INORGANIC CHEMISTRY.MANY papers have been published on orthodox inorganic chemistryduring the last, year, and several of these awaken more than a pass-ing interest. There have appeared, however, four papers by Aston,Rutherf ord, Harkins, and Wendt and Landauer, which outshineall others in importance, for without question they bid fair to revo-lutionise the fundamental conceptions of chemistry. Whilst bysome the signs of the impending change have been recognised, themajority of chemists must now awaken to the fact that a newphilosophy is beicg born. The brilliant discoveries of Soddy andFajans of the existence of isotopes mark the first real step after thediscovery of the production of helium in the radioactive disintegra-tion of atoms. About the same time Thomson, Collie, Patterson,and Masson stated that helium and neon are produced in hydrogen-filled vacuum tubes under the influence of a powerful electric dis-charge.Very soon afterwards appeared the Harkins theory thatall elementary atoms are built up either of helium atoms or ofatoms of helium and hydrogen. Last year the next step wasgained when Rutherford succeeded in disrupting the atom ofnitrogen.It may now be said that the whole story is practically complete,and a wonderful story does it prove to be. Perhaps the moststartling of all the new knowledge gained is that on the oxygenstandard all atomic weights, with the exception of hydrogen, areexact whole numbers, and that the fractional values we haveaccepted as the resnlt of highly accurate work are merely fortui-tous statistical averages due to a mixture of two or more isotopes.Whilst this has been proved by experiment, it also is a necessarycorollary of the theories of atomic structure.In the annual Reportfor 1917 reference was made to Harkins’ theory that all elementaryatoms are built up of helium atoms or helium and hydrogen atom0This theory has now been published in its complete form, and itcarries conviction in its train. An essential feature is that thehydrogen isotope H, plays an integral part in atomic structure, thatit has a definite power of existence, and that very probably it i28 ANNUAL REPORTS ON THE PROGRESS. OF CHEMISTRY.identical with the nebular material called nebulium. First de-tected by Thomson, then more fully confirmed by Aston, H, hasnow been prepared from hydrogen.Then, again, Rutherford has shown that by the disruption of theatoms of oxygen and nitrogen an element of mass 3 is produced,which, however, is an iso€ope of helium.Rutherford considers thatthe atom of mass 3 which enters into the nuclear structure of atomsis this helium isotope and not H, as Harkins assumes.Whichever view may prove to be correct, there can now be littledoubt that all elementary atoms are built up from helium or fromhelium and atoms of mass 3, and, moreover, it is accepted by thenew school that helium itself is built up from four atoms of hydro-gen. The added importance of Collie’s work on the formation ofhelium and neon in hydrogen-filled vacuum tubes is manifest, forit has now become an obvious result from the new theories.Another most interesting aspect of this new knowledge is thatthe synthetic process whereby our elements are known to beproduced during the life history of the stars from the originalnebulium by way of hydrogen and helium can now be understood.It ic; diflicult to write of these discoveries and theories in a calmand measured fashion.They are so great in their achievement, sostupendous in their meaning, and so subversive in their effect thatsome enthusiasm may perhaps be allowed to him who records them.Strange it is that after all these years the old hypothesis of Proutshould rise triumphant, for, in a word, it is this that has occurred.In the Report for 1914, when the discovery of isotopes and Collie’swork had been announced, the writer ventured t o write the follow-ing words: “As did his forefathers of pre-Avogadro days, so alsodoes he (the chemist of today) now await that great generalisationwhich shall co-ordinate and link up all the threads to found a newphilosophy.Radioactivity, enhanced line spectra, the intra-stellarelements, zctive nitrogen and oxygen, atomic disintegration,atomic-weight variation, all ill be unified and embodied in thenew philosophy of the twentieth century. Then will a new chem-istry in its greater meaning emerge as a phcenix from the glowingparental fires of the many chemistries of to-day.”Little apology is needed for making this quotation, since theprophecy seems to be almost complete in its fulfilment.Atomic Theory.In the Reports for 1913 and 1914 reference was made to the workof Thomson, of Collie and Patterson, and of Masson on the pro-duction of helium and neon from hydrogen a t low pressures undeINORGANIC CHEMISTRY. 29the influence of the electric discharge.Negative results wererecorded by Strutt and by Merton, but Collie, using Merton's ownapparatus, obtained definite evidence of the formation of boththese gases. Some further experiments have been carried out dur-ing this year, and once more negative results have been obtained.1I n view of the fact that Collie himself more than once obtainednegative results when using different induction coils, the writersuggested that the explanation of the divergence of the resultsobtained by different observers is t o be found in the fact that aparticular type of discharge is necessary.Piutti and Cardoso,whilst admitting that our rudimentary knowledge does not permitus to discuss this explanation, point out that their resultsstrengthen the probability against it. They say that as in thesomewhat analogous case of active nitrogen, where considerabledivergence of opinion existed, if would be advisable that joint workbe carried out systematically in order definitely to settle thisimportant question.There is little doubt Chat the trend of recent ideas wi1.l createless antagonism t o the formation of helium and neon in vacuumtubes than was the case six years ago. The work of Rutherford onthe disintegration of nitrogen and oxygen atoms has underminedthe old confidence in the immutability of the atom.On the otherhand, all other experimental work has been in the direction of thedisruption of atomic nuclei, whilst Collie's work means a synthesisof atomic nuclei heavier than the pirent hydrogen.There can be no question that on0 of the most complete theoriesadvanced as regards the structure of atomic nuclei is that byHarkins.2 His earlier papers were reviewed in the Report for1917. *4ccording to this theory, the elements are of two kinds,namely, those of even atomic number, the atomic nuclei of whichare composed of helium nuclei alone, or helium nuclei together withcementing electrons, and those of odd atomic number, the nuclei ofwhich are composed of helium and hydrogen nuclei together withcementing electrons.Further, the helium nucleus consists of fourhydrogen nuclei, together with two cementing electrons, the loss ofmass being due t o the packing effect. The helium nucleus is themost stable configuration of all, whilst next in order of stabilitycomes the group of atoms or even atomic number. An interestingfact arises in connexion with the number of hydrogen nuclei whichare associated with the helium nuclei in the second class of elements.I n the case of the lighter elements with odd atomic number thisnumber is always three save in the exceptional case of nitrogen,A. Piutti and E. Cardoso, J . Chim. phys., 1920,18, 81 ; A., ii, 311.W. D. Harkins, Physical Rev., 1920,15, 73 ; A., ii, 47930 ANNUAL REPORTS ON TRE PROGRESS OF CHEMISTRY.where it is two.The extremely frequent occurrence of this groupof three hydrogen nuclei suggests that it probably occurs alone as aunit with a nuclear charge equal to 1 and atomic weight of 3, andtherefore structurally it will be an isotope of hydrogen. If thehypothetic nebulium exists a t all it is probably this form of hydro-gen, and it is interesting that from a study of the Diippler effectthe atomic weight of this element has been found to be about 3.3Now there is one point in connexion with the Harkins theorywhich requires