@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Chemistry, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Morell, Arthur Ernest"@en ; dcterms:issued "2010-03-19T21:35:57Z"@en, "1929"@en ; vivo:relatedDegree "Master of Arts - MA"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "[No abstract available]"@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/22179?expand=metadata"@en ; skos:note "U . B . C , L J 8 R A R Y CAT ^ AN ACCURATE! REDETERMINATION OF THE ATOMIC WEIGHT OF CAESIUM by Arthur Ernest Morell A Thesis submitted for the Degree of MASTER OF ARTS in the Department of CHEMISTRY THE UNIVERSITY OF BRITISH COLUMBIAApril, 1927. - 1 -CONTENTS. I. Introduction. page 2 II. Synopsis of Earlier Determinations. \" 4 III. Chemical Properties of Caesium Salt. \" 12 IV. Physical Constants used as Data. \" 14 V. Purification of Chemicals. \" 14 VI. Description of Apparatus. \" 44 VII. Analysis of the Chloride. \" 43 Analysis No. \" VIII. Analysis of the Bromide. \" 34 IX. Summary. \" 60 X. Bibliography. \" 62 - 2 -AN ACCURATE REDETERMINATION OF THE ATOMIC WEIGHT OF CAESIUM I. INTRODUCTION. A redetermination of the atomic weight of caesium was undertaken on account of the possibility of finding an isotope. As can be readily seen the present accepted value, I 3 2 . 8 I , differs considerably from an integer and yet no isotopes of this metal are known. Other elements, such as H ) chlorine, strontium, mercury and gold, possessing fractional atomic weights, have at least one isotope and yet caesium, also having a gractional atomic weight, possesses none. This may mean one of two things, perhaps both; that either the accepted value of the atomic weight is at fault, or that an isotope, as yet unknown, exists. In addition to this, the value for the atomic weight of caesium as calculated by chemical and physical methods is not (2) the same. Aston in 1921 by making use of the mass spectro-graph found caesium to be a simple element, having a mass of 133 1 .2. However, the possible error in his early work happens to be as large as the difference between the accepted atomic weight and the mass of the atom, .2 unit; yet there are very few simple elements showing such a great difference between the mass number and the atomic weight. The element (1) Aston, Phil. Mag., (6) 49, 11<71 (I923). (2) Aston, Phil. Mag., (6) 42, 4 3 6 (1921). caesium possesses a negative packing and when such is the case, the differences between the true weight of the atoms of an element and an integer, are seldom as large as in the case of caesium. The recent search for element \" 8 7 \" has led to a renewed interest in the atomic weight of caesium. This unknown member, closely related to the latter as it is, has as yet defied (1) discovery; in spite of numerous attempts. Also the \"packing (2) effect\", as discussed by Aston in a recent paper, has thrown much light upon the atomic weight and structure of many elements; caesium included. Re has defined the packing frac-tion as the divergence of the mass of an atom from a whole number, divided by its mass number. By means of the mass-spectra of many elements, he was able to show that the whole number rule, based upon hydrogen as unity, did not hold in all cases. The failure of this law was attributed to the fact, that the protons and electrons in the nucleus were packed so closely together, that their electromagnetic fields interfered, resulting in the loss of a certain fraction of the combined mass. However, outside of the nucleus, owing to the great distances between the charges, nothing such as this, could happen. Therefore, a great loss in energy, corresponding to tight packing of the electrons, would mean great stability and visa versa for a small loss of energy. Caesium is situated in (1) L.M. Dennis and R.W.G. Wyckoff, Journ. Amer. Chem. ^oc., 42,983(1920). Herszfenkiel,uompt.rend.,184,968(1?27). (2) Aston, Proc. Roy.Soc. London, II3A, 4 8 7 ( I 9 2 7 ) . the region of tight packing, since it is only slightly removed from the minimum of Aston's curve. Recently, Millikan and (1) Cameron have concluded, that at one end of this curve the light elements were built up with a loss of mass, corresponding (2) to tight packing of the electrons; and Cabrera concludes that the heavy elements, situated at the other end, break down into lighter elements. On considering all these facts, we saw that a redeter-mination of the atomic weight of caesium might be of some avail; especially when no work had been done in this field for twenty-four years. (As this work was started in 1 9 2 7 , Richards' and Francon's paper had not been published.) Before continuing I wish to express my sincere thanks to Dr. E.H. Archibald, not only for the loan of the material; but also for his liberal and useful advice, as well as the purification of some of the caesium salts used in the analyses. II. SYNOPSIS OF EARLIER DETERMINATIONS. Since the discovery of caesium in 1860 by Bunsen and Klrchhoff, only six investigations as to its atomic weight have (3) been made; two by Bunsen himself and one by each of the (4) (5) following: Johnson and Allan, Godeffroy, Richards and tl) Millikan and Cameron, Science, i\\i.S., b7, 401 ( I 9 2 0 ) . (2) Cabrera, compt.rend., 186,228(1928); 186,501(1928). (3) R. Bunsen, Zeit.anal.chem., 1,137(1862);Pogg.Annal.119,1(1862;. (4) Johnson & Allen, Am.Jour.Sci.R Arts ^2), 33,94. (3) codeffroy, Liebigs Annal. der chem., 181,176^1876). (1) ^2) Archibald, Richards and Francon. In the following year, 18bl, Bunsen published his first; determination of the atomic weight, ^he material used in this investigation was recovered from I30 tons of water, obtained from the mineral spring at imrkeim. The amount of caesium obtained from this large quantity of water was very small indeed, being approximately 12 grams in weight. The satisfac-tory purification of such a small amount of material is a very difficult procedure and probably accounts for the inaccuracy of the results. Bunsen purified his material by recrystallizing the chlorplatinate, which is not nearly as soluble as the corresponding rubidium salt. The first value for the atomic weight, as found by Bunsen, was 1 3 2 . 3 3 , which number was soon after rejected by the discoverer himself, since his caesium salt undoubtedly contained traces of rubidium. In 1862 Johnson and Allen of the Sheffield Scientific School, ixew Haven, having discovered a much richer source of caesium, began their work on the atomic weight of this element. The mineral lepidolite found in Hebron, Maine, contained about 3 % of caesium oxide and from this they prepared a mixture of lithium, sodium, potassium, rubidium and caesium chlorides. These chlorides were then treated with chlorplatinic acid, whereupon the caesium and rubidium were precipitated as chlor-platinates. in order to separate the rubidium from the caesium, (1J T.W. Richards and E.H. Archibald, Proc.Amer. Acad.Arts & Sciences, 3 8 , 4 4 3 , ( 1 9 0 3 ) . (2) W.Richards & M.Francon,Journ.Amer.uhem.Soc.30,21b2(ly28). - 6 -the chlorides were converted through the carbonates into the bitartrates. The acid tartrate of rubidium, being only one-eigth as soluble as the corresponding caesium salt, successive fractional crystallization left the mother liquor nearly free from rubidium. A portion of the salt thus obtained was treated with enough chlorplatinic acid to precipitate all the caesium as the chlorplatinate, which was then washed with water and reduced in a stream of hydrogen. Four determinations of the amount of chlorine in the chloride prepared as above were made for the purpose of finding the atomic weight. Silver nitrate was used to precipitate the chlorine and the resulting pre-cipitate was collected on filter paper. The results ranged from 132.89 to I 3 3 . I 3 ; the average being I 3 3 . O 4 if oxygen is taken as 16.00 and vacuum corrections applied. Very shortly after the publication of Johnson and Allen's paper, Bunsen came forward with the results of his second determinations. His method of purification was now very similar to that used by other workers, differing principally in the fact that several recrystallizations of the chlorplatinate were made, with subsequent analysis of each product. Alsa his method of analysis resembled that of Johnson and Allen. The three results so obtained were 1 3 2 . 