MOORE: THE DEXSlTIES OF KRYPTON AXD XENON. 2182CCXX.-The Densities of Kyypton and Xenon.l3y RICHARD B. MOORE, B.Sc.THE volume of krypton and xenon isolated by Ramsay and Traversduring their classical research on the rare gases of the atmosphereamounted to 12 C.C. of krypton and 3 C.C. of xenon. The two bestdensity determinations obtained by them for krypton were made ontwo samples of gas, one of which had been fractionated from argonand the other from xenon. The former gave a density of 40-83and the latter 40.73. Later, Ramsay, starting with 191.1 kilos. ofair, obtained 7.5 C.C. of krypton and 0.87 C.C. of xenon. This kryptongave a density of 40.81. To make an accurate density determina-tion with the original 3 C.C. of xenon was by no means easy, andthe two highest figures obtained, 63.64 and 64.0, agreed exceedinglywell in the circumstances.The density of xenon has generally beentaken as 64.0, and its atomic weight, on the assumption that it isa monatomic gas, as 128.0 (Phil. T~ans., 1901, 197, A , 66)2182 MOORE: THE DENSITIES OF KRYPTON AND XENON.Owing to the small volume of gas available for fractionation, it isprobable that the krypton in the above experiments contained smalltraces of argon, and that the xenon also was not entirely freefrom krypton. This source of error would give low results for thedensities of both gases. I n addition, the volume of the bulb usedwas only 7 C.C. Recently, during a search for possible new elementsi n the atmosphere (Proc. Roy. Soc., 1908, 81, 195), the authorfractionated in Sir William Ramsay's laboratory the residues from120 tons of liquid air, which gave krypton and xenon in sufficientquantities to make it possible to obtain pure samples of both gases.The density apparatus used was a slight modification of that designedby Ramsay and Travers (Phil. Trans., 1901, A, 197, 54), andis shown in Fig.1. It consists of a U-shaped gas burette fastened toa strip of mirror glass, on the surface of which is etched an accuratemillimetre scale. The tube B is of the same diameter as the upperportion of A, hence a correction for capillarity is unnecessary. Ccontains a roll of silver foil in order to prevent minute globules ofmercury, carried over by the gas, from entering the bulb E. K leadsto a Topler pump. The bulb E is connected with the capillary tube Hby means of a selected piece of pressure tubing wired on.I t wasfound that such a connexion remained perfectsly gas-tight duringa period very considerably longer than was required for an experiment.The volume ofthe sealed counterpoise was the same as that of the bulb. A long-armed Oertling balance sensitive to 0.01 milligram and carefullystandardised weights were used. These weights, although standardisedrelatively to each other, were made absolute, inasmuch as they wereused for determining the weight of the bulb full of wstor. Nocorrection for latitude has been applied.The method of manipulation was as follows: By raising the globeD the mercury was run into the burette up to the stopcocks a and b.The latter were then closed.The reservoir D was then lowered so asto leave a barometric vacuum in the tubes A and B, after which thestopzock d was closed. The density bulb had meanwhile been com-pletely exhausted and carefully weighed. It was then attached to thetube H. The vessel P, surrounding the bulb, was packed with groundice, and distilled water, previously cooled almost to zero, was poured inuntil the level of the water reached the upper surface of the ice. Thestopcocks a and m were then turned, the apparatus thus being put incommunication with the pump. I n this manner, the air in the bore ofthe stopcock a wa3 completely removed. After exhaustion wascomplete, a and m were closed and d opened, the mercury risingin the burette.The gas was then run into the burette through thecapillary tube By which was completely filled with mercury. OnThe density bulb used had a volume of 32.7077 C.CMOORE: THE DEKSlTIES OF KltYPTON AND XENON. 2183turning the stopcocks a and e so as t o put .1 into communicationwith E, the gas entered the bulb. The stopcock 6 mas then openedand the reservoir B adjusted SO that the level of the mercury in AKEHIlay close to the top of the tube. After half an hour, the pressure wasobserved, the stopcock e closed, and the barometer read. The gascontained in C and A w a s removed through the pump, and the globeVOL. sc111. 