consideration. If; for example, the elements witheven atomic numbers are formed from helium nuclei, why is it thatthey are not inore unstable in view of the fact that the heliumatom is thz most stable form? It would seem necessary to con-clude that the elements are metastable, and that they are able toexist owing t o their possessing an external force field. I f this isbroken by the supply of energy, then the atomic nucleus willbecome unstable.I f this principle of external fields is accepted,then it only becomes a question of supplying the right amount ofenergy to the hydrogen atom for the association to become possibleof three or four nuclei to form H3 or helium. On the Harkinstheory, therefore, there is no reason against the production of H3and helium 111. vacuum tubes from hydrogen if the discharge em-ployed produces the suitable type of energy. Indeed, such aphenomenon is rather to he expected than denied in view of thestability of the helium nucleus.The writer is therefore all themore encouraged to insist on the correctness of his suggestion madein 1914 that the contradictory results obtained by Thomson, Collie,Patterson, and Masson on the one hand, and by Strutt, Piutti andCardoso on the other, are due to the absence of sufficient energy ofthe right kind in the latter and negative experiments. There aretwo alternative possibilities as to the nature of the energy requiredto break open the fields of the hydrogen atom. It may either beradiant energy of short wave-length or it may be energy given byrapidly moving particles. The production of either of these in agiven vacuum tube varies remarkably with the conditions. Theimportance of this work has undoubtedly increased, and it is amatter of some moment that the question as to the production ofhelium from hydrogen be decided.Reference may here be made t o a branch of investigation which,although not chemical, must possess great interest for the inorganicchemist, namely, stellar development.According to the modernviews of astro-physicists there is little doubt that in the stars adevelopment process is taking place whereby the chemical elementsare being synthesised from hydrogen and helium as parents. Now3 C. Fabry and H. Buisson, Astrophys. J., 1914,4, 256INORGANIC CHEMISTRY. 31i t would seem fairly certain from a study of the spectra and rota-tional velocities of certain nebulae, particularly the one in Orion,that the original inaterial from which the synthetic process startsis nebulium, which as the first stage in the process forms hydrogenand helium. When it was discovered that the probable atomicweight of this gas is 3, it appeared somewhat incomprehensible thata synthetic process should give both hydrogen and helium.In allprobability, on the basis of Harkins' theory that nebulium is H, thefirst stage is the formation of hydrogen, which then associates togive heliiini, which in its turn associates to give elements of evenatomic numbers. I f this is so, by far the greatest amount of con-densation will take place in the direction of the elements of evenatomic numbers. The great predominance of elements of this classhas been pointed out by Harkins, who offers two explanations ofthe relative scarcity of the elements with odd atomic numbers whichconsist of helium and N, atoms.First, their scarcity may be duets their relative instability, and secondly, there may have beenpresent during the synthetic process relatively little H3. The firstalternative is unsatisfactory, for a t present there seems little, if any,direct evidence that the elements of odd atomic numbers are lessstable than their fellows. The second alternative fits in very wellwith the present suggestion, since, if the first stage is the pro-duction of hydrogen from H,, and the second stage is the formationor" helium from the hydrogen, it is probable that in any laterassociat'ion there will be present only small amounts of H,. TheHarkins theory would therefore fill an undoubted gap in thetheories of stellar development.An important paper has appeared during the year on the massspectra, or positive ray spectra, of the elements by Aston, whodescribes his apparatus in detail and the most recent results hehas ~ b t a i n e d .~ The principle of the method consists in producingthe positive rays with a given element and passing them throughslits. The rays also pass through an electric field and a magneticfield, and then impinge upon a photographic plate. A focussedspectrum is obtained in which the lines depend solely on the ratioof mass to charge. By varying the strengths of the two fields, anydesired line may be brought on to the centre of the plate.All themeasurements of the positions of the various lines are relative, andso one element must be taken as standard, and for this purposeoxygen was selected. The molecule of oxygen carries one charge,whilst the atoms carry one or two charges, with the result that withthis gas three lines are obtained. The three lines are obtained a tthe scale readings 32, 16, and 8 respectively. Direct comparison4 F. W. Aston, Phil. Mag., 1920, [vi], 39, 611 ; A., ii, 34432 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.of the carbon, carbon monoxide, and carbon dioxide lines withthese standards gave C++ (6), C (la), CO (28), and CO, (44).Clearly, therefore, the whole number relation and the additive lawhold within the iimit of accuracy, which is one part in a thousand.The following results have been obtained with eleven elements.Neon, with an atomic weight of 20’2, gives two well-defined lineswhich correspond with masses 20 and 22 respectively.This gas,therefore, consists of two isotopes, with possibly a third, of mass 21,since there was observed a very faint line in this position.Chlorine shows four very definite lines, corresponding with masses35, 36, 37, and 38, with no indication whatever of a line corre-sponding with its atomic weight of 35-46. There is no escape, there-fore, from the conclusion that chlorine is a mixture of isotopes, andthat two of these have masses 35 and 37. Whilst the lines 36 and38 may be due to two more isotopes, it is more probable that theyare given by the hydrogen conipounds of the two isotopes withmasses 35 and 37.Strong lines were also observed a t 63 and 65,due, no doubt, to the carbonyl compounds of the two isotopes.Again, if ordinary chlorine of average atomic mass 35-46 is amixture of two isotopes 35 and 37, it is evident that the line of35 should be stronger than the line of 37, and this was actuallyfound to be the case. A faint line was distinguishable a t 39, whichpossibly is due to a third isotope.Argon shows three strong lines at 40, 20, and 13.33, which clearlycorrespond wikh particles of mass 40, carrying 1, 2, and 3 chargesrespectively. A faint companion was seen a t 36, which is doubtlessdue to an isotope present in small amounts. Tbe presence of about3 per cent. would account for the fractional atomic weight deter-mined from the density.Nitrogen gives a line which cannot be distinguished from that ofcarbon monoxide, and a second line a t 7, due to a doubly chargedparticle. Evidently, therefore, no isotope is present and nitrogenis a pure element.The measurements with hydrogen were more troublesome, owingto the fact that the position of the lines is so far removed from thereference standards. The difficulty was surmounted by comparinghelium with the doubly charged atoms of oxygen and carbon (8 and6), T’homson’s H, with carbon and helium, and hydrogen withhelium.T‘he results show definitely that both hydrogen and heliumare pure elements, and that the mass of the helium atom is 4. Themean value for the mass of H, is 3.026, and that for the mass of thehydrogen molecule is 2.015.The atomic mass of hydrogen, there-fore, is clearly 1.008, and the nature of the H, molecule is settledbeyond questionINORGANIC CHEMISTRY, 33Krypton was found to exhibit perfectly definite evidence of beinga mixture of five isotopes of masses 30, 82, 83, 33, and 86, with aprobable sixth of mass 78. Measurements of these lines were madewith singly, doubly, and trebly charged