9 4 9 , 133.04 and 1 3 2 . 9 8 , averaging 132.99* In 1863, Redtenbacher showed that there was a greater (1) Redtenbacher, Wiener. Acad. Anzeiger f., 1863, 39* difference between the solubilities of caesium and rubidium alums, than between the chlorplatinates; hence a better separation of one from the other could be obtained, if the former method was used, codeffroy made use of Redtenbacher's idea in preparing his pure caesium alums: (a) a caesium alum (pure), (b) caesium alum containing a trace of rubidium, (cj rubidium alum containing a trace of caesium. The latter two were treated with barium chloride and then antimony trichloride. The precipitate, after being washed with hydrochloric acid, was claimed to be free from rubidium. At was now dissolved in water, whereby all the antimony oxychloride separated as a residue. The solution was evaporated and ignited with ammonium chloride, thus removing the last trace of antimony by volatil-ization and leaving pure caesium chloride as the residue, in (1) I875 R. Godeffroy showed that the above separation of caesium from rubidium was not satisfactory. The pure caesium alum, (a), was dissolved in hot water and the aluminium precipitated with ammonia. The solutioh, after filtering, was evaporated to dryness in a platinum dish and the ammonium sulphate driven off by heating. The residue was dissolved in water and the remaining sulphuric acid was precipitated by barium chloride. After filtering off the barium sulphate, the solution was treated repeatedly with ammonia and ammonium carbonate, in order go eliminate all traces of aluminium and barium. The caesium (1) R. u-odeffroy, Joum. uhem. Soc., 28,bl2 ^1873). - 8 -chloride, thus obtained, was fused in a platinum dish; thus driving off any ammonium salts that might have been present. This fused mass when dissolved in water was found to be al-kaline, but the crystals deposited by evaporation were not hydroscopic. The chlorine was weighed as silver chloride and the four results obtained ranged from 132.50 to 132.71, averaging 132.57. The low result obtained by Godeffroy was probably due to the admixture of sodium salts coming from the apparatus employed in the complicated purification processes. Therefore the salt used in these analyses can hardly be re-garded as pure, and no doubt earlier investigators had much purer material. In 1903 Richards and Archibald, by using an entirely different method of purification, obtained a value, which up to the present has been the accepted atomic weight for caesium. (1) Wells showed that a very good separation of caesium from rubidium may be obtained by making use of the difference in solubilities of their trihaliae salts. Of these, the dichlor-ide,(CsClgI), is the best method for not only is the caesium salt from eight to ten times less soluble than the corres-ponding rubidium one, but also will crystallize below 70°C. in ( 2 ) a rhombohedral form,while the other metal gives the corres-ponding salt in the orthorhombic form only. This property of (3) heteromorphy adds greatly to the efficiency of the separation TT) Wells, Am.Jour.Sci.(3),43,17(1901); Chem.News,84,2184,Oct 4 (2) R.W.G. Wyckoff, J.Am.Chem.Soc.,42,1100(1920). (1901) (3) Richards & Archibald, Proc.Amer.Acad.Arts & 3ci.,38,447(1903) - 9 -of the two metals. The greater part of the caesium used in Richards' and Archibald's research was obtained from Pollucite found at Mt. Mica, Paris, Maine. From this material was pre-pared the dichloriodide, which after several recrystallizations from dilute hydrochloric acid, was spectroscopically pure. The crystals, after being dried, were placed in a porcelain crucible and heated in an electric oven to about 100^0. At this temperature the extra chlorine and iodine are slowly driven off, leaving the chloride as a white porous mass. The chloride, so formed, was placed in a platinum dish and dissol-ved in a minimum of warm water; after which hydrochloric acid gas, prepared from boiling concentrated aqueous hydrochloric acid, was passed into the solution through a platinum tube. After cooling, the mixture was decanted into another platinum dish; the crystals being washed with a little pure concentrated hydrochloric acid. The salt was redissolved in warm water, precipitated, washed as before and partly dried by suction in a platinum Gooch crucible. A second sample was prepared from the mother liquors from the first, in the same manner as pre-viously described. A third sample was prepared from an impure caesium alum. Barium chloride was used to eliminate the sul-phuric acid; the dichloriodied was obtained by adding an excess of aqua regia, followed by an excess of resublimed iodine, to the filtrate. The resulting trihalide salt was recrystallized eight times, after which it was converted into the normal - 10 -chloride. The purification process then proceeded in the same manner as Sample I. In order to eliminate all traces of moisture, without decomposition, the caesium halide was fused in an atmosphere of pure nitrogen in the well known \"bottling (1 ) apparatus\". Results were obtained with both caesium chloride and bromide; the chlorine and bromine being weighed as silver chloride and bromide respectively. The average atomic weight as determined by twenty-five analyses, which agreed very closely, was 132.81. The reliability of this figure is unquestionable, as there is no doubt as to the superior purity of the salt employed. In 1928 a redetermination of the atomic weight of caesium was undertaken by Richards and Francoh. They felt that there was too great a gap between the atomic weight of that metal, when determined by physical and chemical methods. Aheir material also consisted of pollucite found in Maine. The ore was first treated with hydrochloric acid; then converted to the sulphate by means of sulphuric acid, where it was recrystallized as alums. The pure caesium alum was treated with barium chloride with the subsequent formation of barium sulphate and caesium chloride. 'A'he chloride was now converted into the perchlorate by adding perchloric acid to the above filtrate. The latter was recrystallized and converted back into caesium chloride by heating in a platinum boat to a temperature of (1) Richards & Parker, Proc. Amer. Acad. Arts & Sci., 32,3^1M9bJ Z. anorg. uhem., 1 3 , 8H18?bj. I - 11 -800°u. This excessive temperature was necessary in order to decompose a trace of caesium chlorplatinate. The residue was filtered and recrystallized, after which it was ready for use. The method of analysis was similar to that used in 1903, with the exception that fusion of the caesium halide was carried out in an atmosphere of nitrogen and hydrogen chloride. The chlorine was weighed up as silver chloride; and from this four values were obtained, which ranged from I 3 2 . 7 8 3 to 1 3 2 . 8 3 2 , averaging I 3 2 . 