664 



NA TURE 



[November iS, 1922 



Letters to the Editor. 



[T/ie Editor does not hold himself responsible for 

 opinio* ' by his correspondents. Neither 



can he undertake to return, or to correspond with 

 'iters of, rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken oj anonymous communications^ 



The Isotopes of Selenium and some 

 other Elements. 



The first experiments with selenium some time 

 ago were not successful. Very satisfactory mass- 

 spectra have now been obtained by vaporising the 

 element itself in the discharge tube. The interpre- 

 tation of these is quite simple and definite, so that 

 the results may be stated with every confidence. 

 Selenium consists of six isotopes, giving lines at 

 74 (/), 76 (c), 77 (c), 78 (6), 80 (a), 82 (d). The line 

 at ; 1 is extremely faint. The intensities of the 

 lines are in the order indicated by the letters, and 

 agree well enough with the chemical atomic weight 

 70.-2. Measurement of the lines shows no detectable 

 deviation from the whole number rule. 



Application of the method to cadmium and 

 tellurium has failed to give the mass lines of these 

 elements. The employment of the more volatile 

 TeCl 3 was also unsuccessful, but incidentally gave 

 evidence of great value, which practically confirms 

 two facts previously suspected, namely, that chlorine 

 has no isotope of mass 39, and that aluminium is a 

 simple element 27. 



During some work requiring very prolonged 

 exposures with a gas containing xenon, two new 

 isotopes of that element were discovered at 124, 

 126, making nine in all. The extreme faintness of 

 both lines indicates that the proportion of these 

 light isotopes in the element is minute. 



It will be noticed that the first of these is isobaric 

 with tin, and that the seleniums 78, 80, 82 are isobares 

 of krypton. All isobares so far discovered, including 

 the radioactive ones, have even atomic weights. 



F. W. Aston. 

 Cavendish Laboratory, Cambridge, November 6. 



Bohr's Model of the Hydrogen Molecules and 

 their Magnetic Susceptibility. 



Bohr's model of the molecules of hydrogen explains 

 very satisfactorily the light dispersion of hydrogen, 1 

 and gives the same value for the moment of inertia 

 as that deduced from the specific heat ; 2 but it is 

 generally believed that the model does not explain 

 the diamagnetic property of the gas. 3 For, according 

 to P. Langevin's theory, 4 the hydrogen molecules 

 must have paramagnetic susceptibility, while as a 

 matter of fact the gas is diamagnetic, as determined 

 by Dr. T. Sone. 6 But, as this note will show, this 

 conclusion is not correct. 



It is well known that, besides three degrees of 

 freedom for translation, hydrogen molecules possess 

 two degrees of rotational freedom. According to 

 Bohr's model, this rotational motion must, from 

 the point of view of symmetry, take place about an 

 axis perpendicular to the magnetic axis of the mole- 

 cules — that is, an axis perpendicular to the line joining 

 two positive nuclei. This rotational motion is uniform 

 and increases with the rise of temperature. Hence 



1 Debve, Miinchener Akademie (1915), 1. 



2 Reiche, Ann. der Phys., 58 (1919). 682. 

 ' T. Kun,', Phys. Rev., 12 (191S), 59. 



' P. Langevin, Arm. de Chan, el de Phys., 8 (1905), 70. 

 1 Si i. Rep. 8 (1919). "5. 



iJ 



the magnetic effect of each molecule due to the 

 revolving electrons vanishes on account of the rota- 

 tional motion. In this case, therefore, Langevin's 

 theorj of paramagnetism is not applicable. Obvi- 

 ously his theory can be applied only when the gas 

 molecules have no degree of rotational freedom, or 

 when they revolve only about their magnetic axes. 



If a strong field acts on a uniformly revolving 

 magnet in its plane of revolution (Fig. I), the rotation 

 begins to become slightly ac- 

 celerated in the half-revolution r'~ ~* \ 

 in the direction of the field 

 and retarded in the other half, 

 tins causing a diamagnetic \ 

 effect. 6 In the case of the v -*_^' 

 molecules of hydrogen the fig. i. 



moment of inertia about the 



axis of rotation is, however, very large on account of 

 the positive nuclei being apart from each other ; hence, 

 during rotation, these two revolving nuclei behave like 

 a large flywheel, and before the revolution of the mole- 

 cules is sensibly accelerated it is newly excited by 

 thermal impacts. Hence we may assume that this rota- 

 tion is not sensibly affected by the action of a strong 

 magnetic field, and therefore, in the mean, remains 

 uniform throughout. The hydrogen gas is then 

 diamagnetic, and its susceptibilitv can be calculated 

 by Langevin's theory of diamagnetism. 7 



Assuming Bohr's new model of the hydrogen mole- 

 cules (in which the electrons have elliptic orbits), 

 we have for the major axis of the orbit 



I' 1 

 a= -(« - n 1-, 



k= - I-e-.- " 2 



V3' (n + //')'-' 



where // is Planck's universal constant, in the mass 

 of the electrons, and e their charge; e is the eccen- 

 tricity of the orbit, n and n' are the azimuth and 

 radial quantum numbers. 



In the case of n + n'=i, the possible orbit is «=i, 

 n' = o, which reduces to a circle, the radius of which is 



a = 0-507 x IO~ 8 , 



— ~ — „ being 0-532 x 10- 8 . 



The magnetic susceptibility of the gas per gram-mole- 

 cule is given by 



iiiii/ 

 12 \m 



)\ 



where n is the total number of electrons and - is 

 to be taken for different orbits. Applying this 

 formula to the above case, we have 



■ 2, n=n'=i corresponds 

 = 3/4, and the equivalent 



In the case where n + n' 

 to the elliptic orbit. Here < 

 radius of the circle is 



a= 1-433 x io -8 cm., 



x= -5-7°* IO ~ 6 - 

 The diamagnetic susceptibility x = 3'96xio- 6 ob- 

 served by Dr. T. Sone lies between these two. In 

 actual cases a certain fraction of the whole number 

 of molecules may have the first orbit (»=i, n'=o), 

 and the other fraction the second orbit («=«'= 1), 

 etc. As the orbit becomes greater there is a greater 

 chance that it will collapse into a smaller orbit; 



1 K. Honda and J. Okubo. Sci. Rep. 5 (1916), 325. 

 ' P. Langevin, I.e. 



NO. 2768, VOL. I IO] 



