.1.3* Proc. Roy. Soc., 1935, [ A ] , 150, 363.39 Ibid., 1933, [ A ] , 140, 535; A., 1933, 762.0.5 unitCARTER AND WUDLAW : THE PERIODIC TABLE.147A careful study of the Periodic Table must raise the pertinentquestion why the system of elements should end so abruptly a turanium. It is most remarkable that this element, which is con-sidered to have the highest possible atomic number, should have anexceedingly long life, although radioactive, and should be by nomeans rare. It was, therefore, with unusual interest that in 1934the announcement was received from Eome that E. Fermi,40 in-vestigating the products of neutron activation of various elements,had reported the possibility that an element of atomic numberexceeding that of uranium had been obtained. In a later com-muni~ation,~~ Fermi expresses the opinion that further experimentsmade by himself and his collaborators support the hypothesis thatthe 13-minute and 100-minute induced activities of uraniumare due to transuranic elements. He states that the simplestinterpretation consistent with the known facts is to assumethat the 15-second7 13-minufey and 100-minute activities are chainproducts, probably with atomic number 92, 93, and 94 respectivelyand atomic weight 239.Fermi’s interpretation has been subjectedto criticism by various writersY42*43 but it must be emphasisedthat the work of Fermi and his school is carried out with greatcare and considerable ingenuity, and that the conclusions whichthey draw are tested, wherever possible, by a variety of chemicaland physical experiments.The reader will naturally enquire atthis stage whether elements of atomic number greater than 92have ever been definitely identified. The answer must be that,although there are indications that such elements may be formedby neutron activition of uranium, yet the principal workers in thisfield are careful to emphasise the difficulty of obtaining conclusiveproof.About the same time as Fermi’s announcement in 1934, 0,Koblic 44 stated that he had obtained in considerable quantity fromJoachimsthal pitchblende a pure silver salt, AgXO,, where X waselement 93. He described the properties of the element and itscompounds, estimated its atomic weight as 240 from an analysisof this silver salt, and suggested the name “ bohemium ” for thenew element.Certain of the compounds were then submittedE. Fermi, E. Amaldi, 0. d’Ago-stino, F. Rasetti, and E. SegrB, Proc. Roy. SOC., 1934, [ A ] , 146, 483; A . , 1934,1284.4l E. Amaldi, 0. d’Agostino, E. Fermi, B. Pontecorvo, F. Rasetti, and E.Segr6, ibid., 1935, [A], 149, 522; A., 1935, 910.4a A. von Grosse and M. S. Agruss, Nature, 1934,134, 773; L4., 7; PhysiccrlRev., 1934, 46,241 ; J . Amer. Chene. SOC., 1936,57,438; A., 659.43 Angew. Chern., 1934, 47, 653..14 Nature, 1934, 134, 66.4O Nature, 1934, 133, 898; A., 1934, 826148 INORGANIC CHEMISTRX.to (Frau) I. Noddack 43 for report. A chemical and X-ray investiga-tion of the supposed silver and thallium salts RXO, revealed thatelement 93 was entirely absent, and that these substances weremixed salts of silver or thallium vanadate and tungstate with excessof tungstic acid.This claim to the discovery of element 93 hasconsequently been withdrawn. S. R. C. w. w.2 . SOME ELEMENTS AXD COMPOUNDS.Despite the ever-increasing mass of published work, it is notalways realised that our knowledge of some of the most familiarinorganic substances is far from complete, and that there stillremains much to discover about some of the best-known elementsand compounds. It is only recently, by the use of refined physicalmethods, that an insight has been obtained into the detailed structureof sulphur and phosphorus, and that the constitution of bleachingpowder has been elucidated, Again, during the past few years,important extensions have been made to the fundamental chemistryof common elements such as nitrogen and sulphur, for, by theutilisation of improved appliances and new experimental technique,a number of simple derivatives of great theoretical interest havebeen isolated.Finally, the application of modern theoretical ideasto the problems of valency and chemical combination has resultedin striking developments in the field of molecular structures. Inthis section of the Report an attempt is made to deal with specificexamples of advances in these various directions.It is well known that below 96" sulphur is stable in the ortho-rhombic form, and the chemical evidence strongly suggests theexistence of an S8 molecule in rhombic sulphur. An X-ray in-vestigation has shown that the structure of rhombic sulphur isdefinitely molecular.The S8 molecule is a puckered 8-atom ringwhich may be considered as made up of two squares, one turned45" with respect to the other (Fig. 1). The planes of the two squaresare separated by 1-15 A., the S-S distance is 2.12 A., and the bondangle is 105". The closest distance of approach of atoms in neigh-bouring molecules is about 3.3 8. A study of the high-temperatureforms of sulphur has also been made by J. J. Trillat and K. H.Meyer. It is common knowledge that when sulphur is heated t o170" it becomes highly viscous and if cooled, say, by being plungedinto water, an amorphous plastic product results. If threads ofthis amorphous product are stretched, they show double re-fraction, and J.J. Trillat and H. Porestier have found that theyB. E. Warren and J. T. Butwell, J. Chem. Phpsies, 1935,3,6 ; A., 285.Bull. SOC. chirn., 1932, [iv]. 51,248; A., 1932, 462CARTER AND WARDLAW: so- ELEMENTS AND COMPOUNDS. 149give a fibre-diagram. K. K. Meyer and Y. Go3 deduce from anexamination of this diagram that the sulphur atoms are arrangedin long chains linked by covalencies arranged parallel to the directionof stretching (Fig. 2). The relationship between rhombic and elasticsulphur which these results disclose is most interesting, but thedetailed mechanism whereby rhombic sulphur is converted intoelastic sulphur is still under discussion.An important potential source of sulphur is the sulphur dioxideliberated in the course of many industrial processes.During theroasting of zinc, copper, lead, and nickel ores, for example, verylarge amounts of sulphur are lost in the form of the dioxide. Theproblem of obtaining sulphur from such waste gases has been underFIG. 1. F I G . 2.consideration at Billingham, and, as a result, a valuable processhas been evolved, which M. P. Applebey has described in a recentpaper.4 Researches, extending over some years in the laboratoriesof Imperial Chemical Industries Ltd., have solved the difficulty ofconcentrating sulphur dioxide from metallurgical gases, containing3-7% SO,, by the ingenious method of using a sulphite-bisulphitebuffer system which can be regulated t o have a moderately highpE in the cold and a much lower one when hot.This is achievedby the addition of a substance such as an aluminium salt, thehydrolysis of which is much increased by rise of temperature. Ithas been further demonstrated that the almost pure sulphur dioxideso obtained may be nearly quantitatively reduced by coke in accord-ance with the equation :so, + c "= co, + sThis reaction, which has been known for a long time, thoughgenerally overlooked by the text-books, furnishes a stl iking exampleHelv. China. Acta, 1934, 17, 1081; A., 1934, 1296.Chem. a d I d , 1034,53,1007150 INORGANIC CHEMISTRY.of sulphur dioxide as an oxidising agent. The reduction, whichtakes place very rapidly and almost completely at 1100", is exo-thermic, so no external energy has to be supplied when the requisitetemperature has been attained, M.P. Applebey deals with thefar-reaching possibilities that arise from this successful process,and the impression is gained that these developments are probablythe most important which have been made in the heavy chemicalindustries for some considerable time.The reaction between sulphur dioxide and oxygen in aqueoussolution, with or without the presence of allrali, has recently becomeof practical importance in relation to the problem of removingsulphur dioxide from the flue gases of power stations. A reviewof previous work showed that the mechanism of the reaction wasobscure, especially with regard to the possible effect of surfaces.In an important series of papers, R. C. Hoather and C. F. Goodevehave recorded the results they have obtained from a detailed in-vestigation of this subject.The formation of dithionate by the oxidation of sulphurous acidand sulphites is dealt with in a comprehensive study by H.Bassettand A. J. Henry.6 They find that oxidation of sulphurous acidand of sulphites by chlorine, iodine, and hydrogen peroxide yieldsbut very small amounts of dithionic acid, and this is realised onlyunder acid conditions. Oxygen gives dithionate in amounts whichmay be large. The yield depends on the sulphite concentrationand the acidity. During the photochemical decomposition ofsulphurous acid into sulphuric acid and sulphur, no dithionateappears to be formed. The theoretical considerations involvedin the different reactions are carefully discussed.In a previous Report an account was given of the preparationof sulphur monoxide by an electric discharge in a mixture of sulphurdioxide and sulphur vapour a t low pressures.P. W. Schenk7 hasnow shown that it is possible to obtain a gas containing 40% ofsulphur monoxide, mixed with the dioxide, by the direct com-bustion of sulphur. I<. Heumanns in 1883 had observed thatsulphur at 200" was slowly oxidised, with a feeble phosphorescence,and that a peculiar smell was noticed which he ascribed to sulphurmonoxide. He tried unsuccessfully to prove the existence of themonoxide by leading the gas into alkali. Schenk 7 now finds thatoxidation of sulphur with air at atmospheric pressure gives only asmall concentration of monoxide when determined spectroscopically,Trans.Paraday SOC., 1934,30, 626, 630,1149,1156; A., 1934,1086, 1157;J., 1935, 914; L4., 1090.Z. anorg. Chem., 1934, 220, 268; A., 1936, 51.A., 1936, 42.a Ber,, 1883, 16, 139CARTER AND WARDLAW: SOME ELEMENTS AND COMPOUNDS. 