8 O 9 . As closely as these results agree with the accepted value, nevertheless they can not be classed as reliable. In the first place, the investigators themselves admit the possible presence of sodium in their caesium salt; this would tend to make their values low. Also the alum method of purification is much inferior to the dichloriodide one; and again their caesium chloride may contain traces of extra chlorine, due to the fact that it was fused in an atmos-phere of nitrogen and hydrogen chloride. The last item caused us much trouble, especially in the case of the bromide. From previous determinations, one might expect the atomic weight of caesium to lie somewhere between I 3 3 . O and 132.b; probably nearer the former than the latter value, ^he (1) value usually selected by Ularke and others, I 3 2.8II, has been in accord with this conclusion and the present investigation shows it to have been fairly accurate. (1) F.W. ularke, The Constants of mature,Vol 34,^0 3 , P a r t V, 134(1910) - 12 -a ) III. CHEMICAL PROPERTIES OF THE CAESIUM SALTS. Caesium chloride is a colorless salt, crystallizing in anhydrous cubes, which are very soluble in both hot and cold water. The aqueous vapor pressure of a saturated solution of caesium chloride is slightly greater than that of the air in the average steam heated, American laboratories; hence the crystals are only deliquescent on damp days. The addition of alcohol to its solution precipitates much of the dissolved salt. In addition to this it has no marked tendency to form an acid salt and so is precipitated by hydrochloric acid. But, when the acid concentration is very high a canary yellow precipitate results, which is rendered colorless by fusing. We investigated this colored salt qualitatively and found free chlorine to be present, most likely in the form of a polyhalide of caesium. o Uaesium chloride fuses to a clear, limpid liquid at b4b losing a trace of its chlorine if moisture is present and hyd-rochloric acid absent, however, the salt shows no traces of alkalinity after being fused in absolutely dry air, especially if it had been recrystallized from hydrochloric acid, in this case, microscopic inclusions of the acid must have been present along with the water and in this way prevented hydrolysis of the salt on fusion. Uaesium chloride forms a double salt with antimonious chloride in concentrated hydrochloric acid. U-odeffroy made use of this fact in preparing pure r n L. Moser & E. Ritschei, Xeit.Anal.uhem. 70,184(1^27). (2) R. Godeffroy, J.Chem.Soc.,27,81b(1874):28,bl2(.l875). (3) C.E. Saunders, J. Chem.Soc., A II, b2,778^18^2). - 13 -caesium salts, but later gave up the idea, since rubidium was found to give an analagous series of double salts. Likewise, the insoluble nature of caesium chlorplatinate has been taken advantage of in preparing pure caesium salts, caesium also forms an extensive series of alums, which up to the present, has been the commonest method of purification. The rubidiu% alum is about four times as soluble as the corresponding caesium alum and if this was the only factor involved a very good separation could be obtained. But, both alums have a (1) common crystal form, which makes complete separation almost impossible. Both caesium and rubidium form an extensive series (2) of trihalide salts,which are relatively unstable towards moderate heat. (Jf these, the most important are the dichlor-iodides, which are heteromorphous below 70°G. and differ widely in relation to their solubility in water. Rence a very good separation of caesium from rubidium may be obtained by making use of the above facts. uaesium bromide greatly resembles the chloride in all respects, both chemical and physical; and also fuses to a colorless, limpid liquid at about 600°^. (1) u. Setterberg, Journ.chem.Soc., A,42,464(1882;. (2) H.L. We..s & S.L. Penfield, Journ. uhem. soc., A ii, b2, 773(1S?2). - 14 -IV. PHYSICAL CONSTANTS USED AS DATA: Substance CsBr. CsCl. AgBr. AgCl. Ag. Br. CI. Density (20^C)M.P. t.380 gms/cc. . (1) 6.473 \" Vac.Correct per gm. At. Wt. + 0 600 C. .00013 646°C. .00016 434°C. .00004 0 455 C. .00007 0 961 C. -.00003 ( 2 ) \" 1C7.880 (2) 79.916 ( 2 ) 35.457 V. PURIFICATION OF CHEMICALS. Bromine.- A volume of 100 c.c. of \"C.P.\" bromine was twice redistilled from a pyrex flask and each time the first and last 15 c.c. portions were rejected. The purified product was stored in a bottle which had fcrmerlv contained bromine. (3) Ioaine.- Iodine of the \"C.P.\"variety, was fused under a solution of calcium chloride, potassium iodide and hydrochloric acid. The iodine so obtained was Crushed, washed and drained. The resulting crystals were twice sublimed with barium oxide over a water bath and finally resublimed by themselves. Formic acid.- A volume of 100 c.c. of distilled water was added to 800 c.c. of 75 %, \"C.P.\" formic acid. The resulting mixture was distilled frog a quartz still; the first and last 100 c.c. portions were rejected. (1) Richards & Archibald, Froc.Amer.Acad.,38,466(1903);38,454(1903). (2) Table of Atomic Weights for 1928. (3) G.P. Baxter & L.W. Parsons, Journ.Amer.Cheqi.Soc., 44,577(1922). Musset, Journ.Chem.Soc., A.,60,392(1891). Meineke, Journ.Chem.Soc., Aii, 64,204(1893). I!!!',',',!<]) ay CD tn OJ - 13 -Hydrobromic acid.- The concentrated \"C.P.\" hydrobromic acid was distilled from a pyrex still and the middle fraction collected. The purified acid was stored in a bottle which had formerly contained liquid bromine. Synthesis of Hydrobromic acid.- Calcium nitrate of the best \"C.P.\" brand and weighing 75 gms. was dissolved in distilled water and the calcium precipitated as calcium carbonate by means of an excess of ammonium carbonate (/'C.P.\") solution. The precipitate thus formed was thoroughly washed after which it was dissolved in redistilled hydrobromic acid. The resulting - solution was evaporated to a volume of 40 c.c. Towards the end of the evaporation process the liquid became dark in color and gave off copious fumes of bromine. About 30 c.c. of doubly distilled bromine were dissolved in this calcium bromide solution, that is, after it had become cold. The resulting solution was diluted to a volume of 200 c.c. and the bromine was distilled off. The still was made of pyrex glass and a bottle, that had previously contained bromine, was used as the receiver, in this way any chlorine present in the pure bromine was left in the still as calcium chloride. Hydrogen, prepared from zinc and dilute \"C.P.\" Sulphuric, in a Ripp generator, (A), and purified by being passed successively through wash bottles containing, potassium hydroxide, (B), alkaline potassium per-manganate, ^C), alkaline pyrogallic acid, ^D), and mercuric chloride, (^ E), was bubbled through the especially pure bromine, - lb -(F), and then passed over heated platinized asbestos, (H), and finally into distilled water, (I), The apparatus, from the bromine bottle to the wash bottle containing distilled water, was made entirely of glass. The hydrogen was generated at a very moderate rate and was passed over the platinized asbestos, o maintained at 3 3 0 ^ .,(approx.j, for eight hours. The asbestos used in the experiment was the very best that could be obtained, nevertheless it was boiled for 3O min. in 30y. hydrobromic acid (\"C.P.\"), after which it was thoroughly washed with hot dist-illed water. The fibres of asbestos, thus obtained, were ignited for half an hour at white heat, after which they were made into a thick paste with distilled water and into this was sprinkled a liberal amount of platinum black. The resulting mixture, when of a homogeneous appearance, was rammed tightly into a 12\" piece of 1/4\" glass tubing. At either end of the column of platinized asbestos was a similar one, (G), of pure asbestos 2\" long. Drying of the material was accomplished by suction and gentle heat, in fact, so tightly was the asbestos packed that suction had to be applied before the hydrogen could be forced through. At the end of an eight hour run the distilled water was so saturated, with hydrogen bromide, that it fumed strongly when exposed to the air. tseyond any doubt, the hydro-bromic acid, prepared in this manner, was very pure. hydrochloric acid.- The concentrated \"C.P.\" hydrochloric acid was diluted with an equal volume of distilled water to - 17 -which, was added two crystals of potassium permanganate, which liberated chlorine and so displaced any bromine present in the acid. The volume of this solution was 2000 c.c. The process of distillation was carried on in an all quartz still; the first and last 200 c.c. portions were rejected, ^he purified acid was kept in a Winchester which had formerly contained hydrochloric acid. nitric acid.- To 1J500 c.c. of the concentrated nitric acid (\"O.P.