-MAGNETIC ROTATION SPECTRA OF NITRIC OXIDE IN THE NEAR INFRARED by Glen Alan Mann AN ABSTRACT ■Submitted to the School for Advanced Graduate Studies Mich i g a n State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Physics and Astronomy 1959 Approved & ■ QJL--- Glen Alan Mann ABSTRACT Observations of the magnetic rotation spectra for the 2-0 and 3-0 vibration rotation bands at 2.7 and 1 .8 ^ , respectively, of nitric oxide have been made. Using a multiple traverse cell of the White type, absorb­ ing paths of .1 to 8 meter-atmospheres were used. A n air cooled solenoid capable of giving 2,400 gauss was con­ structed and the multiple traverse cell was placed in the central third of the solenoid inside a brass tube which held the gas. The magnetic rotation seen in the 2-0 band was limited to the R and Q, branches and only the direction of rotation was determined for the R branch. For the 3-0 band, magnetic rotation was seen for the P, Q, and R branches. In the P and R branches only the direction of rotation for the components was determined. The Q, branch showed large rotations and it was possible to determine the frequency of the rotated lines. The frequencies agree very well with those found in the absorption spectrum. One line appears in magnetic rota­ tion at a frequency predicted by the constants for nitric oxide, but this line is not observed in absorption due to the overlapping of the Q, branches. A small rotation is observed for the first R branch line in the 3-0 band. Any magnetic moment associated with the molecule at such a low value of rotational energy is of the order of .05 fLo • Magnetic rotation spectra, there­ fore, m a y furnish a method of detecting molecules w i t h small magnetic moments. Nitric oxide has a ground state. 3-0 bands consist of two bands, one for p transitions and another for p The 2-0 and q/g “ p TT ijz p t t 3/2 ~ TT 3 / 2 transitions. The magnetic rotation spectra shows that all transitions involving the 2 TT~1/2 state are rotated ia a negative sense while those involving 2 3/2 are rotated m a positive sense, viz. negative rotation is a counter­ clockwise rotation as one looks against the incoming light and positive rotation is clockwise* MAGNETIC ROTATION SPECTRA OF NITRIC OXIDE IN THE NEAR INFRARED by Glen Alan Mann A THESIS Submitted to the School for Advanced Graduate Studies Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Physics and Astronomy 1959 Approv ed__ ProQuest Number: 10008575 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality o f the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest ProQuest 10008575 Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code Microform Edition © ProQuest LLC. ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106 - 1346 ACKNOWLEDGEMENTS The author wishes to express his sincere thanks to Dr. C. D. Hause for suggesting the problem and for his considerable aid and guidance throughout the p r o ­ ject. It was through his and Dr. T. H. E d w a r d s ’ efforts that a grant was obtained from the National Science Foundation to support this project. The author wishes to express his thanks to the National Science Foundation for its grant the research* (NSFG— 5959) to aid TABLE OF CONTENTS Page INTRODUCTION 1 THEORY 4 Hund's Coupling Cases (a) & (b) Intermediate Coupling 4 8 Nitrio Oxide 12 Dispersion 13 Zeeman Effect 16 Anamolous Zeeman Effect 18 The Faraday Effect 23 APPARATUS 28 EXPERIMENTAL DETAILS 34 General 34 3-0 Band 36 2-0 Band 39 RESULTS AND DISCUSSION 40 3-0 Band 40 2-0 Band 60 CONCLUSION 66 BIBLIOGRAPHY 67 LIST OF FIGURES Figure Page 1* H u n d fs Coupling Case (a) 6 2. H u n d ’s Coupling Case (b) 7 3. Transition from Case (a) to Case for a Regular ^ 77- State (b) 11 4. Zeeman Splitting for NO 19 5. Zeeman Splittings and Components 21 6. Absorption and Dispersion Curves 24 7. Spectrograph 29 8. Solenoid 32 9. Fore-Optics 33 10. 3-0 Band Absorption Spectrum 42 10 a. 3-0 Band Absorption Spectrum, P Branch 43 11. 3-0 Band Magnetic Rotation Spectrum -5° 44 12. 3-0 Band Magnetic Rotation Spectrum 0° 45 13. 3-0 Band Magnetic Rotation Spectrum 14. 3-0 Band ft Branch Absorption 15. 3-0 Band ft Branch Magnetic Rotation -25° 50 16. 3-0 Band ft Branch Magnetic Rotation -15° 51 17. 3-0 Band ft Branch Magnetic Rotation -5° 52 18. 3-0 Band ft Branch Magnetic Rotation 0° 53 19. 3-0 Band ft Branch Magnetic Rotation 54 20. 3-0 Band ft Branch Magnetic Rotation f l 5 ° 55 21. 3-0 Band ft Branch Magnetic Rotation t 25° 56 22. 3-0 Band Absorption Spectrum wi t h and without Magnetic Field 57 4- 5° 46 49 23. 2-0 Band Absorption Spectrum 24. 