7 2184 MOORE : THE DENSITIES OF KRYPTON AND XENON.was suspended on the balance. Owing to the small size of the globe,no correction was made for shrinkage under atmospheric pressure.Xenon.--The xenon used in the density determinations bad beenseparated from the krypton by a long series of fractionations.Thegas thus obtained, after it had been sparked with oxygen and theoxygen removed with phosphorus, was condensed in the fractionatingbulb at the temperature cf liquid air. As krypton has a vapourpressure a t this temperature of 17 mm., and the vapour pressure ofxenon at the same temperature is only 0.17 mm., the two can beseparated by removing the krypton from the xenon by means of aTGpler pump. It was found during the preliminary fractionationthat a mixture of solid krypton and xenon could be apparentlypumped '' dry " and yet some krypton would be retained below thesurface of the solid xenon.On vaporising all the gas, however, andredepositing, this krypton could be pumped off. This process wastherefore repeated several times during the final attempt t o get ridof all traces of krypton. The gas then pumped off was apparentlypure xenon, as its spectrum did not show the slightest traces of theprincipal krypton lines. Nevertheless, it was rejected as probablycontaining traces of krypton.A portion of the pure xenon was then fractionated at - 130' bymeans of light petroleum cooled with liquid air. Four fractions wereobtained, and density determinations were made with fractions 2 and3. The results were as follows :Fraction 3.Fraction 2. c A\Volume of density bulb (in c.c.) ...32.7077 32.7077 32.7077Temperature ........................... 0" 0" 0"Hence, density (0= 16) .......... 65.253 65.380 65.328Pressure on gas, corrected (in mm.) 480.0 446'4 521-7Weight (in grams) 0.12044 0.1 1223 0'13106 ....................As the density of fraction 2 was lower than that of fraction 3, theformer probably still contained a very small trace of krypton, and theexperiment may therefore be rejected. It is probable that this tracewas contained in the first portion of fraction 2, and as the wholevolume of the fraction was 40 c.c., we may assume that any kryptonin fraction 3 would have no effect on its density within the limits ofexperimental error. The mean of the two determinations on fraction3 (65-35) may therefore be taken as the density of xenon.Krypton.-It is easier to obtain pure xenon than pure krypton.I nthe former case it is only necessary to free the gas from krypton, I nthe latter, both argon and xenon must be removed. In the fractiona-tion of a mixture of three gases, it is always easier to obtain puresamples of the gases which possess the lowest and highest boilingpoints than it is to obtain a similar sample of the gas with an interMOORE: THE DENSITIES OF KRYPTON AND XENON. 2185mediate boiling point. Consequently, special p,Lins were taken topurify the krypton. The gas obtained during the progress of thework already referred to had been repeatedly fractionated, and itsspect)rum showed none of the argon or xenon lines. I n order to besure that the krypton WAS free from these gases, i t mas refractionatedat the temperature of liquid air, according to the following plan :'SFractions 1 and 3 were reEractionat.ed separately, fractions 4 and7 being rejected as containing either argon or xenon; 2, 5, and 6were mixed and refractionated, 8 and 10 being rejected, whilstfraction 9, which contained most of the gas, was considered as beingpractically pure krypton.This gas was then fractionated a t - 130°, a bath of light petroleummixed with liquid air being used. Ten frar,tions were thus obtained.As the gas was more likely to be contaminated with traces of argonthan with traces of xenon, fractions 8 and 9 were selected for densitydeterminations.As the volume of No. 8 was not quite large enough,a small portion of No.7 was added. The gas samples were sparkedwith oxygen over sodium hydroxide solution, and the excess of oxygenwas removed by phosphorus.During the early stages of the sepuation of the mixed rare gases,i t was found that they were not only Contaminated