290 



SCIENCE 



[N. S. Vol. LI. No. 1316 



nesium, and 28.0 for silicon. Among the 

 heavier elements there are probably few ele- 

 ments which are not mixtures of isotopes. 

 Thus there is excellent evidence in the atomic 

 weights that the following elements are mix- 

 tures: nickel, copper, zinc, and practically all 

 of the other elements form atomic number 28 

 to atomic number 80. The radio-elements 

 from thallium (No. 81) to uranium (92) 

 were at that time known to be mixtures of 

 isotopes. 



Experiments on the diffusion of chlorine 

 gas were begun by Dr. W. D. Turner and my- 

 self in 1916, and early in 1917 slight differ- 

 ences in density were detected, but the 

 chlorine was not entirely pure. Chlorine was 

 used because it could be obtained in cylinders, 

 and its flow was therefore very easy to con- 

 trol. However if chlorine consists of two 

 isotopes, chlorine (01) and meta-chlorine 

 (CZ), there are three forms of molecular 

 chlorine: Cl-Cl, G\-Cl, and Cl-Cl, and this is 

 unfavorable to the diffusion. For this reason 

 we have carried out most of our work by the 

 use of hydrogen chloride gas, which, while 

 unfortunately not obtainable in this country 

 compressed in cylinders, at least has the ad- 

 vantage that its molecules contain only one 

 atom of chlorine each, and that the hydrogen 

 of the molecule has little effect in increasing 

 the molecular weight. 



This work was interrupted by the war, but 

 by the summer of 1919 about ten thousand 

 liters of gas had been diffused, a part of this 

 diffusion being done by T. H. Liggett. In 

 October, 1919, I interested Mr. 0. E. Broeker 

 in this problem. He has diffused about eight 

 thousand liters of this gas and we hope soon 

 to have an enlarged apparatus capable of 

 diffusing a thousand liters per day, in the 

 first section. At present we have five large 

 units in operation or ready for operation. 



The separation by diffusion of gases whose 

 molecular weights lie close together is, accord- 

 ing to the diffusion theory of Lord Eayleigh, 

 an extremely slow process. Up to the present 

 time we have concentrated our efforts upon 

 the diffusion itself, and have spent little time 

 in analytical work. The preliminary analyses 



indicate that the density of the fraction 

 which remains inside the diffusion tubes, is 

 increasing at about the rate predicted by the 

 Eayleigh theory of diffusion, if we consider 

 35.0 as the atomic weight of chlorine, and 

 37.0 as the atomic weight of meta-chlorine. 

 We have tested for most of the impurities 

 which might be present except arsenic tri- 

 chloride. The hydrogen chloride gas is gener- 

 ated from 0. P. hydrochloric acid by the 

 action of 0. P. sulphuric acid. The next step 

 in our work of proving that a separation has 

 been effected is to secure larger quantities of 

 diffused material, since ovir final fractions 

 are still small, so that we may be more 

 certain of our purification of the material, 

 and then to make precise atomic weight deter- 

 minations. If on such further purification 

 we obtain an atomic weight for the heavy 

 fraction as high as that already obtained in 

 our preliminary analyses, we will have definite 

 evidence that we have separated chlorine into 

 a heavier and a lighter isotope. This will be 

 of importance in two ways, first, it will be the 

 first experimental separation of an element 

 into parts, and second, it will be one of the 

 strongest links in the proof that the nucleus 

 of the hydrogen atom is actually the positive 

 electron. 



As stated above, I gave in a series of papers 

 published five years ago^ a system of atomic 

 structure which gave very strong indications 

 that chlorine, magnesium, silicon, and the 

 other elements specified, are mixtures of 

 isotopes. In fact this system of structure, 

 for which there was much evidence, depended 

 for its validity upon the existence of these 

 isotopes, and in 1916 I published a prelim- 

 inary notice^ stating that we were working, 

 on the separation of chlorine into isotopes. 

 It is of great interest that Aston in a prelim- 

 inary note written to Nature in December, 

 1919, states that his results obtained by posi- 

 tive rays indicate that both chlorine and mer- 

 cury are mixtures of isotopes, with atomic 



1 Journal of the American Chemical Society, 37, 

 1367-96, especially pages 1390, 1391, 1387. 



2 Ibid., 38, p. 19. 