\") was added I30 c.c. of distilled water, the resulting mixture was distilled in an.all quartz apparatus; the first and last I30 c.c. portions were rejected, The purified acid was tested in a mephelometer for traces of chlorine, but none was found. The resulting acid was kept in a bottle which had formerly contained nitric acid. sulphuric acid.- Concentrated sulphuric acid, of \"C.P.\" brand, was distilled from a pyrex still; the first and last fractions were rejected. Any foreign substances which were lively to contaminate our pure sulphuric acid when distilled from a glass apparatus were completely removed by an elaborate system of recrystallization, since our last crop of caesium chloride crystals showed no evidence of containing any impurity. Ammonium hydroxide.- The concentrated \"u.P.\" ammonium hydr-oxide was heated in a quartz still and the resulting ammonia gas was bubbled into distilled water in a quartz flasn. The middle fraction of ammonium hydroxide obtained in the above - 18 -manner was collected. Barium nydroxide.- The barium hydroxide, of \"C.P.\" brand, was twice recrystallized from hot distilled water and filtered both times while still hot. Ethyl alcohol.- A volume of 1200 c.c. of the \"C.P.\" ethyl alcohol was distilled from a pyrex flask; the first and last 100 c.c. fractions were rejected. Ammonium formate.- Ammonia gas, obtained by heating our purified ammonium hydroxide in a quartz still, was passed into redistilled formic acid, contained in a quartz flasR, until the solution turned into a mass of yellow crystals, which smelled strongly of ammonia, enough distilled water was added to ma^e a saturated solution of the ammoniuip formate. Barium chloride.- Barium chloride, of \"C.P.\" variety, was dissolved in enough water to make an almost saturated solution at 100°G. it was then cooled to about and to it was added an equal volume of redistilled water and treated again as above. The second crop of crystals were washed and dried by suction in a u-ooch crucible after which they were kept in a tightly .corked flask. invert Sugar.- Pure cane sugar was recrystallized cnce from distilled water. The crystals so obtained were about 1/4\" -l/2\" in diameter and were perfectly transparent and very hard. The sugar was washed and dissolved in a minimum amount of (1) L.R. Parks and P.O. Bartlett, Journ.Amer. Chen.Soc. 47, lby8(l?27). - -distilled, water. To each litre of solution was added 30 c.c. of redistilled hydrochloric acid and the mixture kept at 30°c. for two hours. In two days the solution, now invert sugar, was ready for use. charcoal boat.- The charcoal was made from recrystallized sugar by carefully igniting a ground up portion of the latter in a porcelain boat until no more fumes came off. The carbon thus obtained was ground to a fine powder ahd made into a paste by adding 3 c.c. of a solution composed of one part of recrys-tallized sugar to five parts of distilled water. The paste was moulded between two porcelain dishes of slightly different sizes, in this way the boat can be made to any desired thick-ness. The sides should be at least 1/4\" thick. Without remov-ing the inner dish, the mould was transferred to a desiccator, in which it was left until dry. It was then removed and heated very slowly over a gas flame until no more fumes were evolved. After cooling, the inner dish was removed, leaving only the outside one in place. The carbon mould was kept in a desiccator until ready for use. A few small cracks in the boat were of no importance, as the silver, when molten, tended to coalesce into a ball. Lime boat.- Pure calcium nitrate was recrystallized from distilled water, redissolved and the calcium precipitated as calcium carbonate by means of equal volumes of ethyl alcohol (redistilled) and ammonium carbonate. The resulting crystals - 20 -were washed and dried. The calcium carbonate was now made into a thick paste and moulded between two porcelain boats of slightly different sizes. The mould was desiccated and dried, o at 100 ^ for two days, after which the inner boat was removed. The remainder of the mould was heated to redness for half an hour after which it was permanently kept in a desiccator. Since silver was to be fused in this boat, it was therefore necessary to ignite the latter in order to expell any residual nitric acid, which might contaminate the sample of molten silver with oxygen. (1) Silver.- Silver nitrate of the \"a.P.\" brand, was twice recry-stallized from distilled water to which had been added one c.c. of redistilled nitric acid. The resulting crystals were washed, as before, with 3 c.c; of distilled water, and dissolved in one litre of hot distilled water. To this was added a little at a time, with constant stirring, an equal volume of hot, . b 11., hydrochloric acid solution. The solutions were kept hot and very dilute so as to avoid as far as possible the occlusion of any silver nitrate in the precipitate. The silver chloride was washed first with hot water to avoid contraction of the mass, then with cold distilled water, until no appreciable test for (1; R.V.A. Briscoe, Journ.ahem. Soc., 63,1071,1713,!. T.W. Richards & R.C.Wells, Journ.Amer.ahem.Soc.,37,437(1703). C.P. Baxter & L.W. Parsons, Journ.Amer.ahem.Soc.,44,377^1722). G.P. Baxter, Journ.Amer.ahem.Soc., 44,371^1722). h. Rrepelka, Journ.Amer.ahem.Soc., 42,723(1720). N.R. amith, Journ. Amer.Chem.boc., 47,lb42(l727). - 21 -chloride could be noticed. ihe chloride precipitate thus obtained was removed to a large pyrex beaker, to which had been added about one litre of invert sugar solution made alkaline with sodium hydroxide. The mixture was heated to 30°^. and well stirred for 30 min. in this way the silver chloride was re-duced to a greyish-brown powder consisting of silver and silver oxide. The supernatent liquid was poured off and the residue when washed with distilled water, was removed. The mixture of the metal and its oxide was dissolved in a slight excess of redistilled nitric acid, to which was then added a quantity of ammonium formate which was sufficient to leave the mixture just acid. On gently warming, the silver nitrate was reduced to a buff colored residue. The silver precipitate was now thoroughly washed and dried; after which it was fused in a charcoal boat before a blast lamp, the nozzle of which was scrupulously cleaned, if the large buttons of silver, so obtained, are not ke-ot hot too long and are cooled in a reducing flame, they are (1) already very pure. Their purity suffices for ordinary work or even atomic weights of usual accuracy. Nevertheless they are not absolutely pure as they *nust contain about .OOI7, sulphur, (2) traces of carbon from the illuminating gas and oxygen from the air. Therefore when cooled, the silver most lively contains minute cells, enclosing gas. The buttons were now scrubbed (1) Richards, Proc. Amer. Acad., 29, 65(1893). (2) Richards and Wells, Journ.Amer.ahem.Soc., 3 7 , 477^1903). - 22 -with a brush and carborundum powder then washed with distilled water, etched with redistilled nitric acid, washed again in dilute pure ammonium hydroxide and finally in distilled water. The silver was next subjected to an electrolytic process of purification, in which every conceivable trace of impurity must have been removed, except for about . 0 2 - . 0 3 y. of the mother liquor. A button of the metal weighing about b gms. was dissolved in a slight excess of pure nitric acid and when diluted to 30 c.c. served as the electrolyte and was contained in a pyrex dish 3/8\" deep and 4\" in diameter. The silver was electrolyzed between platinum terminals maintained at b Volts for six weeks. The.silver crystals, some of them . 