2-0 Band Magnetic Rotation Spectrum -2° 62 25. 2-0 Band Magnetic Rotation Spectrum 0° 63 26. 2-0 Band Magnetic Rotation Spectrum *t2° 64 61 LIST OF TABLES Table I Page Branch Absorption Frequencies and Magnetic Rotation Frequencies 59 INTRODUCTION The magnetic rotation spectrum of a gas is the spectrum of the radiation transmitted through crossed polarizing elements when the magnetized gas is placed between them so that the light traverses the gas in the direction of the magnetic field. Thus, the spectrum is essentially an enhanced Faraday Effect closely associated w ith the longitudinal Zeeman Effect. Magnetic rotation spectra have been observed in monatomic and diatomic gases. Macaluso and Corbino covered the effect in monatomic gases. dis­ The magnetic rota­ tion spectrum of a diatomic gas, sodium vapor, was first 2 observed by R. W. Wood • Magnetic rotation spectra have 3 4 also been observed by Righi , Loomis , and Carroll . All c of these observations were made on the diatomic vapors of alkali metals, and iodine, bromine, and bismuth. These spectra were confined to the photographic region and were associated w ith electronic band systems in w h ich the upper state alone possessed magnetic properties, or the upper state was perturbed b y some close lying state w h i c h exhibited magnetic characteristics. Since magnetic rotation is closely associated with the Faraday Effect, which in turn depends upon the longi­ tudinal Zeeman Effect, magnetic rotation spectra will only -8be observed for those transitions which show an appreci­ able Zeeman splitting. plane Even though the rotation of the of polarization becomes abnormally large at absorp- tion frequencies, the amount of light transmitted through the analyzer is relatively small. spectra that several The magnetic rotation have been observed required exposures of hours. These spectra were associated with elec­ tronic transitions and the only lines that were seen occurred at band heads and low J values. this is that The reason for at band heads there are several absorption lines veiy close together, and the rotations for the indi­ vidual lines all add together giving a large rotation. The low 1 value terms sometimes observed are favored b e ­ cause of their relatively large Zeeman splittings. Since these spectra were taken, detection techniques in the infrared have been improved by the introduction of Lead Sulphide detectors. Gratings are at present being ruled which give energy distribution curves with maxima in the near infrared. The introduction of multiple traverse cells, such as the White type, give large absorbing paths without the need for high pressures so that pressure broadening need not be the limit of resolution. It was for these reasons and also because the N O molecule has both u p p e r and lower states with magnetic characteristics -3that it was thought the magnetic rotation spectra of nitric oxide in a single vibration-rotation band might be observable. -4THEOHT Hund's Coupling Cases (a) & (b) The electronic states and rotational energies of a diatomic molecule are classified by the m a nner in w h i c h the angular momenta of the electrons are coupled and the way in whi c h they are coupled to the other motions of the molecule. Different schemes of coupling of the internal angular momenta were first considered by Hund. We will be concerned only with the types of coupling known as H u n d ’s case (a) and case (b), and the coupling intermediate between (a) and (b). In H u n d ’s case (a), it is assumed that the inter­ action of the nuclear rotation with the electronic motion is very weak, but the electronic motion itself is coupled very strongly to the line joining the two nuclei. The electronic orbital angular momentum (L ) is coupled very strongly with the electronic spin (S) and their projections along the internuclear axis, - A - and respectively, combine to form a resultant_£Lwhich in turn is coupled with the angular mo m e n t u m N of nuclear rotation to form the resul­ tant total angular momentum The vector J is a constant in magnitude and direction. X L and N precess about J. At the same time L and S precess about the internuclear axis. -5In H u n d ’s case (a), the precession of L and S about the internuclear axis is assumed to be very m u c h faster than the precession of-f^-and N about J* In H u n d ’s case (b), the electron spin S is u n ­ coupled from the electronic orbital angular momentum L* That is, the orbital angular momentum L is coupled to the internuclear axis and combines w ith the angular m ome n t u m N of the nuclear rotation to form a vector K. The angular momentum