EXCHANGE ZIbe IHntversitp of dbtcago nr FORMATION AND LIFE OF METASTABLE HELIUM A DISSERTATION SUBMITTED TO THE FACULTY OF THE OGDEN GRADUATE SCHOOL OF SCIENCE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF PHYSICS BY FABIAN MILLER KANNENSTINE Private Edition, Distributed By THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS Reprinted from THE ASTROPHYSICAL JOURNAL, Vol. LV, No. 4, May 1922 ZTbe Ulniverstts of Cbtcago FORMATION AND LIFE OF METASTABLE HELIUM A DISSERTATION SUBMITTED TO THE FACULTY OF THE OGDEN GRADUATE SCHOOL OF SCIENCE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF PHYSICS BY FABIAN MILLER KANNENSTINE Private Edition, Distributed By THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS Reprinted from THE ASTROPHYSICAL JOURNAL, Vol. LV, No. 4, May 1922 AlKfiO"; EXCHANGE FORMATION AND LIFE OF METASTABLE HELIUM BY FABIAN M. KANNENSTINE ABSTRACT Current- potential curves for alternating and intermittent arcs in pure helium between a Wehnelt cathode and nickel anode were obtained by use of a Braun-tube oscillograph, for pressures of from 0.06 to 2 mm. With alternating voltages, the striking potential for each half-cycle was 29 volts or more for frequencies of 200 or less, but for frequencies of 220 cycles and above, the current started at about 5 volts and soon reached a saturation value until about 25 volts caused a further increase. In all cases the breaking potential was about 5 volts. To form the intermittent arc, the voltage was periodically dropped from 36 to some lower predetermined value. If this lower potential was less than 4.6 volts, the arc was suddenly extinguished; if between 4.6 and 23.7 volts, the arc died out gradually; and if it was 25.3 volts or above, the arc was maintained. All the foregoing values include a correction of i.i volts added to the actual readings to bring them in agreement with accepted results. Evidence as to the life of metastable helium. To explain various experimental results, it has been suggested that atoms of helium, partially ionized by electronic collisions of 20.8 volts or more and hence capable of complete ionization by further impacts of 4.8 volts or more, remain in this metastable condition for an appreciable time. The foregoing results not only confirm this suggestion but also show that under the circumstances of this experiment the metastable form persists for about 0.0024 second. INTRODUCTION Experiments with electron impacts in helium have led to a value for the minimum ionizing potential of 25.3 volts and to two radiating potentials, one at 20.5 volts 1 and another at 21.3 volts. 2 Franck and Reiche 3 and Franck and Knipping 2 conclude that the partially ionized state formed when the atom has been excited by the impact of 20.5 volt electrons is a "metastable" state, which may persist for some time and may be even permanently stable in abso- lutely pure helium. This conclusion is drawn from the fact that the radiation usually observed with electron impacts at this voltage disappeared when the helium was very pure, indicating that the displaced electrons do not return to the normal state. In slightly impure helium they also observed that the radiation, as measured by 1 F. Horton and A. C. Davies, Philosophical Magazine, 42, 746, 1921; J. Franck and P. Knipping, Physikalische Zeitschrift, 20, 481, 1919; K. T. Compton, Philosophical Magazine, 40, 553, 1920; F. S. Goucher, Proceedings Physical Society of London, 33, 13, 1920. 2 J. Franck and P. Knipping, loc. cit. Zeitschrift fur Physik, i, 320, 1920. F. Horton and A. C. Davies, loc. cit. 3 Zeitschrift fur Physik, i, 54, 1920. 345 346 FABIAN M. KANNENSTINE its photo-electric effect, appeared to continue for a brief time after the exciting electron impacts ceased. Experiments by Paschen, 1 which show that in pure excited helium resonance radiation may be caused by the infra-red line 10830 A, also support the conception of a metastable state. All the energy of this wave-length absorbed by the excited atoms was found to be re-emitted as radiation of this same wave-length. In terms of Bohr's model this means that the radiation is absorbed by atoms which have already absorbed the energy corresponding to the radiating potential 20.5 volts, and that the displaced electrons return to this partially ionized state only, and not to the normal state. The semi-stable existence of partially ionized helium atoms is also useful in explaining the phenomena in low-voltage arcs in helium, especially the striking of the arc at 20.5 volts with dense electron currents. 