particles. There wouldseem, 8140, to be five isotopes present in xenon, with masses 128,130, 131, 133, and 135, but as only a minute quantity of this gaswas available these results are only provisional.Mercury was also found to be complex, for the lines observedindiczte the presence of a strong component 203, and a weak one204.There is also a strong band from 197 t o 200, indicating three orfour more componenls, but up t o the present this band has notbeen resolved.Perhaps the most important generalisation that can be made from.this work is the quite remarkable fact that with the exception ofH, and €I, all masses, atomic and molecular, elementary and com-pound, so far measured are whole numbers within the accuracy ofexperiment. The number and variety of substances studied makethe probability of this being true for all elements extremely great.It certainly allows of hypotheses being put forward of atomicstructure far simpler than those which attempted t o explain frac-tional atomic weights, since these now appear t o be merely for-tuitous statistical effeciis, due to the relative quantities of theisotopic constituents.Thus it may now be supposed that an ele-mentary atom of mass M may be changed to one of mass M + 1 bythe addition of a positive particle (H) and an electron. I f bothenter the nucleus a n isotope results, for the nuclear charge is un-altered. I f the positive particle alone enters the nucleus, anelement of the next higher atomic number is formed. When bothforms of addition give a stable configuration the new elements willbe isobares.Apart from the intrinsic value of Aston’s work, its importancebecomes very pronounced when considered along with theories ofthe nuclear structure of atoms. These lead undoubtedly tointegral values of atomic weights, and Harkins explains thedivergence from whole numbers by the existence of isotopes.These isotopes have now been shown by Aston t o exist, and it isof interest to note that Harkins has obtained evidence of theseparation of chlorine into two isotopes by diffusion experimentswith hydrogen chloride.5On the other hand, Rutherford6 has published further experi-mental data which, to a certain extent, do not fit in with Harkins’theory.When the swiftly moving particles from radium4 passW. D. Harkins, Science, 1920, 51, 289.6 (Sir) Ernest Rutherford, Proc. Roy. SOC., 1920, [A], 97, 374 ; A., ii, 541.REP.--VOL. XVII. 34 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.through nitrogen, some of the atomic nuclei of this gas are dis-rupted, and, as is now well known, hydrogen atoms are produced.Hydrogen atoms are not produced in oxygen under the same con-ditions.It is found, however, that both oxygen and nitrogengive slower moving particles of mass 3 with charge 2. Thenitrogen nucleus, therefore, can be disintegrated in two ways, oneby the expulsion of the hydrogen atom, and the other by theexpulsion of an atom of mass 3 carrying two charges. Since theseatoms of mass 3 are five to ten times as numerous as the hydrogenatoms, it appears that these two forms of disintegration areindependent and not simultaneous. It would follow also that thenew atom when it has gained two electrons should have physicaland chemical properties very nearly identical with those of helium,but with mass 3 instead of 4.The spectra of helium and thisisotope should be nearly the same, but, on account of the markeddifference in the relative masses of the nuclei, the displacement ofthe lines should be much greater than in the case of the isotopesof heavy elements like lead. It is very improbable that this isotopeis connected with nebulium.I n dealing with the nuclear constitution of the lighter elements,Rutherford naturally assumes that the new helium isotope formsan integral part of these nuclei. Thus he suggests that the carbonatom consists of four atoms of the helium isotope and that thenitrogen atom consists of four of these isotopes and two hydrogenatoms, whilst the oxygen atom is built up of four helium isotopesand one helium atom. It will be seen at once that there is anessential difference between this view and that put forward byHarkins, who considers that the carbon and oxygen atoms consistof three and four atoms, respectively, of ordinary helium.Now there seems no doubt that the helium isotope discovered byRutherford is a different entity from H,, which forms an integralpart of Harkins' theory, was first discovered by Thomson, nowconfirmed by Aston, and has recently been directly prepared bythe activation of hydrogen.7 Aston has definitely shown that H3carries one charge, and this fact, considered along with its form-ation from hydrogen, shows that i t is an isotope of hydrogen.There thus exist two elements of mass 3, one an isotope of hydrogenand the other an isotope of helium.It is not possible yet to saydefinitely whether either alone or both together take part in atomicnuclear synthesis.I n this connexion, the writer would draw attention t o the veryremarkable permanent contraction suffered by hydrogen when it' G. L. Wendt and R. 8. Lmdauer, J . Amer. Chem. SOC., 1920, 42, 920;A., ii, 425INORGANIC! CHEMISTRY. 35has been activated and lost its activity. This point is detailed inthe section of this Report dealing with the first group of elements.Wendt and Landauer assume, of course, that H,, on keeping,regenerates ordinary hydrogen, but is i t absolutely certain that thisis th’e case? Collie’s results on the formation of helium in vacuumtubes containing hydrogen, his collateral results on the permanentdiminution in the volume of hydrogen in vacuum tubes, consideredin connexion with the theories of atomic nuclear structure, leadinevitably to the conclusion that H,, on keeping, gives little H,,but mainly helium.Although this suggestion may sound veryimprobable to many, it is, in reality, far more probable than anordinary chemical explanation, since it is scarcely possible to con-ceive that H, in the presence of nitrogen would not form ammonia,but prefer to react with the glass of the reaction vessel. Thissuggestion has been privately communicated to Dr. Wendt.Atomic W e i g h t s .The Report of the International Committee recommends onlyone change, namely, that the atomic weight of scandium should beraised from 44.1 to 45.1. The work of Honigschmid, on which thenew value is based, was referred to in last year’s Report.Three series of determinations have been made of the atomicweight of tin.Two of these involved the analysis of tin tetra-bromide by silver,BJg and the third depended on the direct electro-lytic estimation of tin in the tetrabromide.1° The values obtainedwere 118.700, 118.699, and 118.703, respectively, which agree verywell with the accepted1 value.The weight of a normal litre of methyl fluoride has been foundto be 1.54542 grams as the mean of twenty-three determinations.11From this, the atomic weight of fluorine is deduced as 18.996,which is very close to the accepted value of 19.Some determinations have been made of the atomic weight ofsamarium by the anhydrous chloride-silver ratio.12 As the meanof eighteen determinations, the value of 150.43 was obtained.In addition to the above, the following investigations may bereported.A determination has been made of the atomic weightof silicon by the analysis of silicon tetrachloride.13 The mean of* B. Brauner and H. Krepelka, J. Amer. Chem. SOC., 1920, 42, 917; A.,ii, 437. H. Krepelka, ibid., 925 ; A., ii, 437.lo G. P. Baxter and H. W. Starkweather, ibid., 905 ; A . , ii, 436.l1 E. Moles and T. Batuecas, J . Chim. phys., 191 9, 17, 537 ; A., i, 283.12 A. W. Owens, C. W. Balke, and H. C. Kremers, J. Amer. Chcm. Xoc.,1s G. P. Brtxter, P. F. Weatherill, and E. 0. Holmes, jun., ibid., 1194 ;1920, 42, 515 ; A., ii, 316.A., ii, 487.c 36 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.four experiments gave the value of 28.11, but as this is notaccepted as final by the authors, this value must await confirmation.By the analysis of bismuth chloride, a new value for the atomicweight of bismuth has been obtained.l4 The ratio BiC13:3AgClgave 209.024, and the ratio