3 in. in length, were removed from time to time and kept beneath dist-illed water. The electrolyte was finally converted into the chloride to be used for next year's work. The silver crystals were thoroughly washed and in order to free them from all traces of mother liquor were finally melted for 7 min. on a lime boat in an atmosphere of hydrogen within a quartz tube. The hydrogen was prepared from zinc and sulphuric acid and was purified by being passed successively through strong solutions of barium hydroxide, alkaline potassium permanganate, alkaline (1) pyrogallate and mercuric chloride. Lime seems to be a very good substance for a fusion boat, because it is so easily *(T) A. Findlay^ Frac. fhy. Measurments, p.217^1^23). Koscue & Schorlemmer, Treatise on uhem. Non-metallic Elements, Vol.i. p. 6^3. H.V.A. Briscoe & H.F.V. Little, ^ourn.uhem.Soc.,103,l3l0s,lyl4). - 23 -obtained pure and is so difficult to reduce, that no metallic calcium could be dissolved by the silver; of course the latter must be saturated with dissolved calcium oxide, but with an a ; amount wholly negligible ^less than .0001% uaO). nowever, since lime is more or less friable, the buttons may become coated with the former; but due to the high surface tension of silver, the lime remains at the top and can be easily removed. Therefore the silver after fusion must be very pure and Stas states, that the amount of hydrogen dissolved in the silver is not over .0004?.. The buttons, so obtained, were filed with a new file, until all the former surface was removed; then washed with distilled water, etched with pure nitric acid, washed with pure dilute ammoni.im hydroxide and finally with distilled water, if the buttons were not already of the desired size, they were cut into pieces by means of a clean steel chisel before etching. The silver buttons were now placed in a quartz boat in a quartz tube and heated to 400°a. in a vacuum of 10 m.m. Rg. for 13 min., after which they were kept over potassium hydroxide until ready for use. More silver was needed for the following year and the process of purification was the same as before, with only two minor changes, first: the boat or dish containing the elect-rolyte was now composed of quarts, instead of pyrex glass. (1; Richardson and '..ells, tiourn.Araer.ohem.soc.,37,4771^1903;. - 24 -Second: the hydrogen was carefully dried by means of calcium chloride and phosphorous pentoxide tubes, before passing over the molten metal. Also another sample of silver, \"B\", was prepared by recovering the fused silver chloride buttons from our analyses. These were twice reduced by the invert sugar and ammonium formate processes and were not electrolyzed at all. The remainder of their treatment was the same as that given to the other samples. Purification of uaesium chloride.- To begin with , our caesium was in the form of a carbonate, whose original weight was 113 gms. The sample was dissolved in distilled water and the precipitate filtered off. The latter on examination, was found to be only silica. The solution was now made acid with redistilled hydrochloric acid, thereby converting the carbonate into a chloride, which was again filtered to remove the silica, then evaporated to 200 c.c. Enough pure iodine and hydrochloric acid were added to precipitate one half the amount of caesium in the solution; no visible reaction occured. Chlorine, from an all glass generator with grease free stop-cocks, prepared from potassium permanganate and conc. \"C.P.\" hydrochloric acid, was passed into the solution; thus precipitating the orange red caesium dichloriodide. The supernatent liquid was decanted off and a quantity of pure iodine, sufficient to precipitate one half the remaining caesium, was added. A further precipitate, - 27 -of orange red color, was obtained as more chlorine was passed in. When filtered, the mother liquor was evaporated to one quarter of its former volume and the solution again saturated with chlorine, after enough pure iodine had been added to precipitate the remainder of the caesium. The resulting crops of crystals were united, washed, dried by suction and divided into three portions. Each of the above lots was dissolved separately in a minimum amount of hot, distilled water; to which had been added 3 c.c. of pure hydrochloric acid and a few crystals of pure iodine. When cool, the solutions were filtered by suction and the resulting precipitates washed and then united, to give crystals \"A\". The three filtrates were united and evaporated to 100 c.c., whereby a crop of crystals was obtained. \"Table of Recrystallization of Caesium Dichloriodide.\" Precipitate crystals by means of chlorine. Redissolve these crystals in hot R20(HC1) giving:-Crystals \"A\" (164 gms.) Mother liquor Ml^ Evaporate Ml^ to 100 c.c. giving:-Crystals \"B\" ( 3 3 gms.) Mother liquor Ml^ Redissolve \"A\" in hot HpOtHCl) giving:-Crystals 1* - 26 -Mother liquor Ml 3 Evaporate Ml to 100 c.c. giving:-3 Crystals 1\" Mother liquor Ml (Keep 3 c.c. for testing). 4 Unite crystals 1' and 1\" giving (1) Unite Ml^ and crystals \"B\" giving:-Crystals 1 a Mother liquor Ml-Redissolve (1) in hot R^O(HCl) giving:-Crystals 2' Mother liquor Ml^ Evaporate Ml^ to 100 c.c. giving:-Crystals 2\" Mother liquor Ml 7 Unite crystals 2' and 2\" giving (2) Unite 1 a and Mly giving:-Crystals 2 a Mother liquor Mlg Redissolve (2) in hot H20(HC1) giving:-Crystals 3' Mother liquor Ml^ Evaporate Ml^ to 100 c.c. giving:-erystals 3 \" Mother liquor Ml-,,-, - 27 -Unite crystals 3' and 3 \" giving (3) Unite crystals 2 a and Ml-^ giving:-Crystals 3 a Mother liquor M l ^ Unite (3) and 3 a giving III (126 gms.) Unite Mlg, Ml^, Mlg, and Ml-^ Redissolve III in hot ^ ( H C l ) giving:-Crystals 4' Mother liquor Ml^g Evaporate M l ^ to 100 c.c. giving:-Crystals 4\" Mother liquor Ml^^ (Keep 3 c.c. for testing). Unite crystals 4' and 4\" giving (4). Redissolve (4) in hot H^O(HCl) giving:-Crystals 3* Mother liquor M l ^ Evaporate Ml^^ to 100 c.c. giving:-Crystals 3 \" Mother liquor Ml. Unite crystals 3' and 3 \" giving (3) Unite Ml^^ and Ml^^. heat and filter off sili Evaporate Ml^^ and Ml^^ to dryness giving:-Crystals 3 a Unite crystals 3 a and ( 3 ) giving ( 6 ) . - 28 -Before subjecting the caesium dichloriodide to recrystallization 5 c.c. of the filtrate, which should contain one quarter of the caesium chloride, was evaporated to dryness and ignited at 200° U. in a quartz boat, until it was a pure white color. On being examined with the spectroscope, the only impurity found was sodium; ho trace of potassium, reubidium or lithium could be seen. Further, on treating 10 c.c. of the combined mother liquors, MI2, Ml^, Mlg, and Ml^, in the same way, sodium was the only impurity found. Also the 5 c.c. of Ml^ was examined spectroscopically and no impurities were found. Therefore, sodium was eliminated somewhere between MI2 and Ml^. The 3 c.c. from Ml-^ also gave negative results, which showed that our caesium dichloriodide was pure as could be prepared. The crystals of caesium dichloriodide ((6)) were now placed in a large, quartz evaporating dish and gently heated, so as to expell all the iodine. During their heating, they were frequently stirred and once in a while transferred to a mortar and gently ground, in order to facilitate the reaction and to ensure a homogeneous product, which after two days heating, was a pure white color and, weighed bl gms. The caesium chloride thus obtained was dissolved in 30 c.c. of distilled water and filtered to remove silica, nydrogen chloride, generated by the action of conc. \"U.P.\" sulphuric acid upon - 29 -conc. \"C.P.\" hydrochloric acid, in an all glass apparatus, was passed into the solution; whereby caesium chloride was precipitated by the common ion effect. Recrystallization of Caesium Chloride. Dissolve crystals A in 30 c.c. hot R^O giving:-Solution Pass HC1 into giving:-Crystals (1)' Mother liquor MI2 Evaporate Mlp to 20 c.c. and pass in RC1 giving:-Crystals (1)\" (yellow) Mother liquor Mlgt Unite crystals (1)' and (1)\" giving (1). Redissolve crystals (1) in hot H^O and pass in RC1 giving:-Crystals (2)' Mother liquor Ml^ Evaporate Ml^ to 100 c.c. giving:-Crystals (2)\" (yellow). Mother liquor Ml-! Unite crystals (2)' and (2)\" giving (2). Redissolve crystals (2) in hot R^O and pass in HC1 giving:-Crystals (3)' - 30 -Mother liquor Ml, 4 Evaporate Ml to 10 c.c. giving:-4 Crystals (3)\" (.yellow). Mother liquor Ml, 4* Unite crystals (.3)' and (,3)\" giving ( 3 ) . Redissolve crystals ( 3 ) in hot R^O and pass in RC1 giving:-Crystals ^4)^ Mother liquor Ml 3 Evaporate Ml to 10 c.c. giving:-3 Crystals (4)\" (^yellow). ^Set aside for testing;. Mother liquor Ml-, P' Redissolve crystals ^4)^ in hot R2O giving:-Crystals (,3) Mother liquor Ml b Redissolve crystals (,3) in hot H^O giving:-Crystals ^b) Mother liquor Ml^ The crystals ^6) were dried by suction and kept in a desiccator over calcium chloride. The canary yellow crystals viz. ^4)\" were tested, with-out result, for iodine and iron, We now suspected the presence of an acid salt and took steps to verify our suspicion. - 31 -Analysis: Weight of oxalic acid Weight of sodimp hydroxide I7.3 c.c. of xN. Sodium hydroxide = .1302 gms. = I.3OO gms. = .1302 gms uxalic acid. . . Normality of sodium hydroxide = . 