K combines w ith the electron spin S to form the total angular momentum I w h i c h is constant in magnitude and direction. Thus, in case (b), we have S and K precessing about J while - A . and N precess about K. In Hund's case (b) the precession of I and S about J is assumed to be very slow compared with the precession ofVl. and N about K^. Vector diagrams of these coupling schemes are shown in Figures 1 and 2. N HUND'S COUPLING CASE "A" FIGURE I r HUND'S COUPLING CASE "B" FIGURE 2 -8Intermediate Coupling Hund's coupling cases represent idealizations which are never fully realized, although they frequently r e p r e ­ sent a good approximation to the coupling for the molecule. It sometimes occurs that w i t h increasing rotation, the coupling of a molecule m a y change from one idealized case to another. W e want n o w to look at the transition from case to case (a) (b) coupling which is called spin uncoupling. If one assumes a molecule coupled according to case for small rotational values, along the internuclear axis. the molecule rotates. (a) then S is coupled w i t h .A. S precesses around V L while As the rotation increases, its value approaches that of the precessional frequency of S, and if the rotation of the molecule increases still more the influence of the molecular rotation will become the predominant one. W h e n this happens, S can no longer couple itself w i t h - A - along the internuclear axis. Instead, will combine w i t h N to form a resultant K, w h i c h then combines w i t h S to give the total angular momentum of the molecule J, whi c h is case (b) coupling. The rotational term values for the components of doublet states have been obtained by Hill and Van Vleck for any strength of coupling between -A and S, but neglecting -9the coupling between K and S* They obtained, i Ufr+i)\y(w)/£']-K f# ) - H,[(r-fij'tjC+i % where Y - yiy-vAT 7' V - M v A Bv , where A, the coupling constant, is the separation of the two components and is determined from the equation ^ -f / ? A £ where T@ is the elec­ tronic energy of a multiplet term and T Q the electronic energy w he n spin is neglected. By is inversely propor­ tional to the moment of inertia of the molecule about a respectively* (b). -10Figure 3 shows the shift of energy levels for the transition of a regular from case (a) to case (b)* state when the coupling changes -11- J ¥ K Z n t / / \ / / 2 / / / X V ' " _________________________/ - r / / ¥ / "> / ----------------------- - X 1 --- V \ \ \ \ 5 f 6 > / ( / / / / / / / = ) < 42 = = = n = /' / / > / / 3 2 ____________________________ I C A S E "B" i-------------------- '/ //'/// / / / i n t ' // i ---------------------- / ^ / j C A S E "A* TRANSITION FROM CASE"A" TO CASE "B" FOR A REGULAR2n FIGURE 3 STATE -12Nitrio Oxide Nitric oxide is a colorless, slightly heavier- than-air gas w h i c h is toxic in very small quantities and poisonous in moderate quantities* It is a diatomic molecule which has a molecular weight of 30.01 and whose atomic number is 15. Its normal freezing and boiling points are -163°C and -152°C, respectively. Nitric oxide is the only chemically stable diatomic molecule having an odd number of electrons, a resultant electron spin. indicating The projection of the elec­ tronic spin angular momentum along the internuclear axis^JEL^ mu s t be half-integral. with Since this spin vector can combine the projection of L, along the internuclear axis in one of two ways, the ground state of N O must be a doublet state. The type of electronic state, that i s 2 T T , A . . . » is determined by the electron configuration of the molecule. The electron configuration for N O is k Tr^Zp The fifteenth electron is a T T electron. the ground state will be a state of nitric oxide is 2 state. j]~ and This means that Thus, 2 y p (8) the ground -13Dispersion Dispersion can be defined as the change in the index of refraction of a m e dium with the frequency of transmitted radiation. When the index of refraction is smaller for long wave lengths than it is for the shorter wave lengths, we say the dispersion is normal. dispersion can be explained by assuming, materials, Normal for transparent absorption frequencies in the ultraviolet due to electrons and absorption frequencies in the infrared due to ions. In normal dispersion formulas a term appears in the denominator which involves differences between the squares of an absorption frequency and the frequency of the incoming radiation so that when the radiation frequency is equal to this absorption frequency, the value of the index of refraction for the material becomes discontinuous. Whenever