BiC13:Ag gave 209.027.The meanvalue 209.026 is one unit higher than the value a t presentaccepted.Colloids.A few papers have appeared on the preparation and propertiesof inorganic colloids, and mention may be made of the following.A convenient method for the preparation of metallic sols15 is tostrike an arc between poles of the metal under alcohol, usingcapacity in the circuit. With a current of 1.5 amperes and acapacity of 12-8 x MF, colloidal solutions have been obtainedof aluminium, antimony, bismuth, cadmium, copper, gold, lead,platinum, silver, and zinc. The colloidal metal is produced muchmore rapidly than by the earlier Sveciberg method.The stabilityof the sols is fairly great, and although a certain amount alwaysseparates, the bulk of the metal remains in solution. Gold andplatinum are exceptional, since their sols are very unstable.Colloidal rhodium16 has been prepared by the addition of aslightly alkaline solution of formaldehyde to a slightly alkalinesolution of the double chloride, Na,RhCl,, the reduction beingcarried out a t 40°. Under these conditions, a clear, colloidalsolution of rhodium is obtained. This solution absorbs hydrogento the extent of 2510-2960 times the volume of rhodium present.Similarly, the rhodium absorbs 346 times its volume of carbonmonoxide at 12-14O, and 1820 times its volume a t 60°.Thecolloidal solution, slightly alkaline, causes a very slight combinationof nitrogen and hydrogen to give ammonia, the reaction being coii-siderably enhanced if the solution is made just acid with verydilute tartaric acid in the presence of potassium tartrate.Mention may also be made of some work on the preparation andstability of mercury sols.l7 The most coiicentrated solution isobtained by passing a rapid stream of mercury vapour into coldwater, but in every case the sols are not very stable. Theirstability is materially increased by the use of gum arabic as aprotective colloid.l8l4 0.Honigschmid and L. Birckenbach, Zeitsch. Elektrochem., 1920,26, 403 ;A., ii, 549.l5 G. Borjeson and T. Svedberg, Kolloid Zeitsch., 1919, 25, 154 ; A., ii, 31.l6 C. Zenghelis and B. C. Papaconstantinou, Cornpt. rend., 1920,170, 1058 ;l8 A. Gutbier and G. L. Weise, ibid., 1919, 25, 97 ; A., ii, 36.A., ii, 380. l7 I. Nordlund, Kolloid Zeitsch., 1920, 26, 121 ; A., ii, 376INORGANIC CHEMISTRY. 37The Rare Gases.Mention must be made of McLennan’s work on the productionof helium on the large scale from natural gases.19 A large numberof gases from natural sources in various countries was investigated,and the Bow Island gas supplied to the town of Calgary, in Alberta,was selected. This gas consists of methane 91.6, ethane 1.9,nitrogen 6.14, and helium 0.36 per cent., together with traces ofcarbon dioxide and water vapour.It is not possible to givedetails of the experimental plant employed, which followed thelines of the Claude oxygen-producing column. By its means, intwo stages of working helinm, was obtained of 87-90 per cent.purity. By the use of a second plant, this was further purifiedt o 98-99 per cent. From the experience gained with these experi-mental plants, specifications have been drawn up for a commercialplant t o deal with the whole of the Bow Island supply of gas. Sixunits are proposed, each dealing with about 62,000 cubic feet perhour, the average daily supply of gas being 9,500,000 cubic feet.The yearly output of helium of 97 per cent. purity would be about10,500,000 cubic feet, and the working cost would be considerablyless than 219 per 1000 cubic feet.Group T.A most interesting paper has been published on the formation 20of triatomic hydrogen by various means from ordinary hydrogen.Hydrogen a t atmospheric pressure, when submitted to the actionof a-rays from radium emanation or passed through a silent dis-charge tube, is converted into an active form, and a similar resultis obtained when the electric discharge from a large induction coilor transformer is passed through a vacuum tube, through whichhydrogen is passed under a pressure of 2-8 cm.I n each case, asmall amount of an active form of hydrogen is produced, which isat once condensed on passing the hydrogen through a spiral tubecooled in liquid air.This active modification reacts with sulphur,arsenic, phosphorus, mercury, and nitrogen, and also reduces acidand neutral solutions of pota.ssium permanganate. The amountof hydrogen that is converted into the active form in the experi-ments described has not exceeded 0.02 per cent.Very careful experiments have proved that the enhancedreactivity is not due t o the presence of ions, and also the substance2o G. L. Wenat an.d It. S. Landauer, J. Amer. Chem. SOC., 1920, 42, 930;J. C. McLennan, T., 1920,117, 027.L4., ii, 42538 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.differs in its chemical and physical properties from atomic hydrogenprepared by Langmuir, which was referred to in the Reports for1912 and 1915. The polyatomic nature of the substance isindicated by the contraction in volume of the hydrogen when i t isformed, and, moreover, positive ray analysis has shown the un-doubted existence of H, molecules when the electric discharge ispassed through hydrogen a t low pressures.There is little doubtbhat the substance is indeed H,.It is very unstable, and disappears very rapidly after it hasbeen formed. This was shown by the increased reactivity that isobserved when the flow of hydrogen through the silent dischargetube is increased. A t atmospheric pressures it is found that thereactivity disappears within about one minute.Perhaps the most interesting phenomenon in these experimentsis the permanent contraction that takes place in the hydrogen.This was noticed by Usher,21 who carried out experiments on thesynthesis of ammonia by exposing a mixture of hydrogen andnitrogen to the action of a-rays from niton mixed with the gases.I n one case, a contraction of 0*24 C.C.was observed, but only0.006 C.C. of ammcnia had been formed. Collie and’ Patterson22observed a similar disappearance of 3.6 C.C. out of 4.6 C.C. ofhydrogen when the gas was sparked under reduced pressure withcopper or aluminium electrodes. A possible explanation of thisphenomenon is put forward in an earlier section of this Report,and it would, indeed, seem that this may prove to be even moreinteresting than the preparation of H,, great as is the importanceof this advance.Investigation has shown that lithium behaves similarly t o sodiumand potassium iii forming soluble silicates containing a large excessof the acid over the base.23 Lithium metasilicate, Li2Si0,, hasbeen prepared in an insoluble and a soluble modification, theformer having the formula Li2Si0,,R20 .Brief reference may be made to some experiments on the actionof alcohol on the sulphates of sodium.” Dry alcohol acts on drysodium hydrogen sulphate to give the intermediate sulphate,Na,SO,,NaIISO,, and free sulphuric acid, which dissolves in thealcohol. No action takes place with potassium hydrogen sulphate.I n the presence of moisture, sodium hydrogen sulphate is firstconverted into the intermediate sulphate, and then, finally, into21 F.L. Usher, T., 1910, 97, 389.22 J. N. Collie and €1. S. Patterson, P., 1913, 29, 22, 217.28 I<.A. Vesterberg, Medd. K . Vctenakapsakad. Nobel-Inst., 1919, 5, No. 30 :24 G. S. Butler and H. B. Dunnicliff, T., 1920, 117, 649.A., ii, 112INORGANIC CHEMISTRY. 