THE EFFECTS OF PRESSURE OK THE AFTER-GLOW OF NITROGEN. by Henry Hubert Clayton A Thesis submitted, for the Degree of Master of Arts In the Department of Physics. THE UNIVERSITY OF BRISISH COLUMBIA September - 1937. TABLE OE CONTENTS. I. PURPOSE. Page 1 . I I . DISCOVERY OE THE AFTER-GLOW. Page 1. I I I . PREVIOUS EXPERIMENTAL RESEARCHES. A. The Spectrum of the After-glow. Page 1 , B. Spectra excited by Active Nitrogen. Page 3. C. Effects of Temperature^ on the After-glow. Page 4. D. Effects of Pressure on the After-glow. Page 5. E. Effects of an E l e c t r i c F i e l d on the After-glow. Page 6. F. Rate of Decay of the After-glow. Page 7. IV. THE NATURE OF ACTIVE NITROGEN. A. Metastable Molecules. Page 8. B. Metastable Atoms. Page 9. C. Dissociation. Page 10. V. THE PROCESSES OF EXCITATION AND DECAY OF THE AFTER-GLOW. A. Excitation. Page 12. B. Decay (1) The sharp selection of the bands and deactivation of the meta-stable atoms. Page 14. (2) Deactivation of the metastable molecules. . Page 15. VI. DEPENDENCE OF INTENSITY AND RATE OF DECAY ON CONCENTRATION. Page 17. TABLE OF CONTENTS, (Contd.) VII. APPARATUS. A. Source of Nitrogen. Page 24. B. The Discharge Tube. Page 24. C. The Compression Tube. Page 25. D. E l e c t r i c a l Arrangements. Page 26, BIBLIOGRAPHY. ILLUSTRATIONS. Figure 1. Energy l e v e l diagram of Nr>. Facing Page 2. Figure 2. Discharge and Compression Tubes. Facing Page 24. THE EFFECTS OF PRESSURE ON THE AFTER-GLOW OF NITROGEN, I. PURPOSE. The purpose of the work described i n this thesis was to Investigate the effect of pressure on the luminosity of the afterglow of nitrogen. I t was hoped thereby to c l a r i f y the processes taking place when the afterglow i s emitted, and i n particular, to determine whether they involve c o l l i s i o n s of two, or more, bodies. I I . DISCOVERY OF THE AFTER-GLOW. The modification of nitrogen known as "active n i t -(1) rogen" was discovered by E. P. Lewis i n 1899. Lewis found that the passage of a condensed discharge through nitrogen i s followed by the emission of a yellow afterglow, which may persist for some minutes after the termination of the exciting discharge. The glowing gas was found to have a considerable chemical a c t i v i t y , whence- the name "active nitrogen". This discovery iniated a long series of experimental researches, In the course of which the glowing gas was found to have many important properties. A brief outline of these researches w i l l be given I I I . PREVIOUS EXPERIMENTAL RESEARCHES. A. The Spectrum of the Afterglow. To face Page 2 C £ « - I /a /6 J A/, /2 6 -4^ 4 A 4 hern cTS /s/V i3 ' 2 C C*7T V \-h Energy /eve/ aft a^r am of A/a 2 « The spectrum of the afterglow, which i s entirely ( 2 ) different from that of the exciting discharge, consists of nearly a l l the "bands of the F i r s t Positive Group of nitrogen. These bands are ascribed to the transition B TT—»• 3 A 22 (see figure 1.) of the neutral nitrogen molecule. The dist r i b u t i o n of intensity among the bands d i f f e r s markedly from that found i n ordinary discharge, certain bands being selectively i n t e n s i f i e d . These are, at ordinary temperatures, ( 3 ) the bands originating on the vibrational levels IS to 10 of the electronic state B and terminating on the levels 9 to 7 of the state A , and also the bands B g - ^ A ^ and B y — * A 4 . In the f i r s t of these groups the bands B ^ — * A g and B-Q—*-A7 are most intense. At lower temperatures the intensity maximum i s shifted (4) towards the v i o l e t , and the selection i s narrower, the bands ( 3 ) starting on B 1 2 and B & being most intense. These bands are the ox-bands of the afterglow. Three other systems of bands, the yfland > groups, discovered by l e w i s ^ and the group by Xhauss^, were (2) shown by Strutt to belong to the spectrum of NO, and to be due to traces of oxygen, (7) In addition to the above bands Bay and Steiner v ' found lines of N-I and N-II i n the afterglow, Kichlu and to) Achaiya investigated