20 



NA TURE 



[May 7, 1908 



Armet de Lisle at Nogent-sur-Marne. The funds were de- 

 frayed from a grant made by the Goldsmiths' Company to 

 the Royal Society, in 1904, for the purpose of the investi- 

 gation of radium. 



The process of extraction produced about 413 grams of 

 anh^'drous barium chloride, containing radium chloride 

 sufhcicnt to give the salt a radio-activity 560 times that 

 of uranium. 



This salt was received by the Royal Society in the 

 autumn of 1906, and was handed to me in January. 1907, 

 with the request that I would extract the radium chloride 

 from it, and undertake, if possible, a re-determination of 

 the atomic weight of the element. 



The method of extraction was substantially that adopted 

 by Mme. Curie, namely, systematic fractional crystallisa- 

 tion, first from water and then from increasingly strong 

 hydrochloric acid, until finally the acid used was the 

 strongest that could be obtained by distillation. 



Whilst still engaged in the isolation of the radium 

 chloride from the material furnished by M. Armet de Lisle, 

 I received a further small supply of radium from the Royal 

 Society. It was bought in Cambridge, and purported to 

 be radium bromide, but on removing it from the capsule, 

 in which it had been stored since 1903, it was found to 

 be wholly insoluble in water. On treatment with pure 

 dilute hydrobromic acid it readily passed into solution. 

 The salt obtained by evaporation was sent to Prof. Ruther- 

 ford, who had kindly undertaken to make any measure- 

 ments of radio-activity which I needed. He estimated the 

 amount of radium present as equivalent to 33 milligrams 

 of radium bromide. This salt was eventually converted 

 into chloride, and was purified by repeated crystallisation 

 from strong hydrochloric acid. 



Determination of Atomic Weight. — This was effected by 

 ascertaining the amount of silver chloride yielded by a 

 weighed quantity of the anhydrous radium chloride — the 

 principle of the method already employed by Mme. Curie. 



A method was devised whereby the whole of the opera- 

 tions of drying and weighing the radium chloride, pre- 

 cipitating, washing, drying, and w'eighing the silver 

 chloride, might be performed in one and the same vessel, 

 thus obviating the necessity of transferring the silver salt 

 and of separating it by any of the ordinary processes of 

 filtration. 



The vessel in which these operations were made consisted 

 of a thin glass tube with a conical base furnished with a 

 hollow, well-ground stopper. It had a capacity of about 

 15 c.c, and was as light as was consistent with the 

 requisite strength, and could be suspended from the balance- 

 arm by fine platinum wire. In all the weighings a ore- 

 cisely similar bottle of almost identical weight and 

 capacity, suspended in like manner, was employed as a 

 tare. The weighings were made on a very sensitive assav 

 balance, with 4-inch arms, carrying a maximum load of 

 12 grams, and provided with light stirrup pans. 



The washed silver chloride was first dried at 100°, and 

 then heated in the air bath to 160° for about a couple of 

 hours, and, after standing in the desiccator over phos- 

 phoric oxide for about eighteen hours, was weighed in the 

 manner described. 



In order to test the practicability of the method and to 

 acquire experience of its working, as well as to gain some 

 idea of its accuracy before actually making use of it in 

 the case of the radium salt, a series of determinations of 

 atomic weight of barium was made with purified barium 

 chloride. 



The results were as follows : — 



Ag= 107-93. CI =35-45. 



Atomic weight, Ba 



137-5 



I36'9 

 137-6 

 137-5 



The value for barium adopted by the International Com- 

 mittee on Atomic Weights, 1907-^, is 137-4- 



It will be seen from these numbers that a close approxi- 



NO. 2010, VOL. 78] 



mation to the true atomic weight of barium can be 

 obtained by the method described, the maximum error 

 being about half a unit, or less than 0-5 per cent. Con- 

 sidering that the atomic weight of radium is probably 

 nearly double that of barium, the same fortuitous errors j 

 would affect its value to about a unit. 



As the work of isolating and purifying the radium 

 chloride proceeded, determinations of the amount of 

 chlorine were made as described from time to time, and , 

 as soon as approximately constant values were obtained . 

 it was assumed that any barium or other impurity present ■ 

 was too small in amount to affect the results when regard 

 was had to the unavoidable experimental errors. The 

 resulting chloride was then repeatedly and carefully re- 

 crystallised from pure, strong hydrochloric acid, the 

 *' tails," which were comparatively rich in radium, being 

 specially set apart. 



The purified salt finally extracted from the material 

 supplied by .M. .-\rmet de Lisle weighed, when anhydrous, 

 64 milligrams. 



I regard this salt as substantially radium chloride. I 

 am not, however, in a f>osition to say that it was abso- 

 lutely free from barium. .-Xt the same time, I have reason 

 to believe that the amount still present was probably too 

 small materially to influence the result, considering the 

 limited quantity of the salt I had to work with, and the 

 consequent relatively large experimental errors. 



With the aid of Sir William Hoggins, who kindly made 

 the spectroscopic trials for me, I was able to carry out 

 Mme. Curie's test of comparing the relative intensity of 

 the lines of barium and radium in the spark spectrum of 

 the separated r.adium chloride. .Mme. Curie compared the 

 relative strengths of lines 4554-2 of Ba and 4533-3 of Rd. 

 Although these have the advantage of being close together, 

 thev are of dissimilar intensity. Sir William Hoggins 

 advised that a more stringent test would be to take the 

 line 5536-2 of Ba of intensity 10, and compare it with the 

 Rd lines 5813-8 and 5560-8, which are also of intensity 10. 

 On actually making the trials, which were repeated several 

 times, the green Ba line 5536-2, although visible, was seen 

 to be relatively verv feeble — less intense, indeed, than that 

 afforded by the most dilute solution of barium chloride 

 we were able to employ. 



With this material, therefore, I attempted to make the 

 determination of atomic weight. Accordingly, the greater 

 portion was transferred to the vessel already described, and 

 the amount of chlorine in the anhydrous salt determined 

 with all possible care. The result was : — 



Radium chloride, 

 milligrams 



627 



Silver chloride, 

 milligrams 



60 '4 



Atomic weight, Rd 

 2268 



The radium was recovered from the solution, re-converted 

 into chloride, added to what remained of the original 

 quantity, and the amount of chlorine again determined in 

 the anhydrous salt. The second result was : — 



Radium chloride, 

 milligrams 



63-9 



Silver chloride, 

 niiltigr -ms 



618 



Atomic weight, Rd 



225-7 



The purified chloride obtained from the Cambridge 

 material amounted to 24 milligrams. 



It w^as added to the main bulk, and the whole was re- 

 peatedly crystallised from strong hydrochloric acid, about 

 6 milligrams being thus removed in the mother liquors. 

 .'\fter being dried at 150° it was again analysed, with the 

 following results : — 



Radium chloride, 

 milligrams 



7S-4 



Silver chloride, 

 milligrams 



75-3 



Atomic weight, Rd 



2277 



The mean value is 226-7, <"■. ^° '^"^ nearest unit, 227. 

 This, it will be observed, is in very close accord with 

 Mme. Curie's latest number. 



I think, therefore, it is reasonably well established that 

 the atomic weight of radium is now known to within a 

 unit which, considering the relatively high number, is, in 

 the present circumstances, as fair a degree of exactitude 

 as could be anticipated. 