2 The phenomena of successive impacts as dis- cussed by Compton 3 are of much more importance when the par- tially ionized atoms have a relatively long life. The metastable helium should be completely ionized by electrons with a speed corresponding to a potential of 25.3 20.5 =4.8 volts. If we assume that the metastable state may be produced by the arc radiations, 4 then an arc might be maintained with dense enough electron currents down to 4.8 volts. It is of interest that in this way Comp- ton, Olmstead, and Lillie have maintained an arc with potentials as low as 8 volts between the electrodes. Horton and Davies 5 report an unsuccessful attempt to observe the ionization of abnormal helium atoms by bombardment with electrons of 4.8 volts. In this paper experiments with helium arcs are described which show that helium may exist temporarily in a metastable form, and that arcs may be maintained in it with 3.5 volts. An approximate determination of its life is also reported. 1 Annalen der Physik, 45, 625, 1914. 2 K. T. Compton, E. G. Lillie, P. S. Olmstead, Physical Review, 16, 282, 1920. 3 Physical Review, 15, 476, 1920. O. W. Richardson and C. B. Bazzoni, Nature 98, 5, 1916; F. Horton and A. C. Davies, op. cit., 42, 752, 1921. 5 Ibid., 42, 763, 1921. METASTABLE HELIUM 347 ARCS WITH LOW-FREQUENCY VOLTAGES The arcs were formed between a Wehnelt cathode and nickel anode in a pyrex tube containing carefully purified helium. The helium was purified by passing over heated copper oxide and charcoal immersed in liquid air. Before passing into the experi- mental tube the helium passed through two additional charcoal tubes and a liquid-air trap. The tube and liquid-air trap were baked out at 450 C. for several days before the helium was admitted. In the first experiments the helium was excited by an alternat- ing e.m.f. An electromagnetic oscillograph was used to obtain curves showing the variation of current and voltage with time. From such curves the current-voltage curve may be obtained. A decided dissymmetry in the growth and decay of the arc current could be observed, and also an apparent continuation of the arc current down to very low voltages. To obtain greater sensitiveness and to avoid the possible distortion from the inertia of the oscillograph elements, a sensitive Braun tube oscillograph was used in most of the later experiments. The electrostatic deflecting plates of the Braun tube were connected across the anode and cathode of the helium tube, while the arc current passed through the electromagnetic deflecting coils. The electrostatic and electro- magnetic deflections were at right angles, so that the figure on the screen of the Braun tube was a graph in rectangular co-ordinates of the impressed e.m.f. and the arc current. The deflecting coils were so adjusted that the thermionic current alone did not produce any observable deflection. Figure i shows the kind of curve obtained with a 6o-cycle e.m.f. in helium at a pressure of 2 mm. It is seen that at a certain voltage the arc strikes, and as the voltage increases the current increases also. When the voltage decreases the current decreases at the same time, but continues with considerable magnitude until the voltage reaches a value of about 4 volts, when the current ceases to flow. Careful observation showed a sudden drop to zero at this lower voltage. This continuation of the arc to very low potentials may be attributed to the persistence of a metastable form of helium in which an arc may be maintained down to about 4 volts. Two other breaks in the down curve may be observed 348 FABIAN M. KANNENSTINE just below the striking potential of the helium itself. These breaks will be referred to later. This type of curve was observed with 60 cycles at pressures down to 0.06 mm. Down to i mm pressure the striking potential was about 28 volts, and at the dying-out potential of about 4 volts a very decided drop of the arc current to zero was observed. As the pressure was decreased below i mm the striking potential t* FIG. i FIG. 2 increased and the dying-out potential became less certain, i.e., the drop to zero current was not so sudden as at higher pressures. At 0.06 mm the striking potential had increased to about 40 volts and the dying-out potential to about 20 volts. Figure 2 shows the type of curve obtained at 0.06 mm pressure. At