1 3 6 3 n. Weight of caesium chloride = .1643 gms. But .1 c.c. of . I 3 6 3 N. sodium hydroxide are necessary for neutralization. this shows that no acid salt was present; but just a trace of residual hydrochloric acid. However, on heating the yellow salt the color disappears and does not return on cooling. Further, when the latter were tested with starch potassium iodide solution, we obtained a positive test for the presence of free chlorine. It now seems probable that some chlorine was actually in feeble combination with the caesium chloride molecule and the formula for the salt will be of the nature of CsCl^Cl cf. (.CsCl^l). An Examination of the Yellow Caesium uhloride.- The yellow colored caesium chloride previously mentioned and prepared by passing hydrogen chloride into a solution of caesium chloride was 30W investigated in larger quantities. About a gram of this salt was placed in a clean porcelain boat within the fusion tube of the \"bottling apparatus\". After reassembling the apparatus, the salt was heated so as to just fuse in an atmosphere of pure, dry nitrigen and hydrochloric acid. The - 32 -chloride was kept in a molten condition for about 2 min., after which time the hydrochloric acid bottle was cut out, leaving only a stream of pure nitrogen passing through the apparatus. The salt was still maintained at its melting point in the atmosphere of nitrogen for an additional 3 min., after which it was allowed to cool. This fusion process was performed in the same manner and with the same care as was given to any atomic weigh analysis. However, subsequent chloride analyses did not employ an atmosphere of both nitrogen and hydrochloric acid; but only the former. When cool, the boat was removed and its contents inspected. The original canary yellow color of the caesium chloride had not completely vanished, as the salt near the middle of the boat was still of a faint yellow tinge. As a result the boat and its contents were replaced in the fusion tube and melted again for 10 min. When cooled, the now color-less salt was dissolved in 20 c.c. of redistilled water and the resulting solution was found to be absolutely neutral. There-fore no error could be introduced into an atomic weight determination by using the yellow, instead of the pure white, caesium chloride. Recovery of Caesium Chloride and its Conversion into .Bromide.-The caesium, after the analysis, is in the form of nitrate, being contained in the filtrate along with some silver nitrate. The latter is almost completely removed by the addition of a small quantity of pure hydrochloric acid and subsequent filtration. - 33 -However, the solution is not free from silver as silver chlor-ide is slightly soluble in a salt solution and quite soluble in strong hydrochloric acid or brine solution. The combined filtrates were evaporated to a small volume, diluted to 230 c.c. and left over night. On examination, a large mass of long,white, silky,needle-like crystals were found to have been deposited. The supernatent liquid was decanted off and treated with pure hydrochloric acid; whereby, a white precipitate of silver chloride formed. Some of the above mentioned crystals were kept for examination, the remainder dissolved in water. The rate of solution of these crystals was very slow, even in 23O c.c. of water. The decanted liquid and the above mentioned solution were saturated separately with hydrogen sulphide. The latter was made from iron pyrites and dilute \"C.P.\" hydrochloric acid and before being passed into the solution was thoroughly washed with water, in order to remove any soluble impurities contained in the gas. The solutions were filtered free from silver sulphide and again saturated with the gas; but with no result. The filtrates were now united and evaporated to 73 c.c. upon cooling, the same needle-like white crystals were deposited, similar to those mentioned before. 1 3 c.c. of conc., redistilled sulphuric acid were added and the mixture evaporated, until the white fumes of sulphur trioxide were evolved. The resulting compact, crystalline, mass was dissolved in water, containing 1 c.c. of pure sulphuric acid and again evaporated as above and the process repeated. The crystals so obtained were dissolved in o 500 c.c. of distilled water and heated to 80 C., whereupon a hot solution of recrystallized barium chloride was added, until no further precipitate formed. The solution was filtered and to the filtrate was added a dilute solution of pure sulphuric aoid, until precipitation ceased. The solution, containing caesium chloride and sulphate; but no barium salt in dissolved form, was filtered; then evaporated to dryness. The salt was covered and placea under a bell-jar for the summer months. We now decided to work with the bromide instead of the chloride, as we had done last year; hence our conversion from chloride to bromide. The caesium chloride was dissolved in distilled water and added to the mother liquors of last years chloride purifications, viz. Ml III, IV, and V, which contain nothing but the chloride and dichloriouide of caesium. In this way no foreign impurity was annexed. To the resulting solution 19 c.c. of conc. pure sulphuric acid were added and evaporation continued, until white fumes were evolved. The crystals were o dissolved in 400 c.c. of distilled water at 80 C. and to this was added an excess of a hot solution of recrystallized barium chloride. After filtering, the solution was treated with a slight excess of pure sulphuric acid and again filtered. The solution was made acid with hydrobromic (conc. and redistilled) with an excess of 15 c.c.; filtered ana evaporated as below. 35 -Recrystallization of Caesium Bromide. Evaporate the original solution of CsBr givingt-Crystals 1 Mother liquor Ml^ ( 10 c.c.) Dissolve crystals 1 in hot H20(E Br); filter and evap-orate giving:-Crystals 2 Mother liquor Hl^ ( 7 c.c. ) Dissolve crystals 2 in hot HgO(H Br); evaporate to o dryness; heat to 120 C. for 4 hrs; dissolve in KgO and filter off silica; add 10 c.c. of HBr and evaporate giving:-Crystals 3 Mother liquor Mlg ( 5 c.c. ) Dissolve crystals 3 in hot HgO and evaporate giving:-Crystals 4 Mother liquor Ml^ ( 5 c.c. ) Dissolve crystals 4 in hot HgO and evaporate giving:-Crystals 5 Mother liquor Ml^ ( 5 c.c. ) Dry,crystals 5 in a desiccator over phosphorous pentoxide. The drying process lasted about 3 weeks, during which time many peculiar things took place. The caesium bromide first turned a - 36 -yellowish red around the edges and the crystalline mass a mauve (1) color. A few days later a brown crust formed over the surface, covering a pure white interior and when the lid to the desicc-ator was removed, dense clouds of hydrogen bromiae were liber-ated. Seeing the possibility of an acia. salt, a further crop of crystals was obtained by evaporating Hlg to 3 c.c., at which volume the solution became a beautiful, light purple color. The solution was cooled, filtered by suction and drieo. in this manner for one hour; the crystals were a pure white color. A portion of the crystals, weighing .2014 gms., was neutralized by .3 c.c. of .1363 N. sodium hydroxide; thus eliminating the possibility of any acia salt. Since the bromide had lost hydrobromic acia, even after being twice recrystallizea from water, we suspectea that the