39ordinary sodium sulphate. When an alcoholic solution of sulphuricacid (20 per cent. or less) acts on sodium sulphate in the cold, theintermediate sulphate is formed. Nitre cake consists ofNa2SO,,NaHSO4 alone or mixed with either NaHSO, or Na,SO,,according as the acidity is equal to, greater than, or less than,18 per cent. H,SO,.A process has been patented for the preparation of metallicpotassium by heating potassium hydroxide and sodium in exactlyequivalent proportions a t 670° in the absence of air.25 Hydrogenis produced and the potassium is volatilised and may be condensed.Some further and, it may be said, conclusive work has beencarried out on the possible existence of an alkali metal of higheratomic weight than cmium.2G The alkalis were separated from3500 grams of pollucite, which contains more than 30 per cent.ofmsium oxide, and the mixture was carefully tested for the presenceof the next higher homologue to czesium. There is no need todescribe the experimental details, but no indication whateverwas found of the presence of a new element.Group 11.A study has been made of the equilibrium conditions whichobtain between arsenic oxide, calcium oxide, and water a t 3 5 O forthose mixtures in which the arsenic oxide is in excess.27 Evidencewas found of the existence of two orthoarsenates of calcium, namely,dicalcium orthoarsenate monohydrate, CaHAs0,,H20, and mono-calcium orthoarsenate, CaE,(AsO,), .The former is identical withthe mineral haidiiigerite, and is stable in contact with a solutioncontaining more than 27.5 per cent. of arsenic oxide, whils't thelatter is stable with a lower percentage of arsenic oxide in thesolution.Mention may be made of the fact that chlorine has no action oncalcium carbide, whilst liquid bromine slowly reacts to give hexa-bromoethane and calcium bromide.28 The reaction is very slow,and 4.5 grams of the finely-powdered carbide treated with 45 gramsof dry bromine for five weeks gave 22 grams of hexabromoethane,5.8 grams of calcium bromide, and 0.2 gram of unchanged carbide.Reference was made in the Report for last year to the fact thatthe decomposition of bariuw? peroxide takes place a t a much lowertemperature in the presence of silica, a certain amount of bariumF.C. Wickel and TV. Loebel, D.R.-P. 307175 ; A . , ii, 32.26 L. M. Dennis and R. W. G. Wyekoff, J. Amer. Ckem. SOC., 1920,42,98528 E. Barnes, Cham. NEWS, 1919, 119, 260 ; A . , ii, 33.A., ii, 431. 27 C. N. Smith, ibid., 259 ; A . , ii, 37540 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.silicate being formed. The influence of a great number of otheroxides has now been studied by examining the heating curves ofthe mixtures in molecular proportions.29 Cuprous oxide reactsviolently with barium peroxide t o give cupric oxide, which decom-poses barium peroxide catalytically, the optimum t'emperature beingabout 660O.Biagnesiuin 2nd calcium oxides start the decompositiono i the peroxide at 250" and 310° respectively, whilst zinc oxidecauses slow deconiposition between 200° and 370° and forms bariumzincate. Zirconium oxide, stannous oxide, and stannic oxide haveno action, but the oxides of cadmium, lanthanum, and cerium actas pure catalysts. Aluminium oxide forms barium aluminate, andtitanium oxide in niolecular proportions gives a titanate, probablyBaTi03. With twice the molecular proportions of barium peroxidea basic titanate is produced. Litharge and barium peroxidebetween 300° and 400° evolve no oxygen, but form a brown sub-stance of unknown composition. Above 500° much oxygen isevolved, with the probable formation of Ba,PbO,.Vanadium pent-oxide reacts vigorously with barium peroxide. When eqnimolecularproportions are used, the reaction begins a t 2 1 5 O and is endeda t 530*, Ba(V03), being formed. With 2Ba0, the metavansdate isfirs6 formed, but ab 3 7 5 O a second, very vigorous, reaction startsand the colour changes from brown to white, the product apparentlybeing Ba,V,07. Tantalum pentoxide also reacts vigorously t o givea tantalate. With arsenious oxide and three moleciiles of bariumperoxide, arsenic oxide is first formed a t 310° to 410°, and above4 6 5 O oxygen is evolved and barium arsenate is formed. Withantimony oxide at ZOOo oxygen is evolved with almost explosiveviolence. Bismuth oxide stsrts a gradual evolution of oxygen a tabout 250°, and higher bismuth oxides, or compounds of these withbarium peroxide, appear t o be formed.With chromium oxide nooxygen is evolved, and barium chromate is produced. The oxidesof molybde~ium, tungsten, and uranium all cause evolution ofoxygen and form molybdates, tungstates, and uranates respectively.The kwer oxides OF manganese are all oxidised and give bariummanganate. Ferric oxide acts catalytically, and gives bariumferrate, whilst nickel and cobalt also act catalytically and arechanged into higher oxides, which do not agree in their propertieswith the known peroxides o l these metals.It has besn found that strontium sulphide is readily hydrolysedby water t o give equiinolecular proportions of the h ydrosulphideand the hydroxide.30 These two compounds do not form a niixed2 0 J.A. Hedvall and N. von Zweigbergk, Zeksch. anorg. Chem., 1919, 108,119 ; A., ii, 35.3 0 M. Bruckner, Zeitsch. Elelctrochem., 1920, 26, 25 ; A., ii, 251JNORGANlC CHEMISTRY. 41compound, and the hydroxide may be separated by crystallisation.When strontium sulphide is extracted with hot water and the clearfiltrate cooled, pure strontium hydroxide, Sr(OH),, crystallises.The case is different with barium sulphide, as the hydroxide andhydrosulphide . form an additive compound, OH*Ba*SH,5H20.31Under no conditions can pure barium hydroxide be crystallisedfrom the solution obtained by the action of water on bariumsulphid e.From a study of the equilibrium between zinc oxide, phosphoricoxide, and water a t 2 5 O and 37O, the following phosphates of zinchave been found to exist : Zn3(P0,),,4H,0, ZnHP0,,3H20,Zn(HZP0,),,2H,O, whilst a t 3 7 O an additional salt, ZnHPO,,H,O,is obtained.32 Similar investigations with sodium hydroxide solu-ticns and zinc oxide have established the existence ofNa20,Zn0,4H,0 8s a stable compound.33Group 111.An investigation has been made of the equilibrium conditionsbetween alinminium nitrate, nitric acid, and water a t 25O, and itwas found that three hydrates of the salt have a stable existence.34The first, Al(N0&,l8H2O7 is most stable in contact with the solu-tion containing 73 per cent.or less acid, the second,Al(N03)3,16H,0, is stable with 73-81 per cent. acid, whilst thethird, A1(NO,),,12H2O, is stable in the presence of more than 81 percent.of nitric acid.Some phjsical measurements have been made of the solutionsobtained by dissolving aluminiwm in aqueous solutions of sodiumhydroxide and of ainmonium hydroxide.35 Whilst the physicalaspect of this work does not fall within the purview of this Report,the results have some value for inorganic chemists. It is shownthat aluniiniuni hydroxide aeulralises the alkalis as a monobasicacid, and that the aluminates are salts of the acid HAl(OH),, thatis, Al(OII),,H,O. Ammonium aluniinate, NH,Al(OH),, is quitestable in solution.Some further work may be reported on scandium fluoride andthe scandifluorides.36 The best method for the preparation of the31 K. Bruckner, Zeitsch.Elektrocherrz., 1920, 26, 1 ; A., ii, 252.32 N. E. Eberly, C. V. Gross, and W. S . Crowell, J . Arne).. Chein Xoc., 1920,33 F. Goudriaan, PTOC. K. Akad. Wetensch. A4msterdam, 1919, 22, 17934 K. Inamura, J. Tokyo Ckern. SOC., 1920, 41, 1 ; A., ii, 625.35 J. Heyrovskf, T., 1920,117, 1013.36 J. Stgrba-Bob, Bull. SOC. chim., 1920, [iv], 27, 185 ; A., ii, 315.42, 1433 ; A., ii, 545.A., ii, 113.C42 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.pure fluoride is by the action of hydrofluoric acid on scandiumoxide, the product being finally heated a t 150-180O in order t oremove the excess of hydrofluoric acid. Whilst the free scandi-fluoric acid does not exist, two new ammonium scandifluorides havebeen prepared. The salt, (NH4),ScF,,37 is hydrolysed in thepresence of ammonium fluoride to give in quadratic crystals thesalt, (NH&3cF5.I n warm or cold water alone the salt,(NH4)ScF,, is always obtained as a microcrystalline powder. Bythe dissolution of scandium fluoride in a concent,rated solution ofsilver fluoride a scandifluoride of silver is formed.It is known that lead nitrate and nit,rite interact in solution togive well-defined compounds containing both salts. Similarly, thal-lium nitrite and lead nitrate react to give basic compounds of bothsalts.,' In the case of potassium nitrite and thallium nitrate nosuch double salts are obtained, but thallium nitrate-nitrites areformed which are stable and can be crystallised unchanged. Withtwo molecules of potassium nitrite and one molecule of thalliumnitrate the compound formed has