the index of refraction is greater for longer wave lengths than it is for shorter wave lengths, the dispersion is called anamolous. general at absorption frequencies. This takes place in It is anamolous d i s ­ persion in which we are primarily interested. In order to overcome the difficulties in the ex­ pression for normal dispersion which arise at absorption frequencies, one introduces a term similar to the damping term in the equation for a damped harmonic oscillation. -14This term makes the theoretical value for the index of refraction remain finite w h e n one passes through an a b ­ sorption frequency. Helmholtz approach, One can use either the mechanical or the electro-magnetic approach or the quantum mechanical approach. In any case, one ob­ tains similar relations between the index of refraction and the frequency. The difference arises in the meaning attached to the symbols. In the mechanical and electro-magnetic approaches, one considers forced oscillations of the electrons and the ions by the incoming light wave and the natural oscil­ lations of the electrons and the ions in the material. the quantum mechanical case one considers the frequency of the incoming radiation, the frequency corresponding to transitipns of electrons from one energy state to another, and the transition probabilities. The quantum mechanical representation of the index of refraction as a function of frequency for normal dispersion is (9) where N is the number of dispersion electrons per unit volume, e is the electronic charge, m is the electronic A,. is the induced transition probability from In -15state j to ll by the frequency is the spon­ taneous transition probability from the state J to ^ ■»W is the frequency difference between the states and- , and » J ]) is the frequency of the incoming radiation, A classical expression for the index of refraction for anamolous dispersion and the absorption that takes place at the same frequency is given by the following expressions: for the index of refraction, „ ,„ and 4.4 4 . 4 n+x 0 + 7 •, tVt ~ V 1-+ 4 x f(t (I ' for the absorption where ft (1 0 ) ' is the average contribution to the index by all the other resonance frequencies hJ ( l except tOQ f a is ■------- ^ >t 0 is the dielectric constant for vacuum, g is the damping constant, than one a number m u c h less to make the absorption peak sharp, AC. absorption coefficient for the medium, and is the X “ ^ — — -° where (aJ0 is the natural frequency of the dispersion electron, tion. and U) is the frequency of the incoming radia­ -16Zeeman Effect The Zeeman Effect is the effect an external m a g ­ netic field has on the radiation of an atom or molecule. When the radiation is observed perpendicular to the field the effect is called the transverse Zeeman Effect. The longitudinal Zeeman Effect is that observed when one looks along the direction of the magnetic field. The presence of the magnetic field removes the degeneracy of the space quantization by giving a preferred direction around which the total angular momentum of the molecule can rotate. The classical Zeeman Effect is obtained by reducing the motion of the electrons in the molecule to three compound motions, to be elastic: in each of which the binding is assumed one is a linear harmonic motion along the field direction and the other two are uniform circular motions in opposite directions in a plane perpendicular to the field direction. We will concern ourselves only with the longitudinal Zeeman Effect. In the normal pattern, only two lines are seen and these lines are right and left circularly polarized. classically, Only two lines are seen since, there will be no radiation emitted along the field by the motion of the electron which is vibrating parallel to the field direction. The two lines that are -17seen are circularly polarized since they come from the two circular motions perpendicular to the field* One of these has a higher and the other a lower frequency than the line seen without the field, since one of the rotations is aided by the magnetic field while the other rotation is retarded by the field. The shift in frequency from the field free case is given by VJ- W 0 = Bohr magn et on equal to (L U. — ----- H iTtnc A m where jdLa is the ,and is referred to as -18 Anamolous Zeeman Effect In a magnetic field the total angular m omentum J of the molecule is space quantized such that the com­ ponent of J along the field is M"fc where M = -J, - J + 1 ...I. The states with a different M have somewhat different energies. The energies are given by W = W q ~ j H, where Wq is the energy of the molecule in the