1014 Enhanced Spectrum of Magnesium in Low Voltage Arc. 



The enhanced spectrum of triply ionized magnesium — or, 

 more explicitly, a doubly ionized atom with a third electron 

 displaced to various outer orbits, so that we have to deal 

 with a core of net charge -f Ze — might be expected upon the 

 production of L -radiation, if the current density is sufficiently 

 high to keep the two valence electrons completely ejected. 

 The spectrum should be extremely complicated, resembling 

 that of Neon. We were unable, however, to detect any 

 sharp transition in the spectrum between '20 and 60 volts. 



This type of arc is probably less subject to the presence 

 of impurities than most forms of spectroscopic sources. 

 Any metal occurring as an impurity would have to have 

 about the same vapour-pressure temperature-relation as the 

 metal employed. Otherwise it would be either not vaporized 

 or so rapidly distilled and condensed that no appreciable 

 concentration of the vapour would be present after continued 

 operation *. The magnesium employed was not of high 

 purity, yet the only lines of other elements present were 

 zinc and hydrogen. The 30-volt arc shows the first 8 terms 

 of the diffuse triplet series and the first 4 terms of the sharp 

 triplets of zinc. At 9 volts (below the ionization potential 

 of zinc, 9'35 volts) but two lines appear, \3076 and A, 2139. 

 This is to be expected from the behaviour of magnesium, as 

 shown in the second spectrogram of PL XXVII. These two 

 lines constitute the two-line spectrum of zinc, 1*5 S — 2p 2 and 

 1-5 S — 2 P, observed by McLennan f, the first to recognize 

 the importance of the study of low-voltage arcs. At still 

 lower voltages the line A, 3076 appears alone as an impurity, 

 i. e. the single-line spectrum of zinc. 



The presence of hydrogen as an impurity is of especial 

 interest. The magnesium was distilled with a gas-pressure 

 of the order of only O'OOOl mm. Hg throughout the experi- 

 ments. Yet ten terms of the Balmer series X 4861 to A, 3721 

 showed on a 30-volt exposure. The fact that no trace of 

 the secondary spectrum was observed indicates that the 

 hydrogen was present in the active or atomic state. Possibly 

 it may be accounted for by the dissociation of magnesium 

 hydride occurring as an impurity. 



It will be noted from Table II. that the lines of the arc- 

 spectrum 2p l — bd and 1*5 S — 2 P, although representing 

 entirely different types of interorbital transitions, in wave- 

 length are separated by only 0*46 A, about the separation of 



* The filament should be welded to tungsten or molybdenum leads. 

 If hard soldered, lines of impurities may arise from the vaporization of 

 the solder. 



t McLennan and Ireton, Phil. Mag. xxxvi. p. 461 (1918). 