the formula Tl,N,O,.With dif-ferent proportions other salts are obtained, in which the ratiobetween nitrate and nitrite is not a simple one.Since last year's Report was written Sir Charles Parsons haspublished a complete account of his experiments on the artificialproduction of diamond.39 It is shown beyond any doubt that highpressure alone is not sufficient to cause the conversion of graphiteinto diamond, and it is also shown that iron must be present. Ex-periments in which a mixture of acetylene and oxygen is highlycompressed and a temperature produced in excess of that requiredto vaporise carbon, accompanied by a momentary pressure of 15,000atmospheres, prove that the failure to produce diamond is not dueto lack of temperature.Many of the experiments, in which it hasbeen claimed that diamond is produced, have been repeated, andnegative results were obtained unless iron played a part. Experi-ments under vacua from 75 mm. up to X-ray vacua have showngenerally that as the pressure is reduced the yield of diamond isdiminished. On the other hand, when alloys, previously boiledunder atmospheric pressure, are quickly heated in a high vacuum,violent ebullition takes place, due t o the large volume of gasesliberated, and some of the contents of the crucible are ejected before37 R. J. Meyer, Zeitsch. anorg. Chem., 1914, 86, 257 ; A., 1914, ii, 369.38 L. Rollo and G. Belladen, Gazzetta, 1919, 49, ii, 217 ; A . , ii, 34.lS (Sir) C.A. Parsons, Phil. Trans., 1919, [A], 220, 6 7 ; A., ii, 110INORGANIC CHEMISTRY. 43they have time to part with their occluded gas, and diamond occursin the spherules so ejected. There is no doubt that these gases,possibly containing a ferro-silicon carbonyl, are necessary for theproduction of diamond. It seems alniost certain that the chieffunction of quick cooling in the production of diamond in an ingotor spherule is to bottle up and coiicentrate into local spots the gasesoccluded in the metal which, under slow cooling, would partlyescape, whilst the remainder would become evenly distributedthrough the mass. The necessity of subjecting the iron to a tem-peratnre above 2000° before cooling would imply the necessity ofcarbides of silicon, magnesium, etc., being present t o ensure thenecessary chemical reactions with the gases a t high pressure withinthe ingot.The greatest percentage of diamond was obtained whenthe atmosphere round the crucible consisted of 95 per cent. ofcarbon monoxide, 1 per cent. of hydrogen, 2 per cent. of hydro-carbons, and 2 per cent. of nitrogen. The weight of diamond wasabout 1/20,000 that of the iron. It seenis probable that the rateof cooling might be so prolonged as to obtain much larger crystalsand a larger total yield.The presence of crystals of SiO,, Al,O,, and MgO, the spinels, andpyrope, associated with diamond in rapidly cooled iron alloys,appears to have a bearing on the presence of similar crystalsfmnd in association with diamond, and to be compatible withBonney’s view that eclogite is the parent rock of the diamond inSouth Africa.It seems probable that both the eclogite and thediamond may have been siinultaneously crystallised from an ironalloy. Since the average weight of diamond in the blue groundof South Africa is 1 in 5,400,000, there has been produced in coolediron more than 270 times this amount.Investigations were made during the war of the absorptive poweror” various vegetable charcoals and the improvement that is causedby heat treatment. These have now been published in part, andin the first paper the effect of heat treatment on the absorptivepower of sugar charcoal for sulphur dioxide is de~cribed.~O Afterheating the charcoal for forty-five hours the amount of sulphurdioxide absorbed per gram was increased from 97 C.C.to 288 C.C. I na second paper exactly analogous results were obtained, and apossible explanation is suggested.41 The main experiments werecarried out with birch-charcoal, but other wood charcoals were used.The absorptive powers were measured with sulphur dioxide, carbondioxide, and also aqueous soluticns of methylene blue. It wasfound that the absorptive power is very materially increased by40 R. M. Winter and H. B. Baker, T., 1920,117, 319.41 J. C. Philip, S. Dunnill, and (Miss) 0. Workman, ibid., 362.a* 44 ANNUAL REPORTS ON TRE PROGRESS OF CHEMISTRY.the heat treatment, with the result that the activity of animalcharcoal can be paralleled and even surpassed by wood charcoal.It was noticed that the heat-treatment is not the only factor inenhancing the activity, and the clue was found in the decrease inthe bulk density of the charcoal during the heating process.Ifthe heating is carried out in the absence of oxygen little or no im-provement jn the activity is produced, and oxygen must be presentfor the activation to take place. The explanation is probably thatin the case of a freshly prepared sample the capillaries through thematerial are exceedingly small, so that they are soon blocked whenabsorption takes place. When the charcoal is heated in the presenceof oxygen some oxidation takes place, and the capillaries becomewider, so that the effective surface is enormously increased.A convenient method has been described for the removal of carbonmonoxide from its mixtures with other gases for analytical andhygienic purposes.42 The carbon monoxide is very rapidly oxidisedby chromic acid solution to which some mercuric oxide has beenadded.Some further work on the derivatives of the silicon hydricles maybe reported .43 It was previously shown that dibromomonosilanereacts with water to form polymerides of protosiloxane, O-SiH,.The unimcllecular form has now been obtained as a gas by the actionof the required amount of water-vapour on dichloromonosilane in avery large flask under greatly reduced pressure.The compound hasan extraordinary tendency t o polymerise, in consequence of whichthe flask nimi be perfectly clean and smooth. Liquid and solidpolymerides are formed immediately on condensation.The liquidones are like benzene, and c m be conveniently obtained as a solu-tion by shaking a benzene solution of dichloromonosilane withwater. They correspond approximately with the formula (SiH,O),.The solid polymerides are insoluble. All the polymerides react withsodium hydroxide in accordance with the equation SiH,O +2NaOlI = Na,Si03 + ZH,.The behaviour of disilane, Si,H,, towards halogen acids has beeninvestigated, and is found closely t o resemble that of monosilane.Disilane does not appear to react yith hydrogen chloride a t theordinary temperature or a t 120°, bnt in the presence of a little sub-limed aluminium chloride a reaction occurs more or less readilyaccordng to the general scheme :Si,K, + zHCl= Si,H,-, Clx + zH.A mixture of chlorides is invariably produced, the equilibrium lying42 K.Hofmann, ~9.22.-P. 307614 ; A , , ii, 309.43 A. Stock and K. Somieski, Ber., 1919,52, [B], 1851 ; 1920,53, [B], 759 ;A., ii, 31, 429INORGANiC CHEMISTRY. 