brown portion might be alkaline. However, on titration, this portion was founa to possess the same aegree of aciaity, as the sample previously mentionea. But when a small amount of the brown resiaue was aissolvea in water a transparent sanay precipitate was left behina. This meant one of two things, either silica was present or phosphorous had combined with the caesium to give a phosphide. The latter is decomposea by water, liberating small amounts of phosphine ana hyarogen ana aepositing a hyariae of phosphorous. The solution was immeaiately testea for the latter; but with no result. The (1) H.L. Wells ana S.L. Penfieia, Journ. Chem. Soc., A II 62, 773(1892). - 37 -whole sample of caesium bromide was now heated in an oven at o 125 C. for 2 hours, after which it was dissolved in distilled water and filtered. The residue was at first a brown color, but in 5 minutes became colorless, evidently, having lost bromine. The filtrate became a milky color when diluted and hence was again filtered; further dilution producer no precipi-tate. The solution was evaporated to a small volume giving:-Crystals 6 Mother liquor Ml^ The caesium bromide was filtered and dried by suction for an hour; the so formed crystals were perfectly white. The latter were removed to a vacuum desiccator and kept at a very low pressure for two days. Repurification of Caesium Bromide.- The repurification process for caesium bromide was, in most respects, very similar to the procedure used in its purification. During the process, whenever we saw the possibility of an objectional chloride or sulphate of caesium being present a recrystallization was made. In this way we made sure, beyond any aoubt, that our final product was absolutely pure. Also great care was taken to purify all the chemicals used in the procedure from chlorides an sulphates. All the unused caesium bromide was added to the solution of caesium nitrate, containing a small amount of dissolved silver bromide, obtained from the previous unsuccessful bromide - 38 -analyses. The resulting solution was evaporated to 100 c.c., during which the liquid assumed a dark red color and gave off abundant fumes of bromine and hydrogen bromide. The solution became a pale yellow color as its volume approached 100 c.c.. It was then diluted to 250 c.c. and saturated with hydrogen sulphide prepared from iron pyrites and dilute \"C.P.\" hydro-chloric acid in a Kipp generator. Before being passed into the caesium solution the gas was twice washed with water. The scanty greyish precipitate forged was filtered off and the filtrate again saturated with the gas, but with no result. A volume of 30 c.c. of concentrated, redistilled sulphuric acid was added and the solution evaporated until the white fumes of sulphur trioxide were evolved for at least five minutes. This process was repeated three times and each time an additional 2 c.c. of the concentrated acid was anded to compensate for that which was lost during the evaporating process. The last evap-oration was continued for an hour after the first white fumes made their appearance. When cool, the viscous solution was diluted with 10 c.c. of distilled water; immediately a white precipitate formed. The mother liquor, containing any unde-composed caesium chloride, was drained off. The resulting crystals, crystals \"A\", were dissolved in 600 c.c. of hot distilled water, to which was then added a solution of twice recrystallized barium hydroxide, until no further precipitate formed. The barium sulphate thus formed was filtered off and to t&e filtrate was added a dilute solution of pure sulphuric acid, until the latter was present in a slight excess. The solution was then filtered and evaporated in a large platinum dish to a volume of 135 c.o. During the evaporation a white crystalline precipitate was deposited, presumedly barium sulphate. After the solution of caesiuw hydroxide had cooled, a few c.c.^ of pure barium hydroxide were added to get rid of the slight excess of sulphuric acid. There was no danger of having adaed too much of the former, as much of the caesium hydroxide had been converted into the carbonate during the evaporation, and by virtue of this took care of an excess of the latter, precipitat-ing it as barium carbonate. The solution was filtered and to o the filtrate at 40 C., was added enough doubly distilled bromine to leave the solution highly colored. A white crys-talline precipitate, probably caesium bromate, was formed which dissolved when the solution was heated. Evaporation was contin-ued until the solution reached a volume of 50 c.c.; it was then cooled and to it was added five drops of bromine. The resulting mother liquor was poured off and along with it went any trace of barium bromide or sulphate that pight have been present, leaving the crystals, *B\", as pure caesium bromate, bromide and hypobromite along with some silica. The above mentioned crystals were placea in^a porcelain evaporating dish and heated to a temperature of 350°C. for 20 min., after which they were fused for 5 min. The fused mass was of a clear color and no fumes of - 40 -bromine or hydrogen bfomice were given off. The resulting product was known as crystals \"B\",. These were dissolved in distilled water and the silica filtered out. The filtrate was acidified with 2 c.c. of synthesized hydrobromic acid of our own preparation and a crop of crystals obtained. Table of the Recrystallization of Caesium Bromide. From conc. H-SO obtain:- - H * * * * * * o .H CM CM -M M eg * * * * * * > tQ t o c\\I to to to S-t d o a # r-4 p- t o r-t 00 s t o co ^ r-t -F 43 3 t o to c - E- ^ CM Mr-t 3 tQ tO O tO CO to to ! r-t g * > CM to r-t to CM CO CO ^ E- cy* * CM t-t W t-t t-t E- CO - 36 -fusion chamber of the apparatus was thoroughly washed and dried before being used for the bromide. The fusion was still carried out in an atmosphere of pure dry nitrogen. The first sample was fused in a platinum boat, but the salt on melting turned a dark, reddish-brown color; it was therefore kept in a molten condition for five minutes, during which time the salt became much lighter in color and gave off abundant fumes of hydrogenbromide and bromine. The bromide was bottled in the usual manner and upon examination was found to be a dirty, greyish-black color. When brought into solution with water a very fine, black precipitate settled out; but the solution, on being tested with indicators, was neigher acid nor alkaline. However, the platinum boat had lost .00283 grams during the process. Since it was impossible for us to use a platinum boat we then chose a quartz one. As was expected, the bromide still turned a dark brown color, but upon five minutes further heating was absolutely colorless. Caesium bromide, just like the chloride, contracts considerably upon cooling and as a result our quartz waw badly fractured; in fact it was recovered by filtration, followed by ignition and weighing. The salt, when cooled, was a beautiful, white, crystalline mass. As a last resort a porcelain boat, placed on a piece of plat-inum foil, was used for the fusion, and served the purpose splendidly; for it was neither crached nor attacked by the molten salt. - 37 -Since our platinum boat was badly corroded by the molten salt and since the latter, when dissolved in water, was neutral to indicators; we came to the conclusion, that the caesium bromide contained a considerable amount of extra bromine in combination. The peculiar behavior of the salt when drying over phosphorous pentoxide also supports our conclusions as to its composition. One would expect the salt to become alkaline upon loosing any of its constitutional bromine; but since the fused salt was neutral, then the bromine liberated must have been extra or added bromine. Fusion appears to be the one and only way to get rid of this additional bromine, as we had tried nearly every other way and failed. The fused salt was now weighed and was found to have lost twenty milligrams during the fusion. No doubt the greatest part of this was caused by the loss of