the infra-red spectrum of the after-glow. The found an extension of the F i r s t Positive system, but no trace of atomic lines in this region, The absorption spectrum of the afterglow has been to) investigated by Sponer and by Ruark, Foote, Rudnick and 3» Chenault^ 1 0^. Sponer f a i l e d to obtain any trace of the transition A3Z>-»X1'Z> in that region of the u l t r a - v i o l e t i n which i t was to be expected. She also examined the u l t r a -v i o l e t spectrum of nitrogen i n emission, but f a i l e d to find bands corresponding to t h i s t r a n s i t i o n . Ruarx, Foote et a l investigated the region between 6,500A and 3,200A and found no absorption by the afterglow. B. Spectra excited by Active Nitrogen. I t was early found that active nitrogen i s able to excite other gases or vapours with which i t comes i n contact to the emission of their characteristic spectra. This excitation i s often accompanied by chemical reaction. Thep, yand£"bands mentioned above, which belong to the spectrum of NO and are due to the presence of oxygen, are an example of t h i s . Other instances are the excitation of metallic l i n e spectra, which i s usually accompanied by the formation of n i t r i d e s . Strutt and Fowler examined the spectra of several metals and compounds. They found the spectra of Na, K, Hg, Th to approximate to the arc spectra, while that of Mg contained no arc l i n e s , but was nearly the flame spectrum. The admixture of CuCl produced many lines of copper, besides the bands of the chloride. Similarly admixture of stannous and stannic chlorides produced lin e s of t i n . (12) Mullllcen examined the spectra of the copper halides when excited by active nitrogen. He found, confirming Strutt and Fowler, the entire arc spectrum of copper produced. The 4 . ionisation potential of copper i s 7.7 vo l t s . The bands of Cul excited were those corresponding to electronic levels with energy of from 2 . 4 4 to 2.96 vo l t s . Foote, Ruark and Chenault ' found that lines of mercury requiring up to 9.52 volts for their excitation were strongly developed, while exposures of 150 hours f a i l e d to. . (6) bring out line s requiring 9.66 volts or higher. Knauss experimented with hydrogen, oxygen, n i t r i c oxide and carbon monoxide. The f i r s t electronic l e v e l of the hydrogen molecule has energy of 11.1 v o l t s . No bands due to hydrogen molecules were obtained. In the case of carbon monoxide, bands whose i n i t i a l levels have energyfrom 8.2 to 9.0 volts were photographed. Those requiring higher excitation were (14) not observed. Kaplan x "*' found that the auroral green l i n e of the oxygen molecule can be excited by active nitrogen i n the presence of argon. This l i n e requires 4.17 volts for (15) i t s excitation. Okuba and Hamada studied many metallic spectra excited by active nitrogen. Their conclusion was that the highest energy available i s about 9.5 volts. For example, i n the case of mercury the highest l e v e l excited required 9.51 vo l t s . Evidence of the excitation of other levels requiring s l i g h t l y higher energy was not obtained. C. Effects of Temperature on the Afterglow. S t r u t . t h e a t e d l o c a l l y a glass tube through which active nitrogen was flowing. He found that the luminosity of the afterglow was diminished or extinguished at the point of application of the heat, and that the luminosity and 5. attendant a c t i v i t y reappeared after the streaming gas had passed on to a cooler part of the tube. On the other hand he observed that l o c a l cooling caused an increase i n luminosity, accompanied by a faster decay of the afterglow and loss of a c t i v i t y . Heat however applied to a vessel containing the glowing gas diminished the luminosity, and quickly destroyed the a c t i v i t y ; while cooling caused an increase of luminosity and similar faster reversion to ordinary nitrogen. The application of heat i n the two cases apparently causes contra-dictory