absence of the field, ^ Cjj is the mean value of the component of the m a g ­ netic moment in the field direction and H is the value of the field. In order to determine the Zeeman Effect for a molecule one must evaluate ykc ^ for that particular molecule. If the molecule is coupled according to H u n d ’s case the value of ytc ^ is found to be where jji q is the Bohr magneton. according to H u n d ’s case Ah r ' "** ^ (a), A 1 |U.0 If the molecule is coupled (b) the value of j/. jj is given by -f a. ft, K ( i c + 1) Since most molecules are neither pure case pure case (a) nor (b), one must obtain an expression for jJL ^ or the Zeeman splitting for the intermediate coupling case. 11 12 This has been done by Hill and has been used by Crawford Using the equations given by Crawford, Figure 4 shows the m ax im um Zeeman splitting^M = i J as a function of J or K for nitric oxide in terms of the splitting for the **19* .5 - A z/n o — PURE CASE A T= - M -.5 — K= I 2 3 4 5 6 7 8 9 10 12 13 1.0 5 PURE CASE A A vH 0 — -.5 — - 1.0 — T-.M ZEEMAN SPLITTING FOR v =o, i,2,3 FIGURE NO 4 14 15 16 -20 normal Zeeman Effect. The curves in Figure 4 also show tbat w i t h increasing rotation the coupling scheme for NO goes from case at (a) to case (b). For the upper curve K ® /, Z ss 1/2, there is no splitting, meaning that S is antiparallel t o , case (a). As the rotation of the molecule increases S uncouples from coupled to the vector K, case (b). J = + M has the greater energy, and becomes Since the condition S is parallel to K. In the lower curve for small values of J, S is parallel to or nearly so since I 5 4 M has the higher energy, case (a)* As the rotation increases a point is reached where the sum of and S is zero resulting in no magnetic moment for the molecule. This occurs at about K = 10. As tie rotation increases still more S becomes coupled to K, case (b), and is actually antiparallel to K, or nearly so, since J “ - M has the higher energy. For the longitudinal Zeeman Effect the selection rule for transitions is ^ M = ± 1. The transition 21 M - + 1 gives lines shifted to higher frequencies and left cir­ cularly polarized, while transitions of 4 M = - 1 gives lines shifted to lower frequencies and right circularly polarized. Figure 5 indicates the appearance of the Zeeman components for several values of J or K, depending on whether the molecule approximates case (a) or case (b) - 21- C A S E "A" 27 CASE 68 __L B T I ZEEMAN SPLITTINGS AND COMPONENTS FIGURE 5 -22 coupling* The separation between components is given only as a function of the various M values divided by J U - f i) or K(K-t- 1 ) and it is assumed tbat the other terms are a constant. 2 TT i/e state, For the case (a) type coupling of the the quantity 2 = 0 , so that this substate will not show any splitting. “XT’ 3/2 state is shown. Thus, only the It can be seen that for case (a) splitting the number of components is 2J4-1 and as J increases, rapidly. splitting. the separation between components decreases For high J values there is practically no For case (b) coupling the number of compon­ ents is 2(2K-t-l) and as K increases the separation b e ­ tween components rapidly decreases. For large K values the components are no longer resolved and the resulting splitting is practically that of the normal Zeeman Effect. With the magnetic fields normally employed, usually around 20,000 gauss, the Zeeman components of the individual lines in a vibration-rotation band are not resolved except for low J values and general broadening for higher J values. A picture of the Zeeman Effect in the 5-0 band of N O for H = 2,400 gauss is shown in Figure 22. The only notice­ able change in the appearance of the spectrum is a broaden­ ing of the Q, branch. -23The Faraday Effect The Faraday Effect is observed by passing plane polarized light through a substance which has been placed in a magnetic field and observing the transmitted radi a­ tion through an analyzer w he n the direction of propagation of the radiation is along the magnetic field. Effect is observed in all transparent media. The Faraday The Effect is us ua l ly small since most transparent substances are diamagnetic. It is most easily observed in a paramagnetic m e d i u m near natural absorption frequencies for that medium. W hen the magnetic field is turned on there will be for each absorption frequency without the field at least two absoiption frequencies, one for right and the other for left circularly polarized light according to the theory of the Zeeman Effect. rotation, curve. For each of these directions of one may draw an absorption curve and