some of the bromide by sumlim-ation, as a white coating was noticed on the inside of the fusion tube. Precipitation was accomplished in the dark room and the resulting mixture allowed to stand overnight, as had been done in the case of the chloride. The solution was then decanted through a platihum crucible provided with a filtering mat of platinum and iridium. The silver bromide was shaken for an hour with 73 c.c. of redistilled water, after which it was decanted off and added to the first solution. The operation was again repeated, after which the bromide was assumed to have been thoroughly freed from any excess silver or caesium nitrate. The pretty, canary yell&w silver bromide was treated from here just - 38 -exactly as was the chloride. The precipitate of silver bromide is of a fine, granul-ated nature and does not coagulate as does that of silver chloride. As a result, in our second analysis, some very fine particles of the bromide were noticed to be passing through the platinum-iridium filtering mat. However, the filtrate on exam-ination appeared to be perfectly clear; yet the platinum cruc-ible, when dried on a watch glass, deposited a thin, yellow ring of the bromide. This showed conclusively that the silver pre-cipitate was being lost; and in order to counteract this, two layers of filter paper were placed upon the mat before filter-ing. This latter procedure proved to be very effective. An equal amount of filter paper must be added to the tare crucible in order to obtain accurate weighings. The results from four analyses proved to be somewhat discouraging as they fell womewhere between I33.322 and 133.338. As far out as the values were, yet nevertheless they agreed remarkably well among themselves. Since all possible precautions had been taken this shows that some constant error was present in all four determinations. We could not say for sure what this was, but thought that it might possibly be a trace of chlorine. The only place at which the latter could be introduced would be along with the hydrobromic acid, which had been previously distilled in order to eliminate any trace of chlorine. The only way out of our difficulty was to repurify the caesium bromide - -which we did with the utmost care. The purpose of purification was twofold; first, to eliminate any impurity, chlorine included; second, to obtain the bromide in such a condition that platinum and not porcelain vessels could be used for the fusion. The process of analysis for the repurified bromide was, with the exception of the fusion, exactly the same as that which had been used previously. The fusion of the caesium salt was now made in an atmosphere of nitrogen and hydrobromic acid; the latter being used to prevent any possible decomposition of the (1) molten salt. The exact procedure used was as follows:' the caesium bromide, contained in a platinum boat, was placed in the pyrex fusion tube; after which the remainder of the apparatus was assembled in the customary way. A current of pure nitrogen, bubbled through concentrated redistilled hydrobromic acid, and dried by means of freshly fused calcium bromide, was then passed through the apparatus for 20 min., When all the air had been swept out of the lat.er, the caesium bromide was heated to about 130°C. for an additional 20 min., after which time the temperat-ure was increased so that the caesium bromide just fused in the atmosphere of nitrogen and hydrobromic acid. While the bromide was still molten, the hydrobromic acid bottle was cut out; thus allowing a stream of pure, dry nitrogen to pass over the salt, which was kept fused for 3 min. before it was allowed to cool. The purpose of the last step was to prevent the hydrobromic (1) T.W. Richards & E.H. Archibald, Proc. Amer.Acad., 3 8 , 4b5 (1903). - 60 -acid from combining with the caesium bromide resulting in the formation of a polybromide of the metal. After the tube and its contents had cooled to room temperature a stream of pure, dry air was passed through the apparatus for another 20 min., after which the caesium bromide was bottled in the usual manner. This time the caesium salt fused to a clear, limpid liquid, without showing any signs of decomposition and when dis-solved in water gave a clear solution. IX. SUMMARY. The two salts of caesium have been prepared in a very pure form and the amount of chlorine and bromine combined with the former has been determined by a series of accurate analyses. The values obtained are as follows: From chloride analyses 1 to 6 I32.%8j? \" \" \" 7 \" 8 1 3 2 . 5 7 4 0 \" bromide \" 1 \" 4 132.3703 Mean value I 3 2 . 9 7 I O Therefore I32.57I represents the atomic weight of caesium as determined by analysis of the chloride and bromide of the metal. - 33 CO W o E-< CO o w co Ew 4-t o xo ox * Ox PQ <0 r-t a s <-t O +3 P O CO SttDtn o^ -t Q) CO O 5 O O 4-t o ' r-t Q) rW 6 M -H O > -r) O +3 p -r-t ^ -rt S !)D 03 O O CO ^ O CO ^ g O M S-t o o o O !! +S**pq co t)D m m u S-) o M o o o*. H -HiLtiL) m w 4-t o CO ^ a o oo CO o S-) CO o c{ -r-t M CQ g o M cO ,-) 3-4 K! M p^ o U ! m d *<-< 03 O -3{ - 62 -X. BIBLIOGRAPHY. 1. Aston, Phil. Mag.(6) 42, 4 3 6 ( 1 9 2 1 ) . \" , \"\" \" (6) 49, 1191(1923)., \" , Proc. Roy.Soc.Lon. 113,A, 487(1927). 2. G.P. Baxter & L.W.Parsons, Journ.Amer.Chem.Soc., 44, 3 7 7 (1922). 3 . H.V.A. Briscoe, Journ.Chem.Soc.,63,107(1913). 4. H.V.A. Briscoe & H.F.V. Little, Journ.Amer.Chem.Soc., 103, 1310(1914). 3 . R. Bunsen, Prog.Annal 119, 1(1862). \" \" , Zeit. anal. chem. 1, 1 3 7 ( 1 8 6 2 ) . 6. Cabrera, Compt. rend., 186, 228(1928). \" , \" \" 1 8 6 , 3 0 1 ( 1 9 2 8 ) . 7 . F.W. Clarke, The Constants of Nature, Vol. 3 4 , No.3, p t . v . , 1 3 4 ( 1 9 1 0 ) . 8. L.M. Dennis & R.W.G. Wyckoff, Journ.Amer.Chem.Soc., 42,983(1920). 9. A. Findlay, Pract.Phy. Measurments, pg. 2 1 7 ( 1 9 2 3 ) . 10. F. Godeffroy, Leibigs Annal der Chem. 181, 1 7 6 ( 1 8 7 6 ) . 11. R. Godeffroy, Journ.Chem.Soc., 2 7 , 8 1 6 ( 1 8 7 4 ) . \" \" \" \" \" , 2 8 , 6 1 2 ( 1 8 7 3 ) . 12. Herszfenkiel, Compt. rend., 1 8 4 , 9 6 8 ( 1 9 2 7 ) . 1 3 . Hulett, Ibid. 3 7 , 3 8 3 ( 1 9 0 1 ) . \" 47,337(1904). 14. Johnson & Allen, Amer.Journ.Sci & Arts, ( 2 ) 33,94. 15. H.C. Kremers & C.W. Blake, Journ.Amer.Chem.Soc., 40,601 (1918). 16. H. Krepelka, Journ.Amer.Chem.Soc., 42,923(1920). 1 7 . Meineke, Journ. Chem.Soc., A ii 64, 204(1893). 18. Millikan & Cameron, Science N.S., 6 7 , 401(1928). 1 9 . L. Moser & E. Ritschel, Zeit. anal, chem., 7 0 , 1 8 4 ( 1 9 2 7 ) . 20. Museett, Journ. Chem. Soc., A., 60, 3 9 2 ( 1 8 9 1 ) . - 63 -21. Oswald, Ztschr. Phy. Chem. 22, 307(1897). \" , Ibid, 34, 493(1900). 22. L.R. Parks & . P.G. Bartlett, Journ.Amer.Chem.Soc., 49,1698(1927). 2 3 . Redtenbacher, Wiener Acad. Anzeiger f., 1 8 6 3 , 3 9 . 24. T.W. Richards, Proc.Amer.Acad., 2 9 , 6 3 ( 1 8 9 3 ) . \" \" \" , Journ.Amer.Chem.Soc., 22, 144(1900). \" \" \" , Zeit. Ph s.Chem., 3 3 , 6 0 3 ( 1 9 0 0 ) . 2 3 . T.W. Richards & E.H. Archibald, Proc.Amer.Acad. Arts & Sci., 3 8 , 4 4 3 ( 1 9 0 3 ) . 26. T.W. Richards & M. Francon, Journ. Amer.Chem, Soc., 30,2162,(1928). 2 7 . T.W. Richards & Parker, Proc. Amer.Acad. Arts & Sci., 32,33(1896). 28. T.W. Richards & E.P.R. Saerens, Journ.Amer.Chem. Soc., 46,939(1924). 2 9 . T.W. Richards & R.C. Wells, Journ. Amer.Chem.Soc, 3 7 , 439(1903). 3 0 . Roscue & Schorleamer, Treatise on Chep. Non-metallic Elements Vol.1, pg.693. 3 1 . C.E. Saunders, Journ.Chem.Soc., A ii, 6 2 , 7 7 8 ( 1 8 9 2 ) . 3 2 . C. Setterberg, Journ. Chem.Soc,m A, 42,464(1882). 3 3 . N.H. Smith, Journ.Amer.Chem.Soc., 49,1642(1927). 3 4 . Wells, Amer.Journ.Sci.,(3), 4 3 , 1 7 ( 1 9 0 1 ) . \" Chem. News, 84,2184,Oct.4(1901). 3 3 . H.L. Wells & S.L. Penfield, Journ.Chem.Soc.,A ii, 6 2 , 7 7 3 (1892). "@en ; edm:hasType "Thesis/Dissertation"@en ; edm:isShownAt "10.14288/1.0059468"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Chemistry"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "An accurate redetermination of the atomic weight of caesium"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/22179"@en .