with oxygen andnitrogeu, but also with traces of hydrocarbon vapour, derived fromthe pentme used in lubricating the compressors of the liquid-air plant.The mixed gases had therefore been passed twice heated over copperoxide, but on sparking the samples of krypton obtained as describedabove, i t wits found that some of t h e hydrocarbon vapour had escapedoxidation by the copper oxide. This, however, did not vitiate thevalue of the fractionation from argon and xenon. The last trace ofthe hydrocarbon was, of course, removed by the sparking.Two deter-minations were therefore made on fraction 8, with the followingresults :Fruction 8.Volume of density bulb (in c.c.) .........Temperature .............................. 0"Pressure on gas, corrected (in mm. ). .....Weight (in grams) .......................... 0.07566Density .......................................... 41 -50932.70574 i 4 *032.70770"478-30.0763341 5007 F 2186 MOORE: THE DENSITIES OF KRYPTOS AND XENON.As the gas was taken through the pump after the first determina-tion there was a possibility of its being contaminated with a verysmall traco of air. This would make the second result low, a n d41.509 may be accepted, therefore, as being the more correct figure.A determination on fraction 9 gave a low result, probably due toinsufficient sparking.A t this stage of the work, the author wasunfortunately forced to leave England, and Sir William Ramsay andMr. A. T. Cameron very kindly offered to make another densitydetermination with this fraction of the gas. After prolongedsparking and removal of oxygen, the density obtained mas 39.53.It was difficult to understand why fraction 9 should have a lowerdensity than fraction 8. They therefore thoroughly sparked fractions5 and 6, and mixed the resulting gas with No, 9. The whole wasthen fractionated once at liquid air temperature, all that could bereadily pumped off constituting fraction 1. The remainder of thegas was taken off in two fractions (3 and 3), the middle fractionbeing much the larger, A determination on this gave the folIowingresult :Volume of density bulb (in c.c.) ........... 32.7077Temperature ......................................0"Pressure on gas, corrected (in iiim.) ....... 765.9W e i d t (in grams) 0.11833 ..............................nenzi ty ............................................. 41'15They then decided to make one last and extremely thorough attemptto obtain a fraction with a density as high as 41.5. The sparkspectrum led them to suspect argon in the fir& and middle fractions,whilst No. 3 seemed to show traces of xenon. The gas was there-fore refractionated according to the followiug scheme. I n each casethe middle fraction was 19/20ths or more of the whole :A + K r -KrII II i r i SeIIIir + S eI1IA + Jir Krli r A + KrIIK r + X eII< rIIA + l i r_...... ___ . I I Il i r + X r A f Iir Kr- K Y + scCONCEPTION OF HYDROGEN IONS IN CATALYSIS, ETC. 2187Each time the first fraction was withdrawn through the pump.The second was allowed to run into a mercury reservoir under slightlyreduced pressure, whilst the gas which was left in the fractionatingbulb after equilibrium was established constituted fraction 3. Thefinal krypton fraction had not therefore been passed through thepump at any stage of the fractionation, which practically eliminatedcontamination with air.The results obtained were as follows :Volume of density bulb (in c.c.) ............ 32.7077'I'eiuperatnre .................................... 0'Weight ( i n gram>) ............................. 0'12329Ca1cul:Lted density ........................... 41'504They considered this figure correct to one part i n 2,000. Thedensity of krypton may be therefore tnkeii as the mean of 41.504 and41.509, namely, 41.506. On the assumption that krypton and xenonare monatomic gases, their atomic weights would therefore be 83,012and 130.70 respectively. These new figures do not throw them out ofplace in the periodic table.Prcssnre 011 gas, corrected (in iniii.) ........ 772.5I desire to thank Sir William Rnmsay and Mr. A. T. Cameron forthe independent work which they so kindly did, and which consti-tuted a rigorous confirmation of my own result ; also the former forhis many helpful suggestions.UNlVEllSITY COLLEGE,LONDON