151but, by operating a t low pressures with oxygen, concentrationsup to 40% of monoxide may be obtained when the optimum temper-ature for the reaction has been realised. This optimum temperaturemust be determined experimentally. At pressures of 30, 8, and5 mm., the yields of sulphur monoxide have been found to be7, 30, and 40% respectively. Pressures of oxygen lower than5 mm. cannot be used, as the sulphur flame is extinguished. I n arecent communication C. W. Montgomery and L.S. Kassel9 calculatethe equilibrium constants for the following reactions from spectro-scopic data :(a) 2so = s, + o,, ( b ) 2so = so, + +s2They find that, at very low temperatures, the equilibrium liesright over on the left in reactioii (a). In reaction ( b ) they reportthat an appreciable vapour pressure of sulphur monoxide onlycomes into equilibrium with the dioxide and sulphur at about2000" Abs. These findings are in agreement with the qualitativeobservations made by Schenk. As the temperature of the sulphurflame is so low, an equilibrium position in the sense of the reaction( b ) is not attained. It follows, therefore, that the sulphur monoxideformed in (a) is removed from the sphere of reaction before theequilibrium required by ( b ) is operating.Schenk 10 has studied,by density determinations, the decomposition which sulphurmonoxide undergoes into sulphur and sulphur dioxide. His resultsshow, from the pressure changes observed, that when sulphurmonoxide is decomposed according to reaction (b), 64% of the sulphurmonoxide molecules are associated to (SO),.Although the existence of a tetroxide of sulphur has been sus-pected by various investigators, it has only recently been isolated.R. Schwarz and H. Achenbach l1 have now prepared it by theaction of the glow discharge on a mixture of sulphur dioxide andoxygen (in the proportion 1 : 10) at 0.5 mm. pressure. The reactionproducts are passed through two vessels cooled in liquid air, awhite solid being obtained. This is freed from sulphur dioxideand ozone by warming to - 30" in a current of oxygen, the tetroxidethen remaining.The yield from an experiment lasting six hoursis 0.1 g. Sulphur tetroxide is a white solid which begins to de-compose at - 5" with evolution of oxygen, and at 3" it melts withdecomposition, giving oily drops of S,O,. By the freezing-pointmethod a molecular-weight determination in pure sulphuric acidgave a value 95 in close agreement with 96 required by the simplelo P. W. SchenkandH. Platz, 2. anorg. Chem., 1935,222,177; A., 593.l1 Ibibid., 1934, 219, 271; A., 1934, 1183.J . Chem. Physics, 1934,2,417 ; A., 1934,966152 INORGANIC CHE1KISTRY.formula SO4. Schwarz and Achenbach suggest the structuralformula A for the oxide. Although theoretically thistetroxide could be the anhydride of permonosulphuric acid, thereis no evidence that it will undergo hydration to H,S05.Whenthe tetroxide is dissolved in dilute sulphuric acid at O", slow de-composition takes place with oxygen evolution, but no H,SO,appears to be formed. This inability to become hydrated isattributed by Schwarz and Achenbach to the co-ordinative saturatedcharacter of the sulphur. Sulphur tetroxide is an excellent oxidisingagent : it will oxidise aniline to nitrobenzene, and bivalent manganeseto the septavalent state.A. H. Spong l2 has concluded that ordinary sulphur monochlorideis probably a mixture of the two forms (1) S-S<cl and (2) Cl*S*S*Cl.G. Giacomello13 has now studied the reaction of phenol and ofp-naphthol with sulphur monochloride in benzene.From thereaction with phenol he isolated one compound with the probablestructure PhO*S*S*OPh, and from the p-naphthol reaction thederivative C,,H7*0*S*O*C,,H7. On the other hand, from p-chloro-phenol he obtained two products C6H,C~*O~SoS0~*C6H,C1 andC,H4C1*O*S *o*c6H4c1. These results are entirely in agreementwith the view that sulphur monochloride is a mixture of (1) and (2),and if this view is correct, then the hypothetical thiosulphurousacid should exist in two isomeric forms (1) SS<OH and(2) HO*S*S*OH. F. LengfeldI4 some 40 years ago examinedthe action of alcohol-free sodium methoxide or ethoxide on a well-cooled solution of sulphur monochloride in light petroleum. Heobtained one ester, in each case as a colourless oil.A. Meuwsen l5now claims that two isomeric esters can be isolated, one similar tothe colourless oil of Lengfeld, and the other a greenish-yellow oilof different constitution but of the same molecular weight. Thecolourless oil he considers to have the constitution OR*S*S*ORwhereas to the other ester he assigns the formula S=S( OR),.H. Stamm l6 maintains, however, that Meuwsen's greenish-yellowester is merely Lengfeld's colourless ester contaminated with about2.5% of sulphur monochloride, and this would account for thechemical and physical differences which have been noted.R. Scholder and G. Denk17 have prepared the first salt of the0-s=o0-0c1 - +OH - +13 J., 1934, 485; A., 1934, 605.13 Atti R.Accad. Lincei, 1935, [vi], 21, 36 ; A., 614.14 Ber., 1895, 28, 449.1 6 Ibid., p. 673; A., 729.16 Ber., 1935, 68, [B], 121 ; A,, 326.l7 2. anorg, Chew., 1935, 222, 17; A,, 461CARTER AND WARDLAW: SOME ELEXENTS AND COMPOUNDS. 153hypothetical sulphoxylic acid, H2S02. When aqueous solutionsof cobaltous chloride and sodium hyposulphite (Na,S,O,) reactedin the presence of ammonia, ethylenediamine, or pyridine, a darkred solution was formed which, on dilution with water, gave adark brown flocculent precipitate of cobalt sulphoxylate,CoSO,,xH,O, probably in a polymerised form.A. M. Middleton and A. M. Ward 18 have investigated the com-position and properties of precipitated nickel and cobalt sulphides.They find that, with air exclusion, the sulphides formed areNi(SH),, Co(SH),, and Co(SH),, which yield NiS, CoS, and Co,S,when dried in nitrogen.In the presence of oxygen, however,variable addition takes place, initially a t the sulphur atoms in -accordance with the scheme Niaccompanied by intramolecular rearrangement. Drying of theoxygenated sulphides results in a partial elimination of hydrogensulphide and water, and further oxygenation may proceed. Thedried and the undried oxygenated aulphides are the substancesusually obtained in qualitative analytical procedure.Although a peroxide of nitrogen appears t o be formed by theaction of a ailent discharge on a mixture of oxygen and nitrogendioxide,* its existence has hitherto only been indicated by a changein colour of the mixture and the appearance of certain characteristicspectral lines.Now, however, R. Schwarz and H. AchenbachI9have succeeded in preparing pure nitrogen trioxide, NO,, andexamining its properties. The apparatus they employed wasessentially the same as that used by them for their successfulsynthesis of sulphur tetroxide. A mixture of nitrogen dioxide andoxygen in the proportion 1 : 20 at a pressure of 1 mm. was passedthrough the apparatus and submitted to a glow discharge. TheNO, : 0 ratio must be carefully adjusted. Lower oxides of nitrogenare formed unless a large excess of oxygen is present, but too greatan excess gives ozone. The reaction products were condensed invessels cooled with liquid air, and a pale blue condensate wasobtained, apparently a mixture of dinitrogen trioxide and higheroxides. By modifying this apparatus so that the condensing tubeformed part of the discharge tube, they obtained a colourless depositof pure nitrogen trioxide, NO,.It is stable a t - 142", above whichit slowly decomposes into nitrogen dioxide and oxygen. In aqueouBmedia it is relatively stable. For example, at 15-20', the oxidation18 J., 1935, 1459.* The nomenclature used in this article denotes the actual numbers ofatoms of nitrogen and oxygen in the molecule ; e.g., dinitrogen trioxide N,O,,nitrogen trioxide NOb.-Ed.l9 Ber., 1935, 68, [B], 343; d., 457154 INORUANIC CHEMISTBY.value of nitrogen trioxide in sulphuric acid ( N / 5 ) requires about50 hours to sink to zero. Again, when the trioxide is dissolvedin nitric acid (2N) and potassium iodide is added, although iodineis liberated immediately, the reaction reaches completion only after30 minutes.With sodium hydroxide the trioxide reacts thus :ZNO, -+ 2NaOH = NaNO, + NaNO, + 0, + H20The reaction takes place in two stages. In the presence of water,the trioxide is decomposed into the dioxide and oxygen and thedioxide then yields NO,' and NO,'. The trioxide does not function,therefore, as an acid anhydride. Although direct determinationsof molecular weight are impossible owing to the instability of thecompound, the authors consider that it has the monomeric form.They base this opinion on the very low temperature of condensationof the trioxide and its spectroscopic behaviour.Further, theyconsider that it must have a co-ordinatively unsaturated characterfrom the fact that it can be extracted from aqueous media by ether.The distribution coefficient between water and ether is, in fact,1 : 3. As hydrolysis in water is not accompanied by the productionof hydrogen peroxide, the constitution NO,-O-O-NO, is excludedand identity with the dimeric oxide N,O, not possible. Theauthors believe, therefore, that their new trioxide has the formula0 O=N<?, which would be more correctly written as O+N< I 0 0'T. M. Lowry and J. T. Lemon 2o report that when dry dinitrogenpentoxide (N205) is vaporised in a stream of ozonised oxygen andthen passed through a glass tube heated by a small flame, thecolourless gas becomes brown, through the formation of nitrogendioxide, a short distance before the flame is reached.A narrowzone of a dark grey-blue colour is, however, seen hovering at theboundary, and this is preceded by a zone of clear blue. In a longtube the blue flame thus formed " strikes back " from time to timeat the rate of about 10 em. per second to the point at which the gasenters the tube, which is then filled from end to end with brownnitrogen dioxide. When the concentration of dinitrogen pentox-ide is low, the grey boundary between the colourless incoming gas andits pale brown decomposition products remains stationary and doesnot strike back. It is suggested that the formation and disappear-ance of the blue zone may be due to the production and decom-position of it higher oxide of nitrogen; e.g., N20, + 0, =2NO,(blue) + 0,; 2N0, = 2N0, + 0,.The temperature at theboundary is probably below 100". In a study of the binary systemN,O,-N2O5 Lowry and Lemon 21 find that dinitrogen tetroxide2o Nature, 1936, 185, 433 j A., 593. 21 J., 1936, 692; A., 824CARTER AND WARDLAW: SOME ELEMENTS AND COMPOUNDS. 155and pentoxide give a simple freezing-point diagram with a eutecticat 10.8% N,O, and - 1543".During recent years some exceptionally interesting compoundshave been isolated from reactions involving fluorine. G. H. Cady 2,has now investigated the action of fluorine on nitric acid (3N)and thereby discovered the new gaseous compound, NO,F. Inhis experiments, Cady sometimes experienced explosions, but0.Ruff and W. Knasnik23 find that the danger of explosion isminimised by the use of nitric acid of higher concentration, andthat it disappears entirely if pure nitric acid (100%) is employed.At low temperatures, e.g., - 35", the reaction is incomplete, but at + 20" it proceeds quantitatively in accordance with the equationHNO, + F, = NO,F 4- HFThe apparatus used is constructed partly of quartz glass and partlyof ordinary glass. The hydrogen fluoride liberated in the reactionis absorbed by anhydrous potassium fluoride, and the other productsare condensed in receivers cooled with liquid air. The condensateis fractionally distilled at 100 mm. pressure through a quartz-glasscolumn, and the portion passing over at - 79'199 mm. is pureNO,F. Other fractions consist of SiF4, H,SiP,, and H,F,.Amolecular-weight determination by the density method gave thetheoretical value for the molecule N0,F. The fluoride, which isnormally a gas with critical temperature estimated t o be 67.2",condenses to a colourless liquid, b. p. - 45-9"/760 mm., and to awhite solid, m. p. - 175". The density of the solid at - 193.2" is1.951, and that of the liquid is 2.2148 - 0.003114T. Solid orliquid NO,F is exploded by mechanical shock. The vapour pressurehas been measured with a quartz-spiral manometer between - 128"and - 68" and is given by log p = - 1044*9/T + 7.478, and thelatent heat of vaporisation is 4726 g.