a dispersion A total dispersion curve may be drawn for the whole region. This is shown in Figure 6 . one of the dispersion curves is A T In this figure, and the other is ft_ , depending on w h i c h absorption curve belongs to the right circular or left circular component. The total dispersion curve at the bottom of the page is then either ft+— jr\ — . or From this drawing one notices that outside -24 ABSORPTION ZEEMAN COMPONENTS DISPERSION RESULTANT CURVE CURVES DISPERSION FIGURE 6 CURVE -25the absorption lines one of the circular polarized radia­ tions travels faster than the other, while between the absorption frequencies the reverse is true. This means that the plane of polarization will be rotated in one direction outside of the two absorption frequencies, while the plane of polarization will be rotated in the opposite direction between the absorption frequencies when the two circular components recombine to form a plane polarized beam upon emergence from the field region. A n examination of the resultant dispersion curve shown in Figure 6 indicates that rotation would probably only be detected outside of the absorption lines. The rotation whi ch occurs between the two Zeeman absorption components will probably be hidden by the absorption that takes place at these frequencies. For this reason a ba l­ ance must be reached between the amount of absorption and the amount of rotation. If the absorption is increased by increasing the pressure of the gas, then a point will be reached at w hich the absorption line will mask all of the rotation. At this pressure the absorption line will be broad enough to cover the rotation which normally could be detected just outside of the frequencies of the Zeeman absorption components. In Figure 6 it has been assumed that the splitting -26of the absorption line results in two Zeeman components of equal intensity. It m a y happen that the intensities of the two components in the magnetic field are not equal. In this case then the two dispersion curves will not be the same and the resultant dispersion curve which is the difference of the two individual dispersion curves will no longer be symmetric as shown in Figure 6 , but will be assymmetric about the field free absorption frequency. The angle of rotation can be shown to be where £ is the path length in the medium, A l ength of the radiation, and is the wave is the difference in the indices of refraction for right and left circularly polarized light. The above is a description of how one line is formed for a magnetic rotation spectrum. The Faraday Effect, unlike substances possessing natural rotary dispersion, is increased when the radiation is reflected and sent back through the magnetized sub­ stance. The magnitude of the rotation of the plane of polarization for the Faraday Effect is proportional to the length of the sample in the magnetic field and to the field strength. The constant of proportionality is called the Verdet Constant. dependent. This constant is frequency It assumes abnormally large values when one passes through a magnetic absorption frequency for a given material. -27Serber 20 has developed a general expression for the Verdet constant occurring in the Faraday Effect for d i a ­ tomic molecules. He obtains the following expressions: ^ P ^ S ( n n 'J P P'')* + PCh'^r-P^ 7(P0>"')^~P'L) 9 The term w h i c h is most important when one is near an absorption frequency is JJ ^(nn'J Serber has found for this term that the Verdet constant can be given as the following: \j _ j ^ <£- P*Cn'n) P ,(r>'m 'in n j[^K (n m )^'n>') ~ where C is P ' L n ' n l 7- ) ~ is the absorption frequency, the Zeeman frequency shift, B is /1-rr ^ j / ti*j] Js f (/?'/*» '• ntrx) is w / t i "V) and Of'/fcy p/*) ~~ ( hm j ih p '/h ’J ?u ft/HjJ ” * * 0 Ctwtoj P r*\ J • This, then, says that near an absorption frequency the rotation depends only on the magnitude of the Zeeman splitting and the amplitude of the absorption without the field. -28APPARATUS The vacuum infrared grating spectograph used in this w o rk was designed by Dr. R. H. Noble^®. the Pfund pierced flat type. It is of The optical alignment of 14 . this spectograph has been outlined by R. G. Brown A schematic diagram (Figure 7) shows the path of the radia­ tion through the spectograph. The spectograph is arranged to allow two light beams to traverse the monochrometer section at the same time. One beam goes to the Lead Sulphide detector and the other is sent to a Fabry-Perot Interferometer. The Edser-Butler bands produced by sweeping the interferometer w i t h continuous light are detected by a photomultiplier