all these pressures the filament temperature was approximately constant. The mean arc current varied from 25 milliamperes at 2 mm pressure to 10 milliamperes at 0.06 mm pressure. If any impurities are present one would expect their effect to be more marked at lower pressures. Franck and Knipping have assumed that me ta stable helium will not return to normal helium except in the presence of impurities, so that with larger amounts of impurities we might expect the metastable helium to be destroyed in a shorter time. This is probably the interpretation of the gradual dying out of the arc before the voltage had dropped to 4 volts, as was observed when the pressure was very low. Experiments were made directly with the method discussed in the next section to see if the life was affected by impurities obtained by removing the liquid air from the charcoal for a short time till the helium had become very impure. The experiments indicated MET AST ABLE HELIUM 349 that while the impurities suppressed the amount of the metas table helium, the part remaining had the same life as that formed in the pure helium. The increase in the striking potential as the pressure was reduced is more difficult to interpret and is probably connected with the mechanism of the arc formation. ARCS WITH HIGH-FREQUENCY VOLTAGES With a sufficiently high frequency the e.m.f. should pass from +4 volts to its negative maximum and back to +4 volts before the metastable helium has disappeared, and upon reaching 4 volts the arc should strike, due to the ionization of the metastable helium still present. Accordingly various frequencies from 60 to 10,000 cycles were impressed on the arc. An audion oscillator was used as a source for these frequencies. The author wishes to acknowledge his indebtedness to Dr. J. P. Minton and Mr. V. O. Knudsen for the use of their oscillators in this work. The alternating e.m.f. impressed on the arc was obtained from the secondary of an iron core transformer having its primary winding in the plate circuit of the oscillator. Figure 3 shows the type of figure observed above 220 cycles per second at 2 mm pressure. The arc struck at about 4 volts, and the current rapidly attained a saturation value at which the arc current continued until the voltage had reached about 25 volts, when a second sudden increase was obtained. The decreasing part of the curve was above the increasing part but fell to zero at the lower striking voltage. With 200 cycles no appreciable FlG 3 arc formation at the lower voltage could be observed, indicating that the metastable helium practi- cally disappeared in less than approximately T fo of a second. The striking of the arc at about 4 volts with 220 cycles indicates that enough of the metastable helium was left after approximately of a second to give an observable indication of ionization. e. 350 FABIAN M. KANNENSTINE All the figures above 220 cycles were similar, except that as the frequency was increased the striking at about 4 volts became more and more marked. In using the oscillator as a source of alternating e.m.f. smooth curves were not always obtained. As the mean arc current was increased, by raising the electron emission, certain parts of the curve became wavy. This was particularly noticeable on the striking of the arc at about 25 volts. Instead of the arc current increasing as in Figure i, it increased in a wavy form, indicating that the impressed voltage rose and fell during this striking period. This is provisionally attributed to self-induction in parts of the circuit or to the presence of iron in the oscillating circuit. ARCS WITH INTERMITTENT CONSTANT VOLTAGES In another series of experiments intermittent constant voltages from a battery were impressed between the hot cathode, C, and the anode, A (see Fig. 4). When the mercury interrupter, H, was open the voltage impressed on the electrodes was about 35 volts, and when the in- terrupter shorted the section BD, the voltage dropped to a value dependent upon the position of the slider B. In this way it was possible to impress a voltage of about 35 volts and suddenly change this voltage to a lower value. The arc current was passed through the deflecting coils, M, of the Braun tube. The spot was examined directly and also with a rotating mirror. Since it was rather difficult to observe the spot in the rotating mirror the following arrangement was also used. A uniform time deflection was obtained by