-cals. per mol. The gas isstable in dry glass or quartz, but with water it gives oxygen andfluorine monoxide together with nitric acid and hydrogen fluoride.It reacts with dilute sodium hydroxide solution (2%), OF, beingliberated :2NO,F + 2NaOH = 2NaN0, + OF, + H,OIf a more concentrated sodium hydroxide solution (20%) is employed,the initially formed fluorine monoxide decomposes in the usualway and oxygen alone is obtained. These experimental findingscharacterise NO,F as tt derivative of the fluorine monoxide, Ol?,,in which a fluorine atom is replaced by the NO, group, i.e.,22 J .Amer. Chem. Soc., 1934,56,2635; A., 1935, 181.23 Angew. Chem., 1935, 48, 238; A., 715; D. W. Yost and A. Beerbower,J . Amer. Chem. SOC., 1935, 57, 782; A , , 715156 INORGANIC CHEMISTRY.Z)-O-F, or in modern nomenclature h;N-O-F. Although 04the reaction of NO$' with water or sodium hydroxide solutiondoes not give an explosion, contact with alcohol, ether, or anilinecauses one immediately.On the other hand, with glycerol, aceticacid, or acetone no reaction is perceptible. Moreover, NO$'is appreciably soluble in acetone. The odour of the new compoundis described as irritating.K. Gleu and R. Hubold% have reinvestigated the per-acidwhich is formed on mixing a nitrite solut?on with acidified hydrogenperoxide and to which Raschig and other investigators have giventhe formula HNO, and the name pernitric acid. Gleu and Huboldprepared the supposed pernitric acid by adding to a solution ofsodium nitrite (1 mol.) and hydrogen peroxide (> 1 mol.) muchice and then the equivalent amount of sulphuric acid (2N), excessof sodium hydroxide being added after two seconds.In the in-tensely yellow solution produced, 70% of the nitrite may be con-verted into the per-acid under the best experimental conditions.To obtain this result it is most important to observe the correctinterval of time between addition of acid and of alkali. Theexcess of hydrogen peroxide, in the presence of sodium nitrite andthe sodium salt of the per-acid, was determined by titration withsodium hypochlorite (0-1N) in alkaline solution in the presence often drops of potassium iodide (0.1N) until the colour of the ruthen-ium-red indicator was discharged. Excess of arsenious oxide(0.1N) was added immediately and back-titrated with sodiumhypochlorite, the same indicator being used, to determine the activeoxygen in the per-acid.The NO,' was determined by reductionto ammonia with vanadyl sulphate, VOSQ,, and distillation intoacid. Under these conditions any nitrate is unaffected. The ratio,active 0 : NO,', was always 1 : 1, proving the presence of per-nitrite HNO,*O and absence of pernitrate. These findings confirmthe views of J. Schmidlin and I?. Massini25 put forward in 1910.It may be mentioned again that the nitrogen trioxide (NO,) recentlyprepared by Schwarz and Achenbach did not prove to be theanhydride of pernitric acid.Phosphorus crystallises in a number of allotropic forms, three ofwhich, the white, the red, and the black variety, are easily dis-tinguished because of their widely differing physical and chemicalproperties.It is well known that white phosphorus is easilychanged into the red variety by heat, light, or X-rays, but that toconvert white into black phosphorus 26 a pressure of 12,000 atmo-a 5 Ber., 1910,43,1162; A., 1910, ii, 498. 24 2. anorg. Chem., 1935,223,305.26 P. W. Bridgman, J . Arner. Chem. Soc., 1914,36, 1334; A., 1914, ii, 647CARTER AND WARDLAW : SOME ELEMENTS AND COMPOUNDS. 157spheres is necessary at a temperature of 200". Recently, blackphosphorus has been made from white phosphorus by a pressureof 35,000 atmospheres at room temperat~re.~' The density ofblack phosphorus is very high, being 2-6-2.7, as against 1.83 forwhite and 2.34 for red phosphorus. All three varieties when heatedgive a vapour composed of P, molecules, which condenses to whitephosphorus.At lower temperatures (100" in a vacuum), however,the vapour of red phosphorus condenses unchanged, but the vapourdensity is so low that no molecular-weight determination has yetbeen made. These facts seem to indicate that white phosphorusis composed of more or less loosely-bound P, molecules, whichreadily break up to form the more stable structures of red andblack phosphorus. It is most difficult to come to a decision inthis matter, for data on phosphorus are confusing and incomplete,FIG. 3.and until recently none of the crystal structures had been satis-factorily worked out by X-ray methods. An important advancehas now been made by R. Hultgren, N. S. Gingrich, and B. E.Warren 28 as a result of their investigation of the crystal structureof black phosphorus. They find that the unit cell of black phos-phorus consists of two double layers, one of which is shown inFig.3. Each atom is bound to three nearest neighbours at 2-18 A.in agreement with the accepted atomic radius of 1-10 A. and withthe co-ordination to be expected from covalent bonding of phos-phorus. Two of the bonds are in the plane of the layer at 99"from one another; the third is between layer halves at 103" 30'from both, making the average bond angle 102". This agrees withthe tendency of bond angles to be nearer tetrahedral for the lighterelements of the periodic group. The decisively covalent characterof phosphorus is clearly shown by the fact that the distance betweenbonded atoms is only 2.18 A.compared with 3.68 A. as the closest27 P. W. Bridgman, Physical Rev., 1934, 45, 844.28 J . Chern. Physics, 1935, 3, 351; A., 919158 INORCANIC CHEMISTRY.approach between atoms in different layers. In arsenic, which isappreciably more metallic than black phosphorus, the bonds areless sharply differentiated ; the three nearest neighbours are a t2.51 A., with three more atoms at 3.15 A. Atomic distributioncurves of crystalline and amorphous black and red phosphorus,obtained by the method of Fourier analysis, furnished a valuableclue to the structure of red phosphorus, for they showed that eachatom formed three covalent bonds in the normal manner. More-over, the fact that the second peak was at practically the sameposition as in black phosphorus suggests that the bond angles arenearly the same.Ortho-salts, such as Na3N04, Na3N03, Na,Zn04, and Na,CuO,,have been prepared 2o by the action of sodium oxide, Na,O, on thesalts of oxy-acids or on weakly acidic metallic oxides.They areusually decomposed by water or carbon dioxide, so cannot be pre-pared from alkali hydroxides or carbonates. The proof of theexistence of these ortho-salts has been established by X-ray analysis.The preparation of &,NO, brings out an interesting relationshipbetween nitrogen and phosphorus. However, as 4 is the maximumco-ordination number of nitrogen, it will, presumably, not bepossible to prepare the nitrogen analogue of the hypotheticalH,PO,, esters of which have been isolated.30The constitution of bleaching powder has been the subject ofinvestigation and speculation for many years.The older methodsof examination, depending on such properties as solubility, vapourpressure, mode of chemical decomposition with heat or acids of thecarbonic and hydrochloric type, have been used at various timesbut with no great measure of success. The problem has now 31 beenattacked by phase-rule studies, by extended microscopic examin-ation, and by the use of X-ray powder photographs. As a resultof the phase-rule work, all the pure compounds which could beisolated under equilibrium conditions from the system CaO-CaC1,-Ca(OCl),-H,O below 40" have been defined and used as a basis forcomparison with the solid phases present in bleaching powder.It has been established that the first stage, in the reaction betweenchlorine gas and calcium hydroxide in the preparation of bleachingpowder, is the formation of the basic hypochlorite Ca(OC1),,2Ca(OH),and the basic chloride CaCl,,Ca( OH),,H,O.On further chlorination,the former is converted into another substance which appears to be29 E. Zintl, M. Morawietz, and G. Woltersdorf, ~ a t U W i S S . , 1935, 23, 197 ;30 L. Anschutz and W. Broeker, Ber., 1926, 59,2848; A., 1926, 146.31 C. W. Bunn, L. M. Clark, and I. L. Clifford, Proc. Roy. SOC., 1935, [-4],E. Zintl and W. Haucko, Z.physika1. Chem., 1935, 174, 312; A., 1936, 16.151, 141; A . , 1214CARTER AND WARDLAW : SOME ELXMENTS AND COMOUNDS. 159a mixed crystal whose chief constituent is calcium hypochlorite.Ordinary bleaching powder, containing about 35% of availablechlorine, is a mixture of this hypochlorite mixed crystal with thebasic chloride CaC12,Ca( OH),,H,O.On further chlorination, thelatter is partly but never completely converted into hydrated calciumchloride (usually the tetrahydrate, CaC1,,4H20) whilst the hypo-chlorite mixed crystal persists, with gradually changing properties.The most highly chlorinated sample which was examined consistedof hypochlorite mixed crystal, CaC1,,4H20, and CaCl,,Ca( OH),,H,O.The non-deliquescent nature of ordinary bleaching powder and thedifficulty of introducing more than 35% of available chlorine intothe solid are due to the presence of the basic chlorideCaCI,,Ca( OH),,H,O,which appeass to be a very stable substance.R.K. Bahl and J. R. Partingfon3, have reinvestigated thelower oxides and sulphates of iodine. Contrary to some statementsin the literature, which are reproduced in text-books, they provethat the interaction of iodine and cold concentrated nitric acid(d, 1-5) produces the pentoxide I,O, and not the dioxide. Inconformity with the weakly basic properties of iodine, they fmdthat basic saltsare more stable than the normal ones. They wereunable, however, to obtain Chrbtien’s sulphate, Iz0,,S03,&H,0,by heating a mixture of iodic acid and concentrated sulphuric acidtill iodine is evolved. The analysis showed that the product wasprobably I,0,,H,S04 containing some I,O,,H,SO,.Although numerous attempts have been made to prepare com-pounds of the rare gases, practically all investigations have servedonly to show the inertness of these elements.P. VillardF3 however,in 1896 claimed to have obtained an unstable hydrate of argon,and since then hydrates of krypton and xenon have been reported.Prompted by the idea that boron, in the trifluoride, BF,, is veryreactive by reason of its incomplete octet, H. S. Booth and K. S.Willson 35 have examined the system argon-boron trifluoride bythermal analysis. A graph of the freezing points against composi-tion exhibited maxima and minima. The maxima correspondedto the ratios A,BF,; A,2BF3; A,SBF,; A,6BF3; A,8BF3; andA,16BF3, indicating compound formation, but the compoundsare unstable and dissociate above their melting points. From theshape of the curve, the ratio A,2BF3 appears to be the most stable.32 J., 1935, 1258; A., 1334.33 Compt.rend., 1896, 123, 377; A., 1897, ii, 31.34 R. de Forcrand, ibid., 1923, 176, 365; 1925, 181, 15; A., 1923, ii, 239;35 J. Amer. Chem. SOC., 1935, 57, 2273.1925, ii, 812160 