results. Strutt pointed out that this Aexplicable by the assumption of two processes in the decay of the afterglow. F i r s t l y there i s a wall effect by which the active nitrogen reverts to the normal state without the emission of the after-glow. This effect i s increased by heating. Secondly there i s a process taking place i n the body of the gas, the reversion to the normal state being accompanied by radiation. This process must be delayed by heating. Conversely the body process appears to be actually accelerated by cooling. D. Effects of Pressure on the Afterglow. S t r u t t p o i n t e d out that the effect of cooling on active nitrogen, especially i n the case of the flowing gas, " might be attributed to a l o c a l condensation. To test this p o s s i b i l i t y h e^ 1^ compressed the glowing gas in a vessel by means of a column of mercury. The increase of pressure was accompanied by a large increase i n the intensity of the after-glow, and by a decrease i n i t s duration. This dual effect of. pressure i s evidence that the emission of the afterglow i s not a monomolecular effect. Strutt was not able at that time to make any measurements to determine whether the process involves c o l l i s i o n s of two or more bodies, E. Effects of an E l e c t r i c F i e l d on the Afterglow. In order to investigate whether there are ions present in active nitrogen, S t r u t t ^ 1 8 ) passed a stream of the glowing gas through a tube containing two electrodes, between which there was an e l e c t r i c f i e l d . He found that, provided the electrodes were in contact with the glowing gas, a current passed. He was unable to obtain a saturation current, the increase of potential caused the current to increase t i l l a discharge passed. This subject was further investigated by C o n s t a n t i n i d e s ^ * He employed two sets of electrodes, the f i r s t set acting as an ion-trap to remove any stray ions s t i l l surviving from the activating discharge. By this means he was able to show that the large conductivity observed i s not due to ions from the discharge. Moreover, by varying the area of the electrodes he showed that the conductivity i s due not to ions originating i n the body of the gas, but to electrons liberated from the electrodes. The observation of Strutt, already mentioned, that the electrodes must be immersed in the glowing gas for a current to pass, shows that this emission of electrons i s a wall effect, and not photo-electric. Strutt, i n the research mentioned, had found a large increase i n conductivity to result from the admixture of sodium vapour with the active nitrogen, while mercury vapour caused no such increase. This research was extended by Constantinides to hydrogen and iodine. With hydrogen also no increase i n conductivity was found. In the case of Iodine however an Increase was observed, and this increase Constantinides showed to be due to the produetion of ions in the body of the gas, that i s , to ionisation of the iodine and sodium. F. Rate of Decay of the Afterglow. The rate of decay of the afterglow has been studied by several workers, with a view to determining whether the deactivation i s caused by c o l l i s i o n s of two, or more, bodies. (PQ) Rudy found the rate of decay to be increased by increased pressure, and to follow a bimolecular law for the f i r s t 180 (21) seconds of decay. Eonhoeffer and XCaminslcy also showed that the reaction which produces the afterglow i s bimolecular (22) and not trimoleeular. Kneser v , on the other hand, came to the conclusion that the deactivation i s due to a three-body c o l l i s i o n . This question i s complicated by the dual (16) process of deactivation pointed out by Strut t • later (23) he investigated the effects of various coatings on the walls of a large glass vessel i n which active nitrogen was produced by the electrodeless discharge. He found that by using metaphosphoric acid an afterglow of very long duration can be obtained, showing that the wall