connecting the deflecting plates to a potentiometer with a rotating contact, which was mounted on the shaft of the mercury interrupter. The electro- static deflection was therefore proportional to the time and the FIG. 4 METASTABLE HELIUM 351 magnetic deflection proportional to the current through the arc. These two deflections were at right angles. Furthermore, since the rotating contact was mounted on the interrupter shaft, the figure produced was stationary and could be carefully examined as the slider B was moved. With the lower voltage below 3.5 volts and without the time deflection two spots appeared on the Braun tube screen, the upper one corresponding to the current for 35-volt excitation and the lower one for zero current. Using the time deflection, it was seen that the arc was extinguished in a time too short to be so observed, as shown at the right of Figure 5. Only a very faint vertical line appeared, since the spot moved very rapidly from the upper to the lower positions. As the lower voltage was increased to 3.5 volts the lower line suddenly became curved as shown in Figure 6, indicating a gradual u u FIG. 5 FIG. 6 decrease of the current to zero value. This sudden curving of the lower line at 3.5 volts may be attributed to the presence of metas- table helium produced during the 35-volt excitation. That is, the voltage dropped from 35 volts to 3.5 volts in a time short compared with the life of the metastable helium, and the arc was maintained by 3.5 volts as long as the metastable helium was present. The dying out of the arc continued to have the same appearance until the lower voltage had been increased to 22.6 volts, when a third spot appeared slightly above that for zero current as in Fig- ure 7. Using the time deflection, it was seen that the arc was maintained for a brief time with this voltage before dying out. A slight further increase of the lower voltage to 23.6 volts brought out still another spot, as in Figure 8, and the time deflec- tion showed a dying out of the current in two stages as shown. The voltages for these two breaks in the decrease correspond to 352 FABIAN M. KANNENSTINE the voltages for the two breaks observed in the down curve of the alternating e.m.f. figure already described. This decrease in two stages may be connected with the fact of the existence of two radiat- ing potentials of helium 0.8 volts apart corresponding to the first coplanor orbit (helium) and the first crossed orbit (parhelium) on Franck and Knipping's interpretation. With a slightly increased lower voltage of 24.2 volts the arc was maintained. The two lower spots of Figure 8 disappeared, and the arc current jumped between the two values indicated by the two upper spots. FIG. 7 FIG. 8 These phenomena were qualitatively the same at pressures ranging from 2 mm down to 0.06 mm. The above exact determina- tions were all made visually with the Braun tube oscillograph, but for the purpose of illustration photographs made with the electro- magnetic oscillograph show qualitatively the same phenomena, although with less sharpness in the details. Plate VI shows the various changes which occur as the lower voltage is increased from zero to 24.2 volts. These oscillograms were obtained by passing the arc current through an element of the electromagnetic oscillo- graph and photographing the curve while the mercury interrupter was operating. In a the lower voltage was below 3.5 volts, and it is seen that the arc current drops to zero in a very short time. The curving of the lower current line when the lower voltage reaches 3.5 volts is shown in b. The brief maintenance of the arc at a lower voltage of 22.6 volts is shown in c, and the decrease in two stages at a lower voltage of 23.6 is shown in d. The maintenance of the arc at 24.2 volts is shown in e. A number of readings were taken of these voltages, and then the cathode current was reversed in order to eliminate the drop in voltage along the cathode, and another set of readings taken. The PLATE VI . I UNIVERSITY OF CALIFORNIA LIBRARY, BERKELEY THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW Books not returned on time are subject to a fine of 50c per volume after the third day overdue, increasing to $1.00 per volume after the sixth day. Books not in demand may be renewed if application is made before expiration of loan period. Y APR 5 1932 1999 25m 7, '25 Makers Syracuse, N. Y. PAT, JAN 21, 1908 UNIVERSITY OF CALIFORNIA LIBRARY