and recorded. bration. These band contours are used for cali­ This system is referred to as the Fringe Calibration System and is explained in detail by B. H. V a n Horne-^. A 300 watt concentrated zirconium arc was used for the infrared source and a 100 watt zirconium arc was used to obtain the fringe calibration. A Lead Sulphide detector cooled to -50°C was used to detect the infrared radiation coming from the monochrometer section of the spectograph. In order to obtain suitable absorbing paths, multiple traverse cell of the White type was used. a The •29- GRATING V \ DE TE C TO R CHOPPER <■ ,R P? SOURCE 'r // Tr Ij •f=f F- P INTER. FRINGE SOURCE TUBE SPECTROGRAPH FIGURE 7 -30cell was adjusted for 24 traversals, of eight meters, giving a total path and the pressure of N O was varied from 1 to 56 cm of mercury. This multiple traverse cell was placed inside a brass cylinder which was fitted with windows to allow the radiation to enter and leave the tube after multiple reflections in the White cell. The brass tube w a s then inserted in an air cooled solenoid. The multiple traverse cell was placed in the central third of the solenoid. The solenoid consists of 36 pancake windings of insulated copper strips 1/2 inch wide and 1/16 of an inch thick. E a c h coil contains 90 turns. The total solenoid is 82 cm in length, which means the solenoid has 3,950 turns per meter. The 36 coils were all connected in series giving a total resistance of about 4.2 ohms. The magnetic field in the center and at places covering the central third of the solenoid was measured by nuclear magnetic resonance techniques. The variation of the field in the central third of the solenoid was not greater than 5fo of the value of the field at the center. The value of the magnetic field at the center of the solenoid with a cu r­ rent of 38 amperes was measured as 1,650 gauss. If one compares this value with the one obtained by using the formula for an infinite solenoid, the value for the actual -31solenoid is 87% of that for the theoretical infinite solenoid. Magnetic rotation spectra were obtained with fields of 2,400 gauss w h i c h meant there was a current of 55 amperes in the solenoid. The completed solenoid was placed on an aluminum frame to match the height of the spectograph bed. A picture of the completed solenoid is shown in Figure 8 . A schematic diagram (Figure 9) shows the light path from the source through the absorption cell and up to the entrance slit of the spectograph. Two vibration-rotation bands of NO were investi­ gated — the 2-0 band at 2.7^, and the 3-0 band at 1 . 8 ^ ^ For the 3-0 band the polarizers used were type HR polaroids from the Polaroid Corporation. These transmit about 40% of the incident radiation at 1 .8^6 when both polaroids are used w i t h their axes parallel and almost 0% when their axes are crossed. These polaroids are in the form of plastic laminations. 16 For the 2-0 band, multiple plates of silver chloride set at B r e w s t e r ’s angle for 2.7 were used. These polarizers were loaned to us by Professor C. W. Peters of the University of Michigan Physics Department. Figure ^ SOURCE SPECTROGRAPH ENTRANCE SLIT -POLARIZERS SOLENOID AND CELL FORE - OPTICS FIGURE 9 -34EXPERXMOENTAL DETAILS General Since the amount of energy reaching the detector of the spectograph is dependent upon the angle the electric vector makes with the rulings of the grating, the analyzer was placed in position first and rotated until a maximum signal was obtained* was found, When this position the analyzer was not touched again. Both the polarizer and analyzer had an angular scale and pointer so that the angle of rotation fr om the crossed position could be measured. The magnetic rotation spectra were obtained as a function of the angle between the planes of polarization of the polarizer and analyzer. It was decided to call the angle between the polarizer and analyzer 0° when they were in the crossed position since neither the rota­ tion in a plus or minus direction would be aided or retarded in this case. A rotation of the polarizer in a clockwise direction as one looks against the incident radiation was called a positive rotation. A rotation of tie polarizer in a counter-clockwise direction was termed a negative rotation. The magnitude of the rotation is given by the number of degrees the polarizer is rotated away from the crossed position. -35- The direction and magnitude of rotation for a given line can thus be determined. As an example, consider a line whi ch is rotated in the negative direction by 10°* Neglecting the absorption that takes place since the polarizers