effect has been very largely eliminated. He then investigated photometrically the rate of decay and i t s dependence on pressure. Under these circumstances he found that the rate of decay i s consistent with the assumption of a bimolecular process, but the t r i -molecular process, on account of the slight difference in the 8 . curves obtained, i s not excluded. However, an investigation of the dependence of the intensity on pressure gave results i n agreement with a bimolecular hypothesis, and quite incom-patible with a trimolecular hypothesis. He concluded, also, that the excess of neutral nitrogen molecules play no part in the process of deactivation. IT. THE NATURE OE ACTIVE NITROGEN, A. Metastable Molecules. Several theories as to the nature of active nitrogen have been proposed during the course of the experimental researches outlined above. That which i s now most generally accepted and which appears to account best for the experimental (3) facts was proposed by Carlo and Kaplan . According to them active nitrogen consists essentially of metastable molecules and metastable atoms. The f a i l u r e to obtain the A—X trans-i t i o n i n either emission or absorption shows that the A S> state of the molecule i s metastable. Since this l e v e l i s the lower l e v e l of the th, vibrational l e v e l of the A S state being somewhat less 17. than 8 volts, i t appears that a t r i p l e c o l l i s i o n between two metastable and one normal moleaile may be considered. The process would cause the destruction of two metastable molecules and the regeneration of two metastable atoms, thus 2 k0 + normal molecule —> 2 normal molecules + 2P + 2D. Accordingly, i f i t takes place, i t might be also of importance in accounting for the long l i f e of the afterglow. In connection with the long l i f e , a suggestion of (32) Kaplan's i s of interest. Owing to the regions of predissoe-i a t i o n i n the upper vibrational levels of the B ^ T T state, i t i s quite possible that many of the molecules excited to this state by c o l l i s i o n s with metastable atoms do not return •7. d i r e c t l y to the A S state with emission of the afterglow. Instead they dissociate. It i s , as Kaplan puts i t , as though "we allow the molecule to waste i t s time during the process of decay of the afterglow". At lower temperatures, of course, the p o s s i b i l i t y of dissociation i s less, so that the intensity i s greater and the duration of the afterglow i s VI. DEPENDENCE OF INTENSITY AND RATE OF DECAY ON CONCENTRATION. The purpose of the present work was to measure the change i n the intensity of the afterglow due to an instant-aneous change of pressure. We w i l l assume that the meta-stable molecules are deactivated by c o l l i s i o n s of the second 18, kind with normal molecules. The reaction which causes emission of the Of-lands i s then 2A + 2P~^2B + 2S —*2A + 2S -s- 2h —»3A + 2h (1) where A denotes an molecule P denotes a 2P atom B denotes a B^ TT molecule and S denotes a atom. This assumes the atoms to combine at once i n pairs to give an A^Z)molecule. This, of course, i s not so, but the two reactions have been combined as one i n order that a solution of the equations which follow may be obtained. The reaction involving atoms i s precisely similar. For s i m p l i c i t y only one reaction i s considered. The reaction of the metastable and normal molecules i s A + H 2 21T2 + K.E. (2) l e t n]_ = concentration of normal molecules ng = concentration of A^S molecules. Then so that ng = concentration of P atoms, 2n-^ + 2ng + n^ = n a constant. dt dt dt 3- 0 (3) (4) Let f-^ and fg be the c o l l i s i o n frequencies of the two reactions. In (l) two c o l l i s i o n s increase the number of A S 19. molecules by one and decrease the number of 2P atoms by two. In (2) each c o l l i s i o n decreases the number of A^S molecules and Increases the number of normal molecules by one. Thus we have . - f o , fe- £L- f 2 , 5 5 3 - - % . dt " dt 2 dt 1 