45in favour of the intermediate members of the series. Thus withhydrogen chloride (1 val.) and disilane (less than 1 vol.) the mainproduct is dichlorodisilane, very little inonochlorodisilane beingobtained. With the gases in the volume ratio 2 : 1 much trichloro-disilane, in addition t o dichlorodisilane, is obtained. Completechlorination is not effected by a large excess of hydrogen chloride.It was not found possible t o isolate monochlorodisilane in a purestate, and also the final purification of dichlorodisilane could not beeffected, since it forms a mixture of constant boiling point withtrichlorodisilane.There is no doubt that as in the case of thecarbon compounds mixtures of isonierides are formed in the halogen-ation of disilan e.The brominatisii of disilane has been carried out in a preciselyanalogous manlier, and monobromodisilane, m. p. - looo to - lolo,has been isolated in a state of purity.The hydrolysis of the halogenated disilanes corresponds exactlywith that of the similar monosilanes. Thus monobromodisilanereacts with water t o yield the substance (Si,H,),O, a colourlessliquid which can be volatilised without decomposition, and, whendissolved in benzene, instantaneously reduces cold silver nitrate,but not copper sulphate, solution.It reacts slowly, but quantita-tively, with sodium hydroxide solution in accordance with theequation (Si,H,),O + 8NaOH + 3H,O = 4Na,SiO, + 12H,. The solidproducts obtained by the hydrolysis of dibromodisilane and themore highly halogenated derivatives closely resemble silico-oxalicacid, (HOzSi*SiO,H)x. They are only slowly hydrolysed furtherby water, can be dried in a desiccator without marked decomposi-tion, evolve hydrogen when treated with alkali hydroxide, andfinally yield a residue of silicate. Evidently the Si-Si linkingremains intact in them, and appears to be more stable towardsalkali than was a t first thought.Amorphous zirconiuni may be obtained from potassium zirconiumfluoride by means of sodium or aluminium, and the coherent forincan be prepared from the sane salt by aluminothermic reduction.44The coherent metal is much less chemically active than theamorphous variety, and, unlike the latter, is insoluble in all acidsexcept hydrofluoric acid and aqua regia.It has been shown that zirconium monoxide does not exist, theblack powders obtained by the reduction of the dioxide bymagnesium being mixtures of the metal and the dioxide.45The iodates, perchlorates, and a chlorate have been prepared of44 J.W. Marden and M. N. Rich, J . Ihd. Zng. C'hern., 1920, 12, 661 A.,4s R. Schwarz and H. Deisler, Ber., 1919, 52, [B], 1896 ; A., ii, 42.ii, 54746 ANNUAL REPORTS ON THE PROGRESS 03' CHEMISTRY.zirconium.465Zr0(OB)2,8ZrO(103)2, 3Zr0(OH)2,4Zr0(I03)z,2ZrO(OH),,ZrO(I03),,3ZrO(OH),,ZrO(103)2, ZrO(C10,),,HC10,, Zr0(OH)2,9ZrO(C104),,and Zr0(OH)2,3ZrO(C103),.Following the method described in last year's Report for thepreparation of bismuth hydride, tin hydride has also been pre-pared.47 It is a gas that can be condensed by liquid air andvolatilised without decomposition.Some preliminary experimentsseem to show that lead hydride also can exist in the gaseous state.The following are described : ZrO(OH)2,2ZrO(IOa)2,Group V .Investigations have been made of the electrolysis of a solutionof ammonium azide in liqui'd ammonia a t - 6 7 O with anodes ofvarious metals.48 The evolved gases were measured, and the lossof weight of the anode determined. Proof was obtained of theformation of the following azides: CuN,, CuN,, AgN,, CdN,,PbN,, and SbN,.A deep red solution of ferric azide, FeN,, wasobtainea, but the compound was ammonolysed, and yielded anammono-basic ferric azide.The equilibrium between nitric oxide and bromine and theirreaction products has been studied between - 1 5 O and 330O. Withbromine a t pressures below 50 mm. and a t temperatures above140°, nitrosyl bromide is formed, the amount of the tribromidepresent being negligible.,, Independent evidence of the existenceof nitrosyl bromide and nitrosyl tribromide was obtained from thefusion-point diagram. The tribromide, NOBr,, is a brownish-black, almost opaque, liquid, which boils with partial decompositionat 32O.It has been found that red phosphorus acts as a reducing agenttowards many metallic salts in aqueous solution, and! very possiblythe method may prove of use in qualitative analysis.60 The solu-tion of the salt is boiled with 0.2 gram of red phosphorus for afew minutes.Mercuric and mercurous salts are reduced to themetal, gold and silver salts give insoluble phosphides, whilstpalladium and osmium salts yield either the metal or a phosphide.Stannic salts are partly reduced to stannous salts, ferric salts are46 F. P. Venable and I. W. Smithey, J . Arner. Chern. SOC., 1919, 41, 1722 ;4 7 F. Paneth and K. Fiirth, Ber., 1919, 52, [B], 2020 ; A., ii, 41.48 A. W. Browne, M. E. Holmes, and J. S. King, J . Amer. Chem. SOC.,4 s M. Trrcutz and V. P . Dalal, Zeitsck. anorg. Chem., 1920,110, 1 ; A., ii, 308.60 L.Rosenstein, J . Amer. Chem. SOC., 1920, 42, 883 ; A., ii, 428.A., ii, 43.1919,41, 1769 ; A., ii, 31IXORGANIC CHEMISTRY. 47reduced t o ferrous, iridic salts to iridous, selenates to the elementor a phosphide, molybdates to quadrivalent molybdenum salts,vanadates to tervalent vanadium salts, dichromates to chromicsalts, and permanganates to nianganous salts. Bismuth, lead,cadmium, antimony, and arsenic salts, arsenates, and stannous saltsare not reduced, whilst telluretes and platinichlorides are veryslowly reduced.When a few drops of phosphorus trichloride are added to anaqueous solution of alsenious oxide, the solution turns yellow, the11opaque-brown, and finally a copious precipitate of arsenic is throwndown.51 The reaction probably takes place in accordance withthe equation As,O, + 3PC1, + 9H20 = 2As + 3H,P04 + 9HC1.Thearsenic is amorphous, insoluble in carbon disulphide, and is appar-ently a new allotropic modification. The reaction takes place witharsenates and arsenites, and is very delicate, since the presence of0.000075 gram of arsenic per C.C. can be detected.Arsenic trichloride can very conveniently be prepared by passingcarbonyl chloride over a mixture of arsenious oxide (80 per cent.)and carbon (20 per cent.) heated a t 200° to 260'. The yield isalmost quantitative.52Golden antimony sulphide is usually supposed to be a mixture ofSb2S5, Sb,S,, and some free sulphur. The compound, Sb2S5, how-ever, is now shown not to exist, and the golden sulphide, afterextraction of the free sulphur, has the formula Sb2S4.Thissulphide can also be prepared in the following way.53 By theinteraction of Schlippe's salt and zinc chloride, zinc thioantimonateis precipitated. The crude salt contains free sulphur, and, afterremoval of this, the product has the formula Zn,Sb,S,. On treat-ment with dilute acid, an orange-red residue is obtained, which hasthe composition Sb,S,.By the oxidation of bismuth oxide or hydroxide in the presenceof alkali by chlorine, ammonium persulphate, or potassium ferriccyanide, the higher oxides of bismuth have been prepared.54 Thetetroxide was obtained as Bi,O, and Ei,O,,H,O, and of each ofthese there are two modifications, which are brown and purplish-black respectively.A third variety, Bi204,2H,0, which is yellow,has also been prepared. Bismuth pentoxide monohydrate,Bi,0,,H20, is obtained by the oxidation process, but is mixed withthe tetroxide. It can be prepared from sodium bismuthate by5L N. N. Sen, J . PTOC. Asiutic SOC. Bcngal, 1919, 15, 263 ; A . , ii, 308.52 L. H. Milligan, W. A. Baude, and H. G. Boyd, J . Ind. Eng. Chem.,53 F. Kirchhof, Zeitsch. anorg. Chem., 1920, 112, 67 ; A., ii, 693.54 R. R. Le G. Worsley and P. W. Robertson, Z'., 1920,117, 63.1920, 12, 221 ; A., ii, 37248 ANNUAL REPORTS ON THE PROGRES3 OF CHEMISTRY.repeated grinding with glacial acetic acid. The anhydrous oxidedoes not seem t o be capable of existence, as the monohydrate losesboth water and oxygen in a vacuum over phosphoric oxide.Bismuth hexoxide has also been prepared by the oxidation process,and is anhydrous.Group Vil.The solubility has been determined of sulphur dioxide insulphuric acid of various ~oncentrations.~5 Th6 measurements werecarried out a t 20°, and the acid concentration was varied from 55to 100 per cent.It was found that a sharp minimum solubilityoccurs with an acid containing 86 per cent. of H,SO,, and i t issignificant that the monohydrate, M,SO,,H,O, contains 84.5 perThe oxidation of ferrous chloride in presence of hydrochloricacid, and of ferrous phosphate in the presence of phosphoric acid,by sulphur dioxide has been studied.56 I n the first case, the reac-tion takes place in accordance with the equation4FeC1, + SO, + 4HC1= 4FeC1, + 214,O + S.The maximum amount of ferric iron produced was about 9 percent., and there seems little doubt that the reaction is reversible.I n the second case, more ferrous salt is oxidised, and the view isexpressed that the reaction4Fe(H2P8,), + 4€I,PO, + SO,= 4Fe(H,P0,)3 + 2H,O + Sis also reversible, but that it is modified by the formation of thestable complex formed by ferric phosphate and phosphoric acid.cent.of H,SO,.G’roup ??ill.A simple and rapid method has been described for the prepar-ation of iodine pentoxide, which depends on the oxidation of iodineto iodic acid by means of 24-26 per cent. chloric acid solution,the evaporation of the solution, and the dehydration of the iodicacid.57 The solution of chloric acid is prepared as iollows:625 grams of barium chlorate [90 per cent.Ba(C10,),] are dissolvedin 1 litre of nearly boiling water, and the solution is poured intoan earthenware crock. The required amount of hot sulphuric acid(obtained by mixing equal volumes of concentrated sulphuric acidand water) was slowly added. It is very necessary to have a slight55 F. D. Miles and J. Fenton, T., 1920, 117, 59.5 6 W. Wardlaw and I?. €3. Clews, ibid., 1093 ; W. Wardlaw, S. R. Carter,6 7 A. B. Lamb, W. C. Bray, and W. J. Geldard, J . Arner. Chem. SOC.,and F. €1. Clews, ibid., 1241.1920, 42, 1636 ; A., ii, 615INORGANIC CHEMISTRY. 