are no longer crossed, a rotation of the polarizer to •+* 10° will make the radiation from this line have its plane of polarization perpendicular to the plane of the analyzer and this line then should not be seen. All of the magnetic rotation spectra were obtained in this manner* -363-0 Band The 3-0 hand of nitric oxide is located at a p ­ proximately 1.8 . This band has been investigated extensively for three reasons: the grating used has 6,000 lines/cm and is blazed at about 27° or l.S^c in the first order, the Lead Sulphide detector has a rela­ tively high efficiency at 1.8 » and the 300 watt Zirconium arc gives a good amount of radiation at 1 .Qft • The P and R branch lines were obtained in magnetic rotation using slits 100 and 150 wide, corresponding to .60 wave numbers. Gas pressures from 1 to 18 cm of mercury were used in the multiple traverse cell, giving absorbing paths ranging from .1 to 2 meter-atmospheres. It was found that pressures in excess of 20 cm of mercury caused the magnetic rotation spectrum in the P and R branches to disappear. This is probably due to self-absorption as mentioned on page 25. The P branch in the 3-0 band has both sub-band components well resolved, while in the R branch the sub­ bands are not resolved at the slit widths that were used. The magnetic rotation spectra were obtained for the above pressures first with the polaroids in the crossed position and then set at progressively larger positive and negative -37angles until the P branch showed absorption for one com­ ponent and rotation for the other component. It was not possible to determine with high accuracy the frequencies in the P and R branch lines for the magnetic rotation spectra. Only the direction of rotation for the compon­ ents was determined. The magnetic rotation spectrum of the Q, branches were obtained using slits of 25 and 3 0 > ponds to .12 wave numbers. which corres­ With slits of this size it was possible to run the fringe calibration system and the magnetic rotation spectra simultaneously. This allowed a fairly accurate determination of the frequencies for those lines in the Q, branches which are rotated. The Q branch rotations were obtained using a gas pressure of 56 cm of mercury in the multiple traverse cell, giving a total absorbing path of about 6 meter-atmospheres. The magnetic rotation spectra for the Q, branches were taken by starting at the R^(3/2) line with the polaroids at ± 25°. Wh en the R^fl/2 ) line had been recorded the polaroids were turned to the crossed position or to the angle, either positive or negative, that was desired and the Q, branches examined for magnetic rotation at this setting. After the Q, branches had been observed, the polaroids were returned to the original setting and -38two P branch lines recorded* The frequencies of the R and P branch lines are known and w it h the aid of the fringe calibration system the frequencies for the rotated lines in the Q, branches were obtained. -392-0 Band The 2-0 hand of nitrio oxide is located at approxi­ m a t e l y 2.*7ju. * Only the Q, and R branches of this band have been observed. This is due to the rapid decrease in sensitivity of the instrument in this region. The decrease in sensitivity can be attributed to the follow­ ing reasons: both the Zirconium arc and the Lead Sulphide detector have envelopes which become absorbing around 2.7j/. , a Germanium filter must be used in order to absorb the radiation below 2 , and the grating that was used for these spectra had to be employed at an angle far f ro m the blaze of the grating. In order to observe the 2-0 band the slits had to be opened to 200 and 250ji , corresponding to .28 wave numbers. Pressures of 10 cm of mercury were used which gave an absorbing path of about 1 meter-atmosphere. Only the direction of rotation for the sub-bands could be determined for the 2-0 band. With the 2-0 band we have direct evidence of the direction of rotation of the individual R branch sub-bands since several of these lines are well resolved. -40RESULTS M B DISCUSSION 3-0 Band As has been mentioned previously, relatively large slit widths were necessary in order to obtain the m a g ­ netic rotation spectrum of the P and R branches in the 3-0 band. In order to determine which lines appeared in the magnetic rotation spectrum, was employed. the following procedure An overlay was obtained for the P and R branch lines and the components marked so that the m a g ­ netic rotation lines could be identified by their J number. The absorption spectrum of the 3-0 band is shown in Figure 10. The next three figures 1 3) show typical magnetic