l o w f l =^ain 2n 37c| + c| f2 =lr°-a2nln2^Cl + °1 = kgn^g. W h e r e ^ a l = ° L 1 ^ '^a2 = ° L l f 2 2 2 the fir's being the Kinetic Theory diameters and the c's the average v e l o c i t i e s . Assuming thermal equilibrium between the various e n t i t i e s , we have 1^/kg = 1/1.09, so that we may take i t as approximately unity. Hence dn„ ., ,c > — 2 = -1 n9n„ - fc?nTnP (7) Division of (5) by (6) gives dn.g _ k-^ dn-j n 3 k2 n l and integration gives so. •0 n« = n 3 0 [ ^ U r ' 2 (8) where n-^ and n^ Q are the i n i t i a l concentrations. Then substitution i n (3) gives 2n± + 2n 2 + n 3 Q p i . o j k 2 = n (9) and i n (4) 2^1 + 2^2 - £l a 30jf n10) £2 4nx = 0 dt dt kg n-Lol11! / d t whence *1 +1 ^ 2 = 1 fSlS30/3aof2 - 2 ) M x (10) dt 2 ) kgn-j^Q^^ / j dt ^ f 1 n 3 " V l ) n 2 by (7) Substitution i n the l a s t l i n e for n g and n^ from (9) and (8) respectively gives 1 ) ^ 3 0 ^ 1 0 ^ 2 - 2(^1 2 J l c 2 n 1 0 / j dt ^ l f e l ^ o f e l o V ^ -2? j" n - n i - Bgo/£i(f 2 (^2n10 1 / J C 2 2 \ n ] _ / dn-, _ T_ dt 2 i ; >£lj§ ~ n l - f ' 3 0 ( f 1 0 ) k2 In this put k-^ /kg = 1. Then •c2 ~ 2 \ n x / J 3c. I f we assume "both and n^ small i n comparison with n-j_ (Strutt , for example determined the percentage of active nitrogen as 2,46%) then SHi * ^ 1 0 ^ > fso^ i o s o t i i a t t h e l a 3 t equation may be written dn-, »i( ni-|) and integration gives = -kodt 1 log/ 1 ) = »lc<>t + n KB. - n n / 2 • For t=0 n± = n 1 0 , so that - f l o j ^ ) \ n i n + n?n * —30 - n-, n ' C = x10 " Xi20 " g O U " n10 2 -, 2nn n - — l o g -LU n ^ i 2 0 • n 3 0 - - log b, n 2n where b = 10 2 n20 + n30 Then b(n ~ 2n x] = _nkr log v W 1 J = - f f 2 t ————— 0 2n-, 2 2 © and solution for n-, gives bn / * -ax\ 2(b + e J 1 7 ^^aft (11) where a = nk g ~2~ Prom (10), putting l^/kg = 1 as before, we have an 2 . i ( f ^ 3 0 .. 2 j a n i so that n 2 . - | ( f 3 ^ 2 + 8 n i ) + c n20 = ~|( n30 4 2 n l o ) + C and C = n + n + n30 _ £ 10 n20 T~ ~ Z n, 2 " f " nl " n10 n50 (12) 2*1 The intensity of the afterglow i s proportional to the c o l l i s i o n frequency of the f i r s t reaction, so that I = KhgUg = i d - - n-, - n10 n30 j n10 Q50 2 2 = Tr/J n10 n30 + ^10*30 _ 4 4 4 ) e"^ n 2 V b ~ ^ ~ ) 2n 2 n L ~2at | ° 2 8 0 e f (13) b n In the present work the intention was to Investigate the effect of change of pressure on the intensity. The rate of decay was not considered. In this case we put t=0, and have simply, 1 = m20n30 (l4> Thus i f the concentrations of a l l reacting ent i t i e s are changed in the same r a t i o by a change of volume, the intensity should be inversely proportional to the square of the volume, or, assuming an isothermal change of volume, d i r e c t l y pro-portional to the square of the pressure. The change of volume should be carried out as quickly as possible, since i t affects also the rate of decajr. For, on differentiating (13) with respect to t and putting t=0, we have ell = K f-a-n n + 4 a n10 n3p' C D b n — j ^-10^30' —Z-Z * i n _ I O c 2 n n 3 o ( 2 n 2 0 4 n3o) ' ( 1 5 ) Thus the rate of decay, under an instantaneous change of volume, i s inversely proportional to the cube of the volume. It may be remarked here that according to (14) the instantaneous value of the intensity i s independent of the concertration of normal nitrogen molecules, while the rate of decay, by (15), i s approximately proportional to the concent-ration of normal molecules. Thus, i f the reactions are as 24. (37)(21) assumed, the observations mentioned above to the effect that the luminosity i s not changed by an increase in the con-centration of normal nitrogen do not prove that normal nitrogen i s not concerned in the decay. This applies also to (23) Rayleigh's conclusion from his experiments on the effect of pressure that normal nitrogen molecules are not concerned. Rudy, as already noted, found an increased rate of decay under these conditions. VII. APPARATUS. A. Source of Nitrogen. Commercial nitrogen was purified by standing i n tubes containing yellow phosphorus, and dried by passing i t over phosphorus pentoxide. The cleaned and dried nitrogen was stored at about atmospheric pressure i n a 500 cc. bulb„ Thence any desired amount could be admitted to the discharge tube through a c a p i l l a r y tube. 33. The Discliarge Tube. The arrangements for generating and compressing the active nitrogen are shown i n Figure 2. A i s the discharge tube, about 45 cms. long and of 2 cms. internal diameter, with aluminium electrodes. This type of discharge was chosen i n preference to the eleetrodeless, as i t permits generation of active nitrogen over a much larger range of pressure. The discharge tube was connected to the compression tube 33 by means of a short tube 33, at right angles to both A and 33. 2 5 e C. The Compression Tube. The compression tube was selected so as to be, as far as possible, straight and of even bore. The front end, E, of the plunger for compressing the active nitrogen was chosen to f i t as closely as possible i n the compression tube, consistent with smooth running. The back end of the plunger contained a core, E, of soft iron wire, IS cms. long. This end of the plunger was somewhat smaller i n diameter than the front end. A short glass ring about 2 cms. long, attached with wax, was used as a bushing to hold this end of the plunger centered i n the compression tube. The two ends of the plunger were joined by a tube of smaller diameter. This arrangement permitted the use of a f a i r l y close f i t for the front end of the plunger while giving s u f f i c i e n t play to allow for I r r e g u l a r i t i e s i n the compression tube. The plunger was evacuated and sealed off, to prevent contamination of the nitrogen. The front end, D, of the compression tube was closed by a pyrex window sealed i n ; the other end was closed by a window sealed on with wax, so that I t could be removed to give access to the plunger. The plunger was moved i n the compression tube by means of a solenoid C, coaxial with the compression tube. The solenoid was given a reciprocal motion by means of a crank and connecting rod, driven by a l/8 H.P. a.c. motor. The stroke was 13 cms. and the speed about two strokes per second. The length of the plunger was arranged so that i t s front end just cleared the opening at C connecting the two tubes at one 26. end of the stroke, and came within 5 cms. of the window D at the other end of the stroke. To prevent scoring of the glass with formation of glass dust and large increase i n the wall effect, the compression tube and plunger were l i g h t l y coated (23) with Apeizon o i l B, which Rayleigh found to have a some-what less wall effect than has glass. The connection to the pumps was from the under side of the compression tube, direct-l y below the opening at C. D. E l e c t r i c a l Arrangements. To generate the discharge a transformer giving about 20,000 volts i n the secondary for 110 volts i n the primary was used. The primary current was about 25 amperes. A condenser was connected across the secondary, and the spark gap and discharge tube in series were i n p a r a l l e l with the condenser, A stream of ai r kept blowing on the spark gap largely increases the production of active nitrogen. The discharge used was intermittent, lasting for about 1/16 second when the plunger was at the extreme back end of i t s stroke, so that the opening at C was clear. This was arranged by means of a commutator on the crank shaft, which broke the primary c i r c u i t . The commutator was a disc of fibre having a brass sector 9 which made contact with two carbon brushes. This arrangement i s not very satisfactory, owing to the large primary current. The glowing gas diffused instantly into the compression tube, so as to f i l l i t s front end, and was there compressed as the plunger moved forward. The discharge must kept cool, owing to the destructive effect . 27 . of heat on active nitrogen. This was clone by wrapping round the whole length of the discharge tube with thin rubber tubing carrying a stream of water. With the intermittent discharge the tube was only s l i g h t l y warm after an hour's run. As uniform conditions of excitation were desirable, the streaming method was not used. The discharge tube was outgassed by pumping for several hours. Heating should not be used i n outgassing, as this appears to increase the wall effect. Nitrogen was then admitted to the desired pressure. The pressure that gave the best production of active nitrogen was found to be about 1.5 mm Hg., but active nitrogen could be generated at pressures from 7.5 mm to .5mm. At the higher pressures however i t does not diffuse f a r . Under the best conditions i t was found possible to operate the tube for an hour with very l i t t l e change i n the production of active n i t r o -gen. I t was intended to compare the intensities with and without the plunger i n operation by means of a Leeds and Northrup MacBeth Illuminometer. The illuminator telescope was mounted i n li n e with and close to the front end of the compression tube. A sectored disc on a shaft p a r a l l e l to the compression tube was turned by a bevel gear mounted on the crank shaft, so that i t made one revolution to each stroke of the plunger. The sector was arranged to allow l i g h t to pass for about l/8 second when the plunger was at the extreme forward end of i t s stroke, so as to measure the intensity at the maximum compression when the plunger was in operation. So that the two intensities to be compared should both be intermittent, the sectored disc was in front of the eye-piece of the illuminometer telescope. I t was found impossible however to malce any comparison of intensities in this way, owing to the short duration of the l i g h t , and. i t s low Intensity In conclusion I wish to thank Dr. G. M. 3brum for suggesting the problem, and for advice given during the progres of the work. I wish also to thank I J? . W. Eraser for design-ing and constructing the mechanical parts of the apparatus. BIBLIOGRAPHY. Lewis E,p. Astrophysical Journal, 12, 1900 p.18. Strutt R.J, Proc. Roy, Society. A93, 1917 p.254, Carlo G, & Kaplan J. • Zeits,f.Phys. 58, 1929 p.769, Herzberg G. Zeits. f. Phys. Lewis E.P, Physical Review. Knauss H.P. Physical Review. Bay & Steiner Zeits.f.Elec.Chem. Kiehlu & Aehaiya Proc.Roy,Soc. Sfsoneri- H. Proc.Nat.Acad.Sci, Ruark, Eoote, Rudnick & Chenault, Journ.Opt.Soe.Amer. Strutt R.J. & Fowler A. Proc.Roy.Society. Mulliken.R„S. Physical Review. Foote, Ruark & Chenault, Physical Review, Kaplan J. Physical Review, Okubo J. & Ham ada H. Phil.Mag, Strutt R.J. Proc.Roy.Society. Strutt R.J. Proc.Roy,Society. Strutt R.J. Proc.Roy.Society. Constantinides P.A. Phys.Rev. Rudy R. Physical Review. Bonhoeffer K.F. & Kaminsky G. Zeits,f.Phys.Chem, 49, 1918 p.512. 1, 1913 p.469. 32, 1928 p.417. 35, 1929 p.733. A123, 1927 p.168. 13, 1927 p.100. 14, 1927 p.17. A86, 1912 p.105. 26, 1925 p . l . 25, 1925 p.241. 33, 1929 p.154. 5, 1928 p.372. A85'," ,1911,p.219. A86, 1912 p.262. A86, 1911 p. 56. 30, 1927 p. 95. 27, 1926 p.110. 127, 1927 p.385. BIBLIOGRAPHY. Contd. (22) . Kheser H.O. Ann. der Phys. 87, 1928 p.717. (23) . Lord Rayleigh. Proc.Roy.Society.. A151, 1935 © 5 G*7 © (24) . Sponer H. Zeits,f.Phys. 34, 1925 p.622. (25). Wrede E. Zeits.f.Phys. 54, 1929 (26). Compton & Boyce.Physical Review. 33, 1929 p.145. (27) . Kaplan J. Physical Review. 42, 1932 © 9 7 © (28). Okubo J.& Hamada H. P h i l . Mag. 15, 1933 p.103. (29) . Cario G.& Kaplan J. Phys. Rev. 33, 1929 p.189. (30). Strutt R.J. Proc.Roy.Society. A91 f 1^22 p.303. (31) . Okubo J.& Hamada H. Phys. Rev. 42, 1932 p * y s 5 c (32). Kaplan J. Physical Review. 37, 1931 p.1407, (33) . Okubo J.& Hamada H.lohoku Univ. Sci.Reports. 23, 1934 © 28 *5 © (34). Jevons W. "Report on Band Spectra Diatomic Molecules". of Cambridge Univ.Press. 1932 p.200. (35) Ham ad a H. Tohoku Univ.Sci.Reports 1932 J3 e 549 © (36) Lewis B. Physical Review. 31 ^ 1923 p.314. (37) Strutt R.J. Proc.Roy.Society. A86, 1911 p. 2 6 6.