49excess of barium chlorate rather than sulphuric acid, as the latterrenders the iodine pentoxide less stable. The solution of chloricacid may be kept unchanged in glass bottles for several weeks.Itis found that ‘in the presence of 3 per cent. excess of chloric acidthe net reaction with iodine is expressed by the equationI, + 2HC10, = 2EII0, -F C1,.The mechanism of the reaction, however, does not consist of thedirect replacement of chlorine by iodine. A considerable quantityof chloric acid is reduced to hydrochloric acid in accordance withthe eqmtion 312 -1- 5EIC10, + 3H20 = 6HIQ, 4- 5WC1. A solutioncon tainiiig hydrochlcric and iodic acids loses iodine on evaporationaccording t o the equation 21310, + lOHCl= I, + 5C12+ 6H,O. Thisis preveiiteci by an excess of chloric acid, which reacts with thehydrochloric acid, and i t was found that an excess of 3 per cent.is sufficient.The iodine is cxidised in quantities of 500 grams, the reactionbeing finished in about twenty miizutes.The iodic acid obtained011 evapcratioii is heated a t 150--lGO0 for three hours. The finddehydration is carried out a t 235-240° in a slow current of dryair. The iodine pentoxide is pure white, and has practically thetheoretical oxidising value, and the yield is almost quantitative.The process has many advantages over the nitric acid method.With reference to this preparation of iodic acid, it is interestingt o note that iodine replaces bromine when the former acts on anaqueous solution of potassium bromate, and that a similar reactiondoes not occur with bromine and potassium chlorate, whilst thereaction between iodine and potassium chlorate is more coniplex.s*The following changes have been shown t o occur:and KHI,O, + KC1 + HC10 = 2KI0, + H,O + Cl,.Potassium maiiganifluoride, Ii2MnF,,LI,0, has been prepared bythe action of nitrous acid 011 potassium permanganate in thepresence of hydrofluoric acid .59 The permanganate is reduced bythe nitrous acid.A manganous salt may also be used, in whichcase the nitrous acid acts as an oxidising agent.ZMClO,-+ 21 + MZO = KHI20, + KC1 + HC10Group V I I IIt has been found that the yield of sodium ferrate obtained bythe electrolysis of sodium hydroxide solution with iron anodes isvery materially increased by superposing an alternating current5 8 G. Gruber, Zeitsch. physikal. Chern. Uszterr., 1920, 33, 107 ; A., ii, 684.69 I.Bellucci, Gaxzetta, 1919, 49, ii, 180 ; A., ii, 4050 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.on the direct current.g0 I n one case the increase of yield was 160per cent. I f the anode and cathode are separated, and thetemperature of the electrolyte is not allowed to exceed 50°, a twhich the ferrates decompose, and an alternating current is super-imposed on the direct current, saturated solutions of sodium ferrateand the crystalline salt can be obtained.Some further work has been carried out on the influence ofhydrogen sulphide on the occlusion of hydrogen by palladium .61The earlier experiments were discussed a t some length in theReport for last year. I n the earlier paper it was shown that whenpalladium is poisoned by hydrogen sulphide, and then heated a tlooo in a vacuum, an amount of hydrogen is evolved equal involume to that of the hydrogen sulphide previously absorbed inthe poisoning.The sulphur is retained by the palladium, a com-plex of the formula Pd,& being formed. Dr. Maxted believes thatpalladium can dissociate hydrogen sulphide to form this complexand free hydrogen slowly a t ordinary temperatures. When thistakes place, more hydrogen is slowly occluded, and the total volumeso occluded added t o the volume derived from the hydrogensulphide is equal to the true occlusive power of palladium forhydrogen, allowing for the palladium which has formed the Pd,Scomplex. This explanation is based on the observation that asample of palladium which has been completely poisoned byhydrogen sulphide slowly gains a power of absorbing hydrogen upto a fixed amount, and that the rate of absorption is faster thelonger the poisoned palladium is kept before the hydrogen isadmitted.This interpretation may be criticised from two points of view.I n the first place, since the palladium dissociates hydrogen sulphide,i t is probable that this dissociation occurs a t the time of occlusion,and that it is, indeed, the basis of the occlusion.I n the secondplace, if palladium is absolutely completely poisoned by hydrogensulphide, it should not gain, on keeping for an unlimited time, anypower of occluding hydrogen. Dr. Maxted offers no explanationof his view that the occlusive power for hydrogen should beincreased when the hydrogen sulphide is dissociated. True poison-ing must mean the absorption of hydrogen sulphide up t o thepoint when a portion of the palladium is converted into the com-plex Pd,S, and the remainder is saturated with the hydrogenobtained by the dissociation of the hydrogen sulphide. Obviously,when this has been secured, no further hydrogen can be occluded.It would seem far more probable that the poisoning obtained with6o a. Grube and H. Gmelin, Zeitsch. Elektrochem., 1920, 26, 153; A .ii, 377. 61 E. B, Maxted, T., 1920, 117, 1280INORGANIC CHEMISTRY. 51hydrogen aulphide is not complete in the strict sense, but that thepoisoning is concentrated on the surface. On allowing the partlypoisoned palladium to remain, a more equal distribution of thehydrogen takes place, with the result that more hydrogen can beoccluded. This is shown by the fact that, even after the palladiumhas been (‘ completely ” poisoned by hydrogen sulphide, it stillpossesses the power of slowly absorbing more hydrogen sulphide.The data are still too incomplete for accurate calculations of thetrue equilibrium conditions. It appears that 1 gram of palladiumhas the definite power of absorbing 69 C.C. of hydrogen. I s thewhole of thii hydrogen dissociated into atoms, or are there twoprocesses, first the occlusion of hydrogen as atoms, followed by asecondary effect of condensation as hydrogen molecules ? Thesecond alternative seems the more probable, but the question canonly be decided by accurate measurements of the dissociationpressures of hydrogenised palladium.An investigation has been made of the hydrolysis of aqueoussolutions of potassium platinichloride.62 It is shown that N / 50and more concentrated solutions are slowly and completely hydro-lysed in the dark, whilst #/lo0 and more dilute solutions undergohydrolysis only when exposed to light. It is found that the hydro-lysis takes place a t first very slowly, but after a time the rateincreases, and this is attributed to the formation of some substancewhich acts as a catalyst. This view was supported by the factthat the addition of a portion of a photochemically hydrolysedsolution to a fresh, N / 100-solution of platinichloride causes thelatter t o undergo hydrolysis in the dark.The addition of a soluble chloride to the hydrolysed solutioncauses a complete reversal of the reaction, and this reverse reactionis influenced by light in much the same way as is the direct reac-tion. The influence of platinum-black in accelerating both thedirect and reverse reactions in the dark is quite noticeable, but isnot measurable when light is acting on the solutions.E. C. C.‘BALY.62 E. H. Archibald, T., 1920, 117, 1104