INORGANIC CHEMISTRY.The conclusion that the association of argon with the boron fluorideis of the type BF, +-A analogous with BF, ,+-NH, must beaccepted with reserve. The function of the BF, may be similarto that of water in the hydrates of the inert gases. In N. V. Sidg-wick’s 36 opinion, the water is held by the van der Waals forces inthe crystal along with those of the solute, and the simple numericalratio of the different kinds of molecules merely reflects the geo-metrical regularity of the crystal.So far, no exception has been found to the generalisation that allelements with a covalency of 6 have the six valencies arrangedin the form of an octahedron.When, however, quahicovalentatoms are considered, no such simple generalisation is available.It will be recalled that in 1931 L. Pauling 37 applied wave-mechanicalmethods to the problem of covalent linkages and showed that for4-covalent atoms, in which electrons from the first two sub-groups(s and p ) of the outer level are used in chemical bonds, the stableconfiguration is tetrahedral. In the case of the transition elements,however, one or more electrons used in binding may belong to theci! sub-group of the incomplete inner level, and in these cases, thefour covalencies may be distributed in a plane.Moreover, sincethe d electrons are chiefly responsible for the magnetic moment ofthe atom, sharing of them should reduce this property, so that4-covalent nickel compounds should be diamagnetic. Pauling ’sconclusions have been criticised by W. Heisenberg 38 and others,but appear to be essentially correct. The point to be recognisedin the case of the transition elements is that the results are probablypermissive, i.e., they show that the metals in question may have,but not necessarily must have, a planar distribution of valencies.That this interpretation is correct is indicated by some recentwork of H. M. Powell and A. F. They examined the com-plex salt Cs,CoCl, by X-ray analysis, and found that the groupCoC1, had an approximately regular tetrahedral configuration,thereby proving that the transition element cobalt can actuallybe tetrahedral when 4-covalent.To Pauling belongs the credit of first suggesting that nickel inits 4-covalent state may have a planar distribution of valencies,and it is remarkable that Werner omitted it from the list of elementswhich he predicted would give planar structures.Pauling’s con-clusion that 4-covalent nickel should be planar and diamagneticreceived strong support from the isolation, by S. fhgden 40 in 1932,36 “ The Covalent Link in Chemistry,” 1933, p. 29.37 J . Arner. Chem. SOC., 1931,53,1367; A., 1931,670.38 See Ann. Reports, 1931, 28, 367..J., 1935, 359; A., 570. c0 J., 1932, 246; A., 1932, 272CAXTER AND WARDLAW: SOME ELEMENTS AND COMPOUNDS. 161of two diamagnetic compounds of nickel with benzylmethylglyoxime.As a matter of historical interest, it should be mentioned that in1910, L. A. Tschugaeff 41 discovered that nickel bis(monomethg1-glyoxime) could occur in two isomeric forms, but he did not suggestthat a possible explanation of the isomerism lay in the planar dis-tribution of the nickel valencies. Recently, H. J. Cave11 and S.Sugden 42 have shown that the occurrence of pairs of isomeridesappears to be general for the nickel derivatives of unsymmetricalglyoximes, and that the explanation must be that nickel is capableof forming 4-covalent compounds of planar type. F.P. Dwyerand D. P. Mellor43 have recorded the preparation of two innercomplex compounds of palladium with benzylmethylglyoxime, andadduced evidence to show that these substances are cis- and trans-isomerides (I and 11) of bis-anti-benzylmethylglyoximepalladium.C,H,*CH2--#---#--CH, CGH, *CH,-G--G-CH,O t N N-OH O+-N N-OHr\ / sHO-# M->O Ot# #-OHCH,-C---C-CH,-C,H, C,H,*CH2-C-C-CH,Further support for Pauling’s predictions has been obtained from anexamination of the nickel and palladium derivatives of salicyl-aldoxime.44 These compounds have been shown, by X-ray analysis,to be isomorphous and to have a, trans-planar structure. Moreover,the nickel compound has been found to be diamagnetic. A strikingexample of planar nickel in a co-ordination compound of quite adifferent type has been discovered in an examination of the dithio-oxalates of nickel, palladium, and platinum.45 The anhydrouspot assiuin nickelodit hio-oxalat e (111) is completely isomorp houswith the corresponding palladium and platinum derivatives, and a,detailed X-ray examination has shown that the valencies of the4-covalent nickel, palladium, and platinum are in a plane with themetal atom.4 1 J .Rum. Phys. Chem. SOC., 1910, 42, 1466; A., 1911, i, 261; Chem. Zentr.,1911, 82, i, 871.42 J., 1935, 621; A., 980.43 J , Amer. Chem. SOC., 1935, 57, 605; A., 752.44 E. G. Cox, F. W. Pinkard, W. Wardlaw, and K. C. Webster, J., 1935,450 ;4 5 E. (2,. Cox, W. Wardlaw, and K. C. Webster, J , , 1935, 1475.A . , 684.RE(:P.-VOL. XXXJI - 162 INORGANIC CHEMISTRY.In view of the fact that compounds of 4-covalent cobalt may betetrahedral, the question naturally arises whether the 4-co-ordinatedcompounds of nickel, palladium, and platinum are invariablyplanar, If Pauling's prediction is true that 4-covalent compoundsof the transition elements would be diamagnetic when the fourelectron-pair bonds lie in a plane, then some recent work by R.B.Janes 46 on the magnetic susceptibilities of typical palladium deriv-atives indicates that palladium is invariably planar. He examineda wide range of derivatives, of which the following are typical :Group 1. Palladous salts with four other groups in the molecule :PdC12,2H20 ; PdC12,2NH, ; K2PdC1, ; K,Pd(CN),.Group 2.Palladous inner complex salts : palladium dimethyl-glly oxime.Group 3. Palladous salts where a double molecule is present :In every case the magnetic susceptibilities were diamagnetic.The corresponding platinum compounds have been less extensivelyin~estigated,~' but the several salts which have been measured areall diamagnetic. This physical evidence is in entire agreementwith the available X-ray and chemical evidence, and in view of thediverse nature o€ the compounds which have been studied, thereappears to be no doubt that the planar configuration is quite generalfor 4-covalent compounds of bivalent palladium and platinum,and it can only be modified in very special cases, such as may arisewith ter- and quadri-dentate groups.The evidence from magnetic susceptibility measurements in thecase of nickel, however, indicates that 4-covalent nickel is notinvariably planar.Nickelous salts of the type of Group 1 (above)are paramagnetic, except Ni(CO), and K,Ni(CN), which are dia-magnetic. If, therefore, Pauling's reasoning is sound, then in thecase of nickel only a few molecules of the type of Group 1 are planar,in contrast to the cases of palladium and platinum, where the planarconfiguration is always assumed. This conclusion, that the dis-tribution of the valencies in compounds of 4-covalent nickel maybe either tetrahedral or planar, is in agreement with the recentdiscovery that quadricovalent cupric compounds may be planar.In 1926, W. H. Mills and R. A. Gotts48 showed that 4-covalentcopper in its benzoylpynxvic acid derivative gave rise to opticalactivity, and so presumably belongs to the tetrahedral type.Now,E. G. Cox and K. C. Webster 49 have examined by X-ray analysis[Pd(N~,),C1212'4 6 J . Amer. Chem. Soc., 1935, 57, 471; A., 573.47 D. M. BOSB, 2. Physik, 1930, 65, 677; A., 1931, 25.48 J., 1926, 3121; A., 1927, 149.49 J., 1935, 731 ; A., 920CARTER AND WARDLAW: SOME ELEMENTS AND COMPOUNDS. 163a range of chelated 4-covalent compounds of bivalent copper,including the copper salts of acetylacetone and benzoylacetone,and have definitely established that these substances possess c2planar structure. The result is of interest from another point ofview, for S. Sugden has shown that a number of compounds ofthis type are paramagnetic.According to Pauline;, a planarconfiguration is to be expected when s, p , and d electrons are in-volved in valency bonds. It is possible, therefore, that bivalentcopper (at. no. 29) may possess a complete 3, sub-group of tenelectrons, some being shared, instead of an incomplete sub-groupas is usually supposed. This would involve one unpaired electron1 2 2 3 3 3 - 4-21 21 62 21 62 f in the fourth principal quantum level, givingrise to a paramagnetic moment of the same order as that actuallyobserved by S. Sugden.50The structure proposed by Werner for platinous compoundsof the type PtA,B, was that the valencies had a planar arrange-FIG. 4./ Phment. A test of this, and a differentiation between a planar and atetrahedral distribution of valencies, is provided by the substitutionderivatives of bisethylenediaminoplatinous salts [Pt enz] X,.Inan important communication by W. H. Mills and T. H. H. Q~ibell,~lthe preparation of the diphenyl dimethyl derivatives has beendescribed and the optical activity of the compounds (see Pig. 4)leaves no doubt that the planar imangement proposed by Werneris the correct one. The salts investigated proved capable of re-solution into antimeric optically active forms showing a high degreeof optical stability. The regular tetrahedral arrangement is therebyexcluded, and no reason exists for inferring pyramidal rather thanthe more symmetrical planar configuration. The authors mentionthat the iodide and chloride crystallise with water of crystallisation,but express the opinion that there is no reason to doubt that thecentral platinum atom of the complex cation is truly 4-covalent.The planar arrangement of the platinum valencies with the inter-valency angle of 90" is shown to give rise to a practically strainlessring, whilst an arrangement with a tetrahedral valency angle wouldJ., 1932, 161; A , , 1932, 324.61 J., 1935, 839; A., 1057164 INORGANIC CHEMISTRY.create a very considerable strain in a five-atom ring composedof carbon and nitrogen atoms and one 4-covalent platinum atom.The structure and configuration of certain diamminopalladiumcompounds have been studied by chemical and X-ray methods,and the findings are fully in agreement with Werner’s view. Adeep red, highly crystalline form of Pd(NH,),Cl,, has been shown 52to have a trans-configuration like the familiar yellow powder form,their difference being due essentially to different crystal structuresbuilt up from the same (truns-) molecules.Recently, it has been shown 53 by an X-ray analysis that theconfiguration of the 4- covalent compound of quadrivalent platinum,Pt(CH,),Cl, is tetrahedral.This is a result of unusual interest,for it indicates that the principal valency of an atom may be a factorof importance in deciding the configuration of its 4-covalentderivatives.F. G. Mann and D. Purdie 54 have published some valuableexperimental data on the parachors of palladium and mercuryin simple and complex compounds. They find that, in certainseries of organo-metallic compounds, both simple and complex,the metal atom shows an apparent parachor which falls steadilyas the homologous series is ascended, and may ultimately reach aconstant value.For example, in the homologous series ( SR,),PdC12,the parachor of palladium fell from 36 for the methyl to - 7 forthe n-amyl compound. Another point of interest has arisen in thisinvestigation. The dipole moments of the three ethyl compoundsof the chloro-series (SR,),PdCl,, (PRJ2PdC1,, and (AsR,),PdC12,vix., 2-27, 1-05, 1.04 ( x e.s.u.), strongly support the assump-tion that they are all stable truns-compounds. The symmetricaltrans-compounds should have a dipole moment closely approachingzero, whereas the cis-isomerides should have very high moments.This conclusion is in agreement with the X-ray and chemical evi-dence relating to the thio-ether compounds of platinous and palladouschlorides. 55G.T. Morgan,5s in an address to the Chemical Society, has dealtwith some important advances made in recent years in the studyof the rarer elements. He is, himself, an outstanding leader inthis field of work, and the results which he and his colleagues have52 F. G. Mann, (Miss) D. Crowfoot, D. C. Gattiker, and (Mrs.) N. Wooster53 E. G. CoxandK. C. Webster, 2. Krist., 1935, [A], 90, 561.54 J., 1935, 1549.s5 E. G. Cox, H. Saenger, and W. Wardlaw, J., 1934, 182; A . , 1934, 397;s 6 -7.. 1935, 554; A., 716.J . , 1935, 1642.H. D. K. Drew and G. H. Wyatt, J . , 1934,56; A., 1934,284HEDUES : NON-FERROUS ALLOY SYSTEMS.165accumulated form the main topic of the address. Certain seams01 Northumbrian coal give an ash containing up to 1% of germaniumand 0.05% of gallium, and G. R. Davies, working in the Teddingtonlaboratories, has elaborated a process for the extraction of boththese rare elements. The germanium is distilled over with acidas tetrachloride, whilst gallium trichloride remains. Rheniumhas been extracted from Australian molybdenite by a lengthyprocess involving fractional volatilisation, and ultimate separationwith organic reagents such as 8-hydroxyquinoline and dipyridyl.The address concludes with some reference to co-ordination com-pounds of ruthenium, and a plea for a wider investigation of therarer elements ".. . for discoveries will from time to time be madein the application of these materials which will redound to thecredit of our science, and the good of the community."S. R. C. w. w.3. NON-FERROUS ALLOY SYSTEMS.Progress in the knowledge of alloy systems has not previouslybeen reviewed in these Reports. From the viewpoint of themetallurgist, the subject is a vast one, and it has been necessary forthe purpose of this Report to restrict the field to certain aspectswhich are clearly of interest to chemists. The policy adopted hasbeen to disregard a great amount of more or less isolated observationson particular alloys, and to describe what has been done during thelast five years or so on the examination of non-ferrous alloy systemsfrom the constitutional point of view, with special reference to theformation of intermetallic compounds. It is natural, therefore,that mainly binary systems are considered, and that much work ofindustrial importance, of interest to the metallurgist and theengineer, falls outside the scope of the Report.Methods of Investigation.-Although thermal analysis and micro-graphical examination continue to be of fundamental importancein the study of equilibrium diagrams, there is evidence that moreuse is being made of the investigation of other properties of alloysystems as accessory means, and that such wider investigations haveled to the reconstruction of some of the older diagrams.The salientfeatures of these methods have been discussed by H. Scott,lJ.L. Haughton,2 and W. Guertler.3 C. Sykes4 has pointed outthat the standard cooling-curve methods may give unreliable resultswhen applied to transformations involving atomic rearrangement1 First Comm. New Internat. Assoc. Test. Hat., 1930, [A], 339.2 Ibid., p. 316.3 Pmc. Zurich Congr. Intemt. Assoc. Teet. Mat., 1932, 1, 469.4 Proc. Roy. SOC., 1935, [A], 148,422; A., 576166 INORGANIC CHEMISTRY.in a homogeneous solid solution, and has described a modified,double differential cooling-curve method.Outstanding among the more modern methods is the X-raydetermination of alloy equilibrium diagrams. The value of theX-ray mebhod has been emphasised by A. We~tgren,~ inasmuch asit makes possible, not only the determination of the nature of thecrystal structure and dimensions of the crystal lattice of the differentphases existing in the alloy system, but also the fixing of the positionof the phase limits.There is the further advantage that someinsight into the chemical characteristics of the phases is given, whichthe earlier metallographic methods could do only very incompletely.The value of the electrical conductivity method has been discussedby W. Hume-Rothery ; advantages are claimed in the detection ofvery small ranges of solid solubility and in the fixing of phasetransformations which are combined with hysteresis or small thermaleffects. The theory of the electrical conductivity of alloys has beendiscussed,' and the rule that the atomic increase in resistance isgreater as the distance between the elements in the periodic systemincreases has been confirmed for a series of gold alloys.The study of constitution by magnetic measuremente has beenreviewed by A.Kussman,8 G. Grubeyg and E. Vogt.lo The magneticmethod has been used in the study of a series of alloy systems,ll andthe existence of the compounds SnSb, CdSb, ZnSb, and Sn,T1 indicatedby abrupt changes in the slope of the susceptibility-compositioncurves. Y. Shimizu l2 has directed attention to the great effect ofadsorbed gases on magnetic susceptibility results, and it appearsthat the curious results obtained in some of the earlier investigationsare due to the neglect to melt and anneal in a vacuum. Shimizufound that, in alloys which form continuous solid solutions, themaximum deviation of susceptibility from the values calculatedfrom the additive rule is small.With eutectic alloys, with solubilityon both sides, the susceptibility-composition curve is linear in thc?range of eutectic composition and slightly curved in the range ofsolid solution. The magnetic susceptibility-composition curves for5 2. Metallk., 1930, 22, 368; Trans. Amer. Inst. Min. Met. Eng., Inst.Metals Div., 1931, 13; Angew. Chem., 1932, 45, 33; Proc. Zurich Congy,Internat. Assoc. Test. Mat., 1932, 1, 484; B., 1931, 978.8 Metallwirt., 1929, 8, 1243; A., 1930, 1106.7 G. Grube and J. Hille, 2. anorg. Chem., 1930, 194, 179; A., 1931, 158;8 2. Metallk., 1934, 26, 25 ; B., 1934, 407.9 Ibid., 1935, 27, 194.11 F. L. Meara, Physical Rev., 1931, [ii], 37, 467 ; Physics, 1932, 2, 33 ; A.,12 Sci.Rep. T d h h I m p . Univ., 1932, 21, 826; A., 1933, 455.L. W. Nordheim, Physical Rev., 1930, [El, 35, 1430; A., 1931, 1361.lo Ibid., p. 40.1932, 686HEDGES : NON-FERROUS ALLOY SYSTEMS. 167several alloy systems of gold 1, and copper 14 have also beendetermined.A few papers report the application of electropotential measure-ments to the determination of constitution. F. Griengl andR. Baum l5 have constructed isopotential lines for the systemgold-tin-mercury in a triangular phase diagram. Although thecurves give no evidence of AuSn or AuSnz, these compounds areindicated by discontinuities in potential in the binary system,suggesting that they are partly dissociated in mercury. Potentialstudies of thallium-bismuth alloys in fused acetates have beenmade,16 and the existence of Bi,T1 indicated.A. Glazunov hasmade an interesting application of the electropotential niethod tothe determination of the phase structure of metallic coatings byanodic dissolution. Since each phase has its own potential, whichremains constant during its dissolution, the composition and thick-nesses of the phases can be determined from the lengths of theindividual potential values (horizontal positions in the graph) andthe intervals between them, provided the course of the potentialcurves of the binary system in question be known. Thus thepresence of Zn,Fe and probably ZnPe, in hot-dipped zinc coatingson iron has been shown.The interpretation of heats of mixture of molten metals has beenundertaken by M. Kawakami,l* who found that with a few excep-tions heat is evolved when the metals form intermetallic compounds,whilst heat is absorbed when solid solutions are produced.Optical properties have so far played little definite part in theexamination of alloy systems.S. Ueno I9 has shown, however,that the curve of intensity of reflexion of silver-aluminium alloyshas a maximum corresponding with the compound AIAg,, a breakat AlAg,, and a minimum at the eutectic. Similar investigationsof the alloys of aluminium with magnesium20 indicate a highreflectivity for Mg,Al,.The cementation method, in which the constituents are heatedtogether in a vacuum at temperatures near the solidus, has beenapplied to several alloy systems by I;.Loslciewicz 21 and has led tol3 E . Vogt, Ann. Physik, 1932, [v], 14, 1 : A., 1932, 907.1 4 H. F. Seemtmn, 2. Metalllc., 1932, 24, 299.l5 iionatsh., 1932, 61, 330; A., 1933, 127.16 A, Olander, 2. physikal. Chem., 1934, [ A ] , 169, 260; A., 1934, 954.1 7 Trans. Paraday Xoc., 1935, 31, 1262.18 Sci. Rep. Tdholcu Imp. Univ., 1930, 19, 521 ; A., 1932, 296.19 Mem. Coll. Sci. Kydt6 Imp. Ul~i.d., 1930, 13, 141 ; A., 1930, 681.20 J. Wulff, J . Opt. SOC. Amer., 1934, 24, 223; A., 1934, 1086.21 Przeglqd Gorniczo-Hutniczy, 1929, 21, 583 ; Przeglyd Techrziczy, 1930, 09,508168 INORGANIC CHEMIST&Y.suggested modifications of the diagrams. It appears from thiswork that the formation of a eutectic during cementation is possibleonly at temperatures above the melting point of the eutectic.Cementation below the melting point can occur only when thecementing metal has a solid solubility in the metal to be cemented.By similar technique with aluminium alloys, M.Bosshard 22 hasconfirmed the existence of the ternary compounds Al,Pe,Si,,AI,Cu,l?e, A15NiCu2, and Al,Mg,Cu.Structure of the Phases.-Reference has been made to the valueof X-ray analysis in the detection of phases and determination oftheir structure. G. Sachs23 maintains that a transformation inalloys generally consists of two partial processes, vix., change oflattice structure and rearrangement of the atoms, and shows thatthese processes can be studied separately by X-ray investigation.The kinetics of lattice transformation are governed by (1) thermalformation of nuclei and crystal growth (at high temperatures),(2) regular change of the lattice and crystal growth (at moderatetemperatures), or (3) regular change of the lattice alone (at lowertemperatures).The effect of thermal agitation on atomic rearrange-ment in alloys has been discussed by W. L. Bragg and E. J. Wil-liam~.~* The ordered has a lower potential energy than thedisordered structure, but thermal agitation promotes disorder.Order sets in abruptly at a critical temperature, and increases asthe temperature is lowered, becoming complete only at absolutezero. The sudden onset of order causes a sharp inflexion in curvesrelating resistivity, lattice spacing, and specific heat with tem-perature.Such inflexions simulate a phase change, althoughackually there is no such change.The effect of additions of one metal on the lattice structure ofanother has been studied.25 W. Hume-Rothery 26 has shown thatthe addition of cadmium, indium, tin, or antimony to silver changesits lattice structure t o an extent which is proportional to the valencyof the added element. The question whether the change in latticeconstants in the formation of solid solutions depends on the grainsize has beenW. Hume-Rothery2* pointed out that the compositions of the22 Aluminium, 1935, 17, 477.23 2. Metallk., 1932, 24, 241; A., 1932, 1196.24 Proc. Boy. SOC., 1934, [ A ] , 145, 699; A., 1934, 954.25 E. R. Jette, Amer. I n s t . Min.Met. Eng., 1934, Tech. Publ. No. 560, 1-26 Nature, 1935, 135, 1038; A., 919.27 E. Schmid and Q. Siebel, Metallwivt., 1932, 11, 685; A., 1933, 1110; U.28 J . Inet. Metale, 1926, 35, 295; A,, 1926, 356.16; A., 1935, 24.Dehlinger and 9. Wicst, ibid., 1933,12,2; A., 1934, 249HEDGES : NON-FERROUS ALLOY SYSTEMS. 169@-phases of many binary alloys are such that the ratio of valencyelectrons to atoms is approximately 3 : 2. The generalisation hasbeen extended by A. J. Bradley 29 to the y-phases of certain alloys,where the ratio of valency electrons to atoms is found t o be 21 : 13.Similarly, the ratio for a number of &-phases is 7 : 4. The validityof these rules has been confirmed 30 for a number of binary alloys ofcopper, silver, and gold.H. Perlitz 31 points out that, since theratio is always greater than 1, a necessary condition for the existenceof these structures is that one of the metals contributes at least2 valency electrons and the other component contributes not morethan 1 valency electron t o the lattice structure. The possiblesystems of alloys that can be formed in accordance with theserequirements have been worked out.Resulting from the study of a large number of binary alloysystems, Hume-Rothery and his collaborators 32 have been able toshow that, if the atomic diameter of the solute differs by more thanabout 14% from that of the solvent, the solid solution is restrictedto a few atoms per cent.; but when the size factor is favourable asolid solution may be formed, in which the solubility limits generallyobey valency laws, the solubility becoming less as the valencyincreases. Under these conditions the maximum solid solubility isdetermined mainly by the concentration of valency electrons, andthis principle permits the approximate calculation of solubilitylimits in certain ternary alloys.A new approach to the problem of the chemical bond betweenmetals has been suggested by J.D ~ r f m a n . ~ ~ Assuming that theatoms of copper, zinc, aluminium, and tin entering the nickellattice become singly ionised, he calculates that the negative valuesof the magnetic moments of these foreign atoms correspond with thenumber of valency electrons left attached to the corresponding ion.I n an examination of the systems aluminium-copper, antimony-silver, cadmium-tin, cadmium-zinc, copper-silver, and lead-tin,D.Stockdale34 found no evidence for his supposition that in abinary eutectic the atoms of the two elements are present in asimple ratio. Recently, however, he 35 has put forward a series of29 Phil. Mag., 1928, [vii], 6, 878; A., 1929, 125.30 A. Westgren and G. Phragmh, Trans. Paruduy SOC., 1929, 25, 379; A.,1929, 987 ; A. Westgren and W. Ekman, Arkiv Kemi, Min. Cfeol., 1930, [B],10, [ii], 1 ; A., 1931, 900.31 J . Chem. Physics, 1933, 1, 335; Actu Comm. Turtuensis, 1933, [A], 24,(2), 1; A., 1933, 118.92 W. Hume-Rothery, G. W. Mabbott, and K. M. C. Evans, Phil. Trans.,1934, 233, 1 ; A., 1934, 725.33 Nature, 1932,130, 506. 54 J . Ins€. hdetale, 1930,43, 193; A., 1930, 537.3 5 Proc.Roy. Soc., 1935, [A], 152, 81.F 170 INORGANIC CHEMISTRY.empirical rules, which, although not exact, are claimed t o givemore accurate results than the majority of those that have beenobtained experimentally in this connexion. (a) I n a saturatedsolid solution at the temperature of the eutectic or peritectic thereis a simple integral relation between the numbers of solvent andsolute atoms. (2) There is a similar relation in a saturated solidsolution in contact with a second solid phase at the temperature oftransformation of the second phase. (3) I n copper- and silver-richalloys the solubilities a t the above two temperatures are simplyrelated, and it is possible t o predict the solubility at the lowertemperature if that at the higher is known.(4) I n a eutecticmixture the elements are present in a simple integral atomic ratio.( 5 ) In a eutectic mixture the ratio of the numbers of atoms,irrespective of their kind, in the two phases is simple.A different angle on related problems is provided by work onthe atomic heats of alloyed metals in the form of solid solutions orintermetallic compounds.36 In both cases deviations from theadditive rule have been noted. I n a comparison of the heat contentsof 25 intermetallic compounds with the sum of the heat contentsof their components a t different temperaturesY37 an increase in theheat content of the compound was observed in 14 and a decreasein 8 compounds; in the remaining 3 compounds the differencechanged its sign with change of temperature.If in all cases theheat content of the compound were greater than the sum of thoseof the components, the conclusion might be reached that the surplusheat content should correspond with the energy for the vibrationof molecules, Le., the lattice of the intermetallic compounds wouldbe built up of molecules, not of atoms. This is, however, not thecase.The volume changes accompanying the formation of inter-metallic compounds have been studied.38 Published data on themolecular volume of intermetallic compounds and on the atomicvolumes of their component metals seem to show that metals whichcan be strongly compressed mechanically undergo a contractionwhen they enter into an intermetallic compound, whilst metals whichare difficult to compress do not.F.H, Jeffery 39 has applied an interesting thermodynamical36 J. A. Bottema and F. M. Jaeger, Proc. K . Akad. Wetensch. Amsterdam,37 G. Tammann and A. Rohmann, 2. anorg. Chem., 1930, 100, 227; A.,38 W. Biltz, Z . Metallk., 1934, 26, 230; A., 1935, 158.3* Trans. Paraday SOC., 1930, 26, 86, 587, 588; 1931, 27, 136, 137, 188;1932,&8,452, 456, 667, 705; 1933,29, 650; A., 1930, 406, 1360, 1361; 1931,418, 676; 1938, 566, 007, 9S9; 1933,454.1932, 35, 916, 929; A., 1933, 18.1932, 986HEDGES : NON-FEB;ROUS ALLOY SYSTEMS. 171method to the study of a number of binary alloy systems. Hededuces for a two-component, two-phase system consisting of a,solid and a liquid phase, each of which is a dilute solution of onemetal in another, the relation log (1 - n') - log (1 - n) =L/RT - L/RT,, where L is the latent heat of fusion of the solvent,To the freezing point, R and T have their usual significance, n isthe molar fraction of the solute in the liquid, and n' that in thesolid solution.Application of this equation t o solutions of leadin tin, tin in lead, and cadmium in tin shows that in all cases thesolute is monatomic and no intermetallic compounds are present.In the copper-tin system thermodynamic considerations show thatthe a-phase consists, in both the solid and the liquid state, of asolution of Cu4Sn in monatomic copper; the p-phase is a solidsolution of monatomic tin in monatomic copper; the 3-phase inthe solid state is the compound Cu3Sn, but in the liquid state itconsists of a solution of CuaSn in monatomic tin; the €1-phase is asolid solution of Cu,Sn in monatomic tin.When applied to thecopper-zinc alloys, the method indicates that the liquid solutionsin equilibrium with the a- and the @-solid solution consist of CuZn,dissolved in monatomic copper, that the a-solid solution consistsof CuZn, dissolved in monatomic copper, and the p-solid solutionof CuZn, in monatomic copper. The liquid and solid solutions inthe copper-gold and copper-silver systems are derived from mon-atomic molecules of copper and gold or copper and silver,respectively. I n the lead-rich lead-antimoay alloys the existenceof the compound Pb,Sb dissolved in monatomic lead is indicated.The liquid and solid solution phases of copper-magnesium alloysconsist of simple atoms of copper and magnesium; the existence ofMgCu, and Mg,Cu is also confirmed by the thermodynamical method.The thermodynamics of solid-solution alloys has also been treatedby C.Wagner and W. Schottky 40 and by H. S e l t ~ . ~ lAlloys of the Alkali and Alkaline-earth Netnls.-The sodium-potassium equilibrium has been investigated under conditions ofmanipulation that exclude contact with air.42 An unstable com-pound Na,K and a eutectic containing 66 atoms yo of potassiumhave been confirmed. Viscosity determinations of these alloys inthe liquid phase a t 125" show a maximum at the composition&Na, but giveThe solidification diagram of the sodium-rubidium system doesinflexion to indicate the compound Na2K.4340 2.phyaikat. Chem., 1930, [BJ, 11, 163; A., 1931, 157.4 1 J . Amer. Chem. SOC., 1934, 56, 307; A., 1934, 366.42 E. Rinck, Compt. rend., 1933, 19'7, 49; A., 1933, 771.43 R. Kremann, M. Pestomer, and H. Schreiner, Rec. trav. clzirn., 1932, 51,557 ; A., 1932, 800172 INORGANIC CHEMISTRY.not resemble that of the sodium-potassium alloys, and no evidenceof the compound NazRb was found.44 A eutectic occurs a t 75atoms yo of rubidium at - 4.5". The curve tends to becomehorizontal in the vicinity of the composition corresponding withNaRb,. In the sodium-caesium system thermal analysis shows anew compound Na2Cs, and a eutectic at 75 atoms yo of caesium and- 30°.45 The mutual solubilities of sodium and calcium have beendetermined.46 Potassium and rubidium are completely misciblein the solid state; the liquidus and solidus are very close andshow a flat minimum at 32-8", corresponding with 66.6 atoms yo ofrubidium.The system calcium-bismuth forms two compounds, Ca,Bi, andCaBi,.The alloys of this system are attacked by moist air.48Alloys of calcium with smaller amounts of zinc, aluminium, ormagnesium have been prepared as greyish-white powders, which arereadily attacked by exposure to air and moisture.49 The calcium-gold system contains 6 compounds : Au,Ca, Au2Ca, AugCal0, Au,Ca,Au3Ca4, and AuCa2. Calcium is practically insoluble in solid gold,but at room temperature it dissolves 4-5 atoms yo of g0ld.~0The lithium-silver alloys have been examined for the first time,the X-ray method being used.At 500" the metals react violmtlywith evolution of heat. The existence of the compounds AgLi andAgLi, has been established. Since no alteration in the lithium orsilver lattices was observed, it is concluded that solid solutions arenot formed. 51 Similar experiments, coupled with thermal analysis,on the lithium-copper alloys give a, quite different result, for nocompounds or solid solutions are formed.52 Thermal analysis andelectrical-resistance measurements of lithium-cadmium alloys haveconfirmed the existence of LiCd and proved the existence of thehitherto unknown compounds LiCd, and Li,Cd.53 The compoundsLiCd and LiCd, have been confirmed independently by X-rayanalysis,54 but Li,Cd was not detected by this method. The com-pounds LiTl, Li,Tl, Li5T12, Li3T1, and Li,T1 have been detected by44 E.Rinck, Compt. rend., 1933, 197, 1404; A., 1934, 137.45 Ibid., 1934, 199, 1217; A., 1935, 22.4 6 Ibid., 1931,192, 1378; A., 1931, 900.4 7 Ibid., 1935, 200, 1205; A., 693.48 E. Kurzniec, Bull. Acad. polowise, 1931, [ A ] , 31; A., 1931, 1118.49 J. MeyerandR. Goralczyk, 8. angew. Chem., 1930,43,149; B., 1930, 330.F. Weibke and W. Bartels, 2. anorg. Chern., 1934,218,241 ; A., 1934, 838.51 S. Pastorello, Gaxxetta, 1930,60,493; 1931,61,47; A., 1930, 1359; 1931,sa Ibid., 1930, 60, 988; A., 1931, 296.53 G. Grube, H. Vosskiihler, and H. Vogt, 2. Elektrochem., 1932, 38, 869;54 A. Baroni, Atti R. Acmd. Lincei, 1933, [vi], 18, 41; A,, 1934, 137.418.A,, 1933, 18HEDGES : NON-FERROUS ALLOY SYSTEMS.173thermal analysis and electrical-resistance determinations. 55 Ofthese, Li5T1, and Li,Tl can be melted without decomposing. Similarmethods applied to alloys of lithiurn and bismuth show the existenceof Li,Bi and LiBi; 56 the latter is dimorphous.Copper AZZoys.-Certain aspects of the copper-gold alloys havereceived much attention. The transformation of the alloy con-taining 50 atoms % of each metal, from the cubic face-centredlattice with irregular distribution of atoms to the tetragonal latticewith regular distribution, has been studied by X-ray analysis.57With falling temperature the transformation occurs in two steps :(1) change of the lattice symmetry from cubic to tetragonal,(2) transition from an irregular to a regular distribution of theatoms.The first step occurs rapidly and completely, the secondslowly and incompletely. Thus an intermediate state is reached,which is characterised by a considerably increased hardness, tensilestrength, and electrical resistance. 58 In this system the existenceof AuCu and AuCu, has been confirmed.59 A transformation inAuCu, has been studied.60 In addition to the AuCu and theAuh, transformation, J. L. Haughton and R. J. M. Payne 61 havefound another transformation in alloys approximating to thecomposition of Au,Cu,. The conipound Au2Cu, is also indicatedby the electrical conductivity and thermoelectric force curves ofM. Le Blanc and G. Wehner,62 but is denied by W.Broniewski andK. Wesolowski. 63W. R. D. Jones 6* obtained evidence of Mg,Cu and MgCu,, butnot of MgCu. In spite of the old-established use of brasses, theequilibrium diagram of the copper-zinc alloys has not been un-equivocally settled. Examination of numerous properties of awide range of these alloys 65 has confirmed the existence of thecompounds CuZn, CuZn,, and CuZn,, but not Cu,Zn,, Cu,Zn,55 G. Grube and G. Schrtufler, 2. Elektrochem., 1934,40,593 ; A., 1934,1065.5 6 G. Grube, H. VosskiLhler, and H. Schlecht, ibid., p. 270; A57 K. Oshima and G. Sachs, 2. Physib, 1930,63,210; U. Dehlinger and L.Graf, ibid., 1930, 64, 359; A., 1930, 1360; L. Graf, 2. Metallic., 1932, 24, 248;A., 1932, 1196.1934, 724.68 E. Schucb, Metallwirt., 1933, 12, 145; A., 1933, 1238.50 N.S. Kurnakow and N. W. Ageew, J . Inst. Metals, 1931, 46, 481; A . ,60 G. Sachs and J. Weerts, 2. Physik, 1931, 67, 507; A., 1931, 414.61 J . Inst. Metals, 1931, 46, 457; A., 1931, 1224.62 Ann. Physib, 1932, [v], 14, 481 ; A., 1932, 989.63 Compt. rend., 1934, 198, 370; A,, 1934, 248.84 J . Imt. Metals, 1931, 46, 396; A., 1931, 1224.65 W. Broniewski and J. Strasburger, Compt. rend., 1930, 190, 1412; Rev.Aldt., 1931, 28, 19, 70; Pram Zakladu Metall. Pol. Warsaw, 1933, 3, 3; A .1930, 987.1932, 1224174 MOEGANIC CHEMISTRY.Cu1,Zn13, Cu21Zn31, Cu3Zn,, Cu,Zng, or Cu,Zn. On the other hand,C. Rossi 66 claims to have prepared single crystals of cU,Zn3. Thecrystal structures of Cu5Zn8 and Cu,Cd, have been ~ompared.~’The relation between mean atomic volume and composition in thisseries of alloys has been studied.68 Examination of the electricalproperties of the copper-cadmium alloys suggests the existence ofCuCd,, CuZCds, Cu,Cd,, and possibly Cu,Cd.The copper-gallium system has been investigated; 69 it contains7 intermediate phases, including Cu,Ga and CugGa,.The copper-indium system is somewhat similar 70 and contains the compounds&,In, CuJn,, and Cu21n.Further evidencethat the &phase is Cu,,Sn8, not Cu,Sn, has been obtained,72 whilstthe &-phase, formerly written as CuSn, is confirmed as c ~ , S n , . ~ ~It has also been shown that an alloy, identical with Cu,Sn preparedby fusion, is produced by the action of tin on copper sulphatesolutions under certain conditions.74 Equilibria in the ternarysystems of copper and tin with ni~ke1,7~ manganese,V6 lead,77 orberyllium 78 have been investigated.Xilver AZZoys.-X-Ray analysis has confirmed the existence ofAgCa and Ag,Ca, but not of Ag4Ca, Ag,Ca, or AgCa,, formerlyclaimed.79 By the thermal method, the existence of Ag4Sr, Ag5Sr,,AgSr, and Ag,Sr3 and also of Ag,Ba, Ag3Ba2, and Ag,Ba, isindicated. *OThe lattice parameters and densities of solid solutions of aluminiumWork continues on the copper-tin series.716 6 2;. Physsib, 1932, ‘74, 707; A., 1932, 454.6 7 A. J. Bradley and C. H. Gregory, Phil. Mag., 1931, [vii], 12, 143; A.,68 E. A. Owen and L. Pickup, Proc. Roy. SOC., 1933, [A], 140, 179; A.,69 F. Weibke, 2. anorg. Chem., 1934, 220,293; A., 1935,22.7 O F.Weibke and H. Eggers, ibid., p. 273; A., 1935, 22.7 1 J. Vero, 2. anorg. Chem., 1934,218, 402; A., 1934, 953.72 M. Hamasumi and S. Nishigori, Tech. Rep. Tdhoku, 1931, 10, 131: A . ,73 M. Hamasumi, Kinz. no Kenk., 1933, 10, 137.74 H. Kersten and J. Maas, J . Amer. Chem. SOC., 1933, 55, l.002; A., 1933,454.75 J. T. Eash and C. Upthegrove, Trans. Amer. Inst. Min. Met. Eng., 1933,104,221; A., 1933, 119; J. Vera, Mitt. Berg. Hutten. Abt. Hochschule, Sopron,1932, 4, 1.1931, 896.1933, 454.1931, 900.7 6 Idem, ibid., 1933, 5, 1 ; A., 1934, 1301.7 7 Idem, ibid., 1932, 4, 1.7 8 E. S. Rowland and C. Upthegrove, Amer. Tnst. Min. Met. Eng., 1935,79 C. DBgard, 2. Krist., 1935, 90, 399; A., 1198.80 F. Weibke, 2. anorg.Chem., 1930,193,297; A., 1930, 1509.Tech. Publ. No. 613; A , , 1935, 1066HEDGES : NON-FERROUS ALLOY SYSTEMS. 175in silver have been measured.81 The experimental densities areslightly lower than those calculated on the basis of a direct sub-stitution of aluminium for silver a.tioms in the silver lattice, and it isconcluded that the aluminium in the solid solution must be com-bined chemically with the adjacent silver. Re-examination of thissystem by thermal analysis 82 shows three compounds : Ag,Al,Ag3Al2, and Ag,A1.83 The solid solubility of indium in silver is19.4% at room temperature. The system contains the compoundsAg,In, AgJn,, and AgIn,.84 The system silver-praseodymiumcontains Ag,Pr, Ag,Pr, and AgFr.*5 The ternary systems of silverand zinc with copper or aluminium,86 and of silver and antimonywith zinc, cadmium, or copper 87 have been investigated, andsimilar types of alloys have been examined from the viewpoint ofan improved, untarnishable sterling silver.88Magnesium Alloys.-A new diagram for the magnesium-zincalloys89 shows the compounds MgZn,, MgZn,, and MgZn.X-Ray investigation of magnesium-cadmium alloys containing30-80 atoms yo of magnesium has confirmed the existence ofMgCd, and Mg,Cd, but not MgCd,.90 The equilibrium diagram ofthe magnesium-thallium system has been re~ised.~l Earlier inves-tigations had suggested the existence of the compounds T13Mg8 ,TlMg,, and T12Mg,.Of these only TlMg, has been confirmed, theothers appearing to be eutectics. I n addition, T1,Mg5 and TlMgare considered to exist.The alloys with praseodymium includetwo well-gefined compounds, MgPr and Mg,Pr, and probablyMgPr,.92 Electrical-conductivity measurements of magnesium-richalloys with tin show a minimum at the composition corresponding81 R. T. Phelps and W. P. Davey, Asner. Inst. Min. Met. Eng., 1931, Tech.82 F. E. Tischtckenko, Zhur. Obs. Khim., 1933, 3, 549; A., 1934, 21.83 Cf. N. Ageew and D. Shoyket, J . Inst. Metals, 1933, 52, 119; A., 1933,84 F. Weibko and H. Eggers, 2. anory. Chem., 1935, 222, 145; A., 576.as G. Canneri, Met. Ital., 1934, 26, 794.86 S. Ueno, Mern. Coll. Sci. Kyat6 Imp. Univ., 1929, 12, 347; 1930, 13, 91 ;A., 1930, 284, 535.87 W. Guertler and W. Rosenthal, Z. Metallk., 1932, 24, 7, 30; A., 1932,455.88 K.W. Ray and W. N. Baker, Ind. Eng. Chem., 1932, 24, 778; B., 1932,845; L. Guillet, A. Petit, and J. Cournot, Rev. Mkt., 1932, 29, 113, 183; B.,1933, 393.89 A. A. Botschvar and I. P. Velitschko, 2. anorg. Chem., 1933, 210, 164;A., 1933, 219.90 U. Dehlinger, ibid., 1930,194, 223; A., 1931, 167.9 1 G. Grube and J. Hille, 2. Elektrochem., 1934,40,101; A , , 1934, 356.92 G. Canneri, Met. Itat., 1933, 25, 250; A., 1934, 483.Publ. No. 443, 1.1110176 INORGANIC CHEMISTEY.with Mg,Sn.93 The magnesium-antimony system contains onlyone compound Mg,Sb,, which is dimorphous, having a transforma-Lion temperature a t 930" & 2O.94 The solid solubility of nickel inmagnesium is less than O*l% .95 Ternary systems investigatedinclude magnesium-aluminium-copper,g6 magnesium-zinc-silicon,97and magnesium-zinc-calcium.98Mercury AZZoys.-E&dence that dilute liquid amalgams arecolloidal sols in which mercury acts as dispersion medium has beenobtained in amalgams of sodium,99 copper and silver,l and iron.2Thermal analysis of the rubidium-mercury system indicates thecompounds Rb,Hg,, RbHg,, Rb2Hg,, Rb5Hgls, Rb,Hgg, RbHg,,RbHg,, and Rb,Hg,.3 Comparison with other mercury-alkali metalsystems shows that the affinity for mercury increases as the atomicweight of the alkali metal increases, and the compounds formedare more complex.The constitution of silver amalgams has beendetermined ; 4 X-ray investigation shows the presence of Ag,Hg,.5The compounds Li,Hg, Li,Hg, Li2Hg, LiHg, LiHg,, and LiHg,have been detected in lithium amalgams by thermal analysisand confirmed by x-ray^.^ Only Li,Hg and LiHg can be meltedunchanged.Although thermal analysis and microscopical examina-tion have given PO evidence of compound formation in cadmiumamalgams, X-ray analysis indicates that Cd,Hg exists.*Determinations of the parachor of thallium dissolved in mercuryshow that it is not in the monatomic state, but is combined withmercury or other thallium atoms t o form polyatomic molecu1es.gThe solid thallium amalgams have been examined by ihe X-ray93 G. Grube and H. Vosskuhler, 2. Elektrochem., 1934, 40, 566; A., 1934,1065.9 p G. Grube and R. Bornhak, ibid., 1934,40, 140 ; A., 1934, 590.Ofr J. L. Haughton and R. J. M. Payne, J .Inst. Metals, 1934, 54, 275; A . ,1934, 590.9 6 A. Portevin and P. Bastien, Compt. rend., 1932, 195, 441 ; A., 1932, 989;Ghim. et I n d . , 1934, Special No., 490; A., 1934, 725.9 7 E. Elchardusand P. Laffitte, Compt. rend., 1933,197,1125; A., 1934,23.98 R. Paris, ibid., 1933, 197, 1634; A., 1934, 138.99 G. R. Paranjpe and R. M. Joshi, J . Physical Chem., 1932, 36, 2474; A . ,1932,1196; H. E. Bent, ibid., 1933,37,431; A., 1933, 561.R. Kijhler, Kolloid-Z., 1933, 64, 200; A., 1933, 895.N. M. Tchuiko, Ukrain. Chem. J., 1931, 6, 229.W. Biltz, F. Weibke, and H. Eggers, 2. anorg. Chem., 1934,219,119; ,4.A. J. Murphy, J . Inst. Metals, 1931, 46, 507; A , , 1931, 1224.G . Grube and W. Wolf, 2. Elektrochem., 1935,41,675; A., 1314.N. W. Taylor, J .Amer. Chem. SOC., 1932, 54,2713; A , , 1932, 989.1934, 1064.j A. Weryha, 2. Krist., 1933, 86, 335; A., 1934, 16.'J E. Zintl and A. Schneider, ibid., 1935, 41, 771.9 L. Belladen and A. Triolo, Gazzetta, 1934, 64, 461HEDGES : NON-FERROUS ALLOY SYSTEMS. 177metliod.10 There is evidence that many rare-earth metal8 formcompounds with mercury of the type MHg4.11The solubility of iron in mercury at 20" is 0.00007% ; and thatof nickel 0.00014% (by wt.).12 The solubility of gold in mercuryat 280-400° has been determined ; the gold-rich compound isprobably A~~Hg.13 An X-ray investigation of alloys of mercurywith gold, silver, or tin has been carried 0 ~ t . 1 4 The system mercury-manganese-tin has been studied.15A considerable number of binary, ternary, and quaternary com-pounds of metals such as copper, tin, zinc, and iron with or wibhoutmercury have been obtained by A.R. Russell and his collaborators l6by reaction in mercury.A technique for determining the vapour pressure of amalgamshas been described and applied to amalgams of the alkali meta1s.l'AZuminium Alloys.-Much of the work on aluminium alloys con-cerns their properties for structural purposes and is of interestmainly to metallurgists and engineers. A summary of availableinformation on the constitution of binary alloys of aluminium hasbeen given by E. T. Richards.lsThe solubility of sodium in molten aluminium has been determined ;there is no evidence of solid solubility,lQ nor has a solid solubilityof barium in aluminium been detected.20 X-Ray examination hasestablished one compound A1,Ba in the aluminium-barium system.21The solid solubilities of beryllium 22 and magnesium 23 in alumin-ium have been determined.An investigation of the aluminium-magnesium alloys shows that the two metals are mutually solublein the solid state and can form the compounds AI,Mg,, AlMg, andA13Mg4.24 G. Wassermann 25 has shown that the lattice parameterlo A. Olander, 2. phyeikal. Chem., 1934, [ A ] , 171,425; A., 1935,440.l1 P. T. Daniltchenko, Zhur. Obs.Khinh., 1931, [ A ] , 163,467; A., 1931, 1381.l2 E. Palmaer, Z . Elektrochem., 1932, 38, 70; A , , 1932, 330.l4 S. Stenbeck, Z. anorg. Chem., 1933, 214, 17; A., 1933, 1006.l5 A. N. Campbell and H. G. Carter, Trana. Paraday Soc., 1933, 29, 1295 ;A., 1934, 138.l6 Nature, 1930, 125, 89; 1934, 133, 217; J., 1932, 841, 852, 857; 1934,1750; A., 1930, 177; 1932, 456, 1083; L934, 265.1 7 J.S. Pedder and S. Barratt, J., 1!)33, 537; A., 1933, 669; H. H. vonHalban, jun., Helv. Phyaica Acta, 1935, 8, 65.la Metallb6rae, 1935, 25, 498, 530, 562, 721, 1041.l9 E. Scheuer, 2. Metallk., 1935, 27, 83; A., 928.2o E. Alberti, ibid., 1934, 26, 6; A . , 1034, 482.21 K. R. Andreas and E. Alberti, ibid., 1935, 27, 126 ; A., 1065.22 M. Haaa and D. Uno, ibid., 1930, 22,277; B., 1930, 1072.23 E. Schrnid and G. Siebel, ibid., 1931, 23, 202; B., 1931, 978.24 M. Kawakami, Kinzoku no Kenkyu, 1933,10,532.25 Z . Metalllc., 1930, 22, 158; B., 1930, 717.J. T. Anderson, J . Physical Chem., 1932,36,2145178 NORUANIC CHEMISTRY.of aluminium is increased by about 0-0045 fi.for every 1 atom yoof magnesium added, and by about 0.001 A. for every 1 atom %of zinc. Purther X-ray work indicates that the supposed compoundA12Zn, does not exist.26N. A. Pushin and V. Staji627 claim to have detected the com-pounds A12Ga, AlGa, and AlGa,, but this claim is not substantiatedby the experiments of E. Jenckel.28 In the system aluminium-praseodymium29 the compounds PrAl, PrAl,, and PrAl, are saidto exist, the last in two modifications. A revision of the diagramfor aluminium-rich aluminium-titanium alloys has been under-taken, and the compound TiAl, confirmed.30 It has been shown 31that variations in casting temperature and rate of cooling determinewhether the TiAl, in these alloys separates out as needles or remainsso highly dispersed that the alloy behaves as a solid solution.The solid solubility of antimony in aluminium is less than 0*1% ; 32only one compound, AlSb, is formed in this system.The compounddecomposes in moist air.33Dissolution in dilute hydrochloric acid of an aluminium alloycontaining 1 4 % of chromium leaves a residue of rhombohedra1plates of CrA4.34 In the aluminium-manganese system, A. J.Bradley and P. Jones 35 detected by X-rays the compounds Al,Mnand A1,Mn. Since small amounts of iron profoundly affect theequilibria, this system has been re-examined, metals of very highpurity being used.36 The solid solubility of manganese in aluminiumwas determined, and the existence of Al,Mn and Al,Mn established.An investigation of the system aluminium-cobalt 37 shows that thesolid solubility of cobalt in aluminium is very small.A eutectic at1.45% of cobalt was observed. When the sluminium-rich alloys2e E. Schmid and G. Wassermann, 2. Metullk., 1934, 26, 146; A., 1934,1064.2 7 Z. anorg. Chern., 1933, 216, 26; A., 1934, 138.28 Z. Metalllc., 1934, 26, 249.29 G. Canneri, Alluminio, 1933, 2, 87.30 W. L. Fink, K. R. van Horn, and P. M. Budge, Amer. Inst. Min. Met. Eng.,31 H. Bohner, 2. MetallE., 1934, 26, 268.32 E. H. Dix, jun., F. Keller, and L. A, Willey, Amer. I n s t . Min. Met. Zng.,33 J. Veszelka, Mitt. Berg. Hiitten. Abt. Hochschule, Sopron, 1931, 3, 193;34 W. L. Fink and H. R. Freche, Trans. Amer. Imt.Min. Met. Eng., Inst.35 Phil. Mag., 1931, [vii], 12, 1137; A., 1932, 116.36 E. H. Dix, jun., W. L. Fink, and L. A. Willey, Trans. Amer. I n s t . Min.37 W. L. Fink and H. R. Freche, Amer. Inst. Min. Met. Eng., 1932, Tech.1931, Tech. Publ. No. 393, 1; A . , 1931, 676.1930, Tech. Publ. No. 356, 1; A., 1931, 158.A . , 1932, 1082.Metals Div., 1933, 104, 325.Met. Eng., Inst. Metals Div., 1933, 104, 335.Publ. No. 473, 1 ; A., 1932, 683HEDGES NON-FERROTTS ALLOY SYSTEMS. 179are treated with dilute acids, a residue having the formula Co,Al,is obtained.Ternary systems, the constitutioiis of which have been inves-tigated, include aluminium-magnesium-silicon,3~ aluminium-copper-~ilicon,~~ aluminium-iron-silicon,40 aluminiurn-antimony-rn~gnesium,~l aluminium-silver-rnagnesiumY*2 and aluminium-ni~kel-tin.~~Tin AZZoys.-The equilibrium diagrams of all the known binaryalloys of tin have been compiled and annotated by E.S. Wedgesand C. E. Homer.44 X-Ray analysis of the tin-gold system con-firms the existence of AuSn, AuSn,, and A U S ~ , . ~ ~ Thermalinvestigation of tin-lithium alloys indicates the existence of SnLi,,Sn,Li,, and Sn,Li; X-ray analysis confirms the first two, but notthe last .46 Other thermal and electrical conductivity measure-ments in this system demonstrate the compounds SnLi,, Sn,Li,,Sn,Li,, SnLi2, SnLi, and Sn,Li. The solid solubility of lithium intin is inappre~iable.~' Tin-barium alloys have been prepared byelectrolysis of eutectic mixtures of barium and potassium chlorideswith a molten tin cathode.48 The system contains the compoundsSn,Ba and Sn5Ba. Tin-strontium alloys have been prepared by asimilar method, and the compounds Sn,Sr and Sn,SrThe tin-arsenic alloys have been investigated by X-rays.50There is a wide range of solid solubility at both ends of the series;the compound SnAs exists, but not Sn,As,. Several X-ray investi-gations of the tin-antimony system have been made.51 The38 L. Losana, Met. Ital., 1931, 23, 375; A., 1932, 907; E. H. Dix, jun., F.Keller, and R. W. Graham, Amer. Inst. Min. Met, Eng., 1930, Tech. Publ. No.357, 1 ; A., 1931, 158.39 G. G. Urazov, S. 9. Pogodin, and G. M. Zomornev, illin. Syrie i Zvet.Metally, 1929, 4, 160.40 V. Fuss, 2. Metallk., 1931, 23, 231 ; B., 1931, 929.41 E. Loofs-Rassow, Hausxeit. V.A.1V.u.d. Eqtwerk A.G. Alunvinium, 1031,42 B. Otani, Kinxoku no Kenkyu, 1933,10, 262.43 S. Kato, Suiyokai-Shi, 1931, 6, 529 ; A., 1932, 567.44 Internat. Tin Res. and Dev. Council, 1935, Tech. Publ. B., No. 2, 1-90;45 S. Stenbeck and A. Westgren, 2. physikal. Chem., 1931, [B], 14, 91 ; A . ,413 A. Baroni, Atti R. Accad. Lincei, 1932, [vi], 16, 153; A., 1933, 18.4' G. Grube and E. Meyer, Z. Elektrochem., 1934, 40, 771 ; A., 1935, 23.48 K. W. Ray, Metals and Alloys, 1930, 1, 314; A., 1930, 681.50 W. H. Willott and E. J. Evans, Phil. Mag., 1934, [vii], 18, 114; A.,1934, 953.51 W. M. Jones and E. G. Bowen, Nature, 1930,126, 840; Phil. Maya., 1931,[vii], 12, 441; A., 1931, 33; K. IwasB, N. Aoki, and A. osawa, Sci. Rep.Tdhoku Imp. Univ., 1931, 20, 353; A., 1931, 1364.3,20; B., 1931, 1143.A., 1065.1931, 1223.Idem, I n d . Eng. Chem., 1930, 22, 519; B., 1930, 866180 INORGANIC CHEMISTRY.compound formed is definitely SnSb, not Sn,Sb,. Single crystalsof this compound have been obtained.52 The tin-bismuth systemhas been examined by the X-ray method.53 A marked effect ofbismuth in reflning the grain size of tin has been noted.54The solid solubility of silver, copper, or nickel in tin is very ~mall.~5No intermetallic compound has been found in the tin-cadmiumsystem. 56Re-examination of the tin-iron system 57 gives no evidence ofFe,Sn, but shows the compounds Fe,Sn, FeSn, and FeSn,. Thesecompounds have been confirmed independently by W. F. Ehret andA. F. Westgren 58 and by W. D. Jones and W. E. H ~ a r e . ~ ~ FeSn,,but no other tin-iron compound, occurs in tin-plate.60Lead AZloys.-With calcium, the compounds Pb3Ca and PbCaare formed.61 A complete equilibrium diagram for the lead-lithiumsystem has been established ; the compounds PbLi, Pb,Li,, PbLi,,Pb,Li,, and PbLi, are formed.62Lead and germanium do not form solid solutions or compound^.^^No compound of lead with arsenic is formed, and the solid solubilityof arsenic in lead is only about 0.01% a t room temperat~re.~~There is no evidence for the formation of a solid solution of antimonyin lead, or of any definite compound in the series.65An outstanding achievement in lead alloys is the improvementof mechanical properties and corrosion resistance of lead by theaddition of a small amount of tellurium.66 E. S. H.S. R. CARTER.E. S. HEDGES.W. WARDLAW.G2 H. S. van Klooster and M. 0. Debacher, Metals and Alloys, 1933, 4, 23 ;A., 1933, 344.53 D. Solomon and W. M. Jones, Phil. Mag., 1931, [vii], 11, 1090; A., 1931,676.54 A. A. Botchvar and N. E. Merkurjew, 2. anorg. C'hem., 1933, 210, 161 ;d., 1933, 219; D. Hanson and E. J. Sandford, J . Inst. Metals, 1935, 56, 191.55 D. Hanson, E. J. Sandford, and H. Stevens, ibid., 1935, 55, 115.5 6 D. Hanson and W. T. Pell-Walpole, ibid., 1935, 56, 165 ; A., 1936, 440.57 C. A. Edwards and A. Preece, J . Iron Steel Inst., 1931, 124, 41; B.,58 J . Arner. Chern. Soc., 1933,55, 1339; A., 1933, 562.59 J . lron Steel Inst., 1934, 129, 273 ; A., 1934, 724.6o W. E. Hoare, ibid., 1934, 129, 253; B., 1934, 581.61 R. R. Syromiatnikov, Metullurg., 1931, 6, 466.62 G. Grube and H. Klaiber, 2. EZektroch,ern., 1934, 40, 745.63 T. R. Briggs and W. S. Benedict, J . Physical Chem., 1930, 34, 173; A.,64 0. Bauer and W. Tonn, 2. Metallk., 1935, 27, 183.65 D. Solomonand W. M. Jones, Phil. Mug.,1930, [vii], 10,470 ; A., 1930,1359.6 6 W. Singleton and B. Jones, J . Inst. Metals, 1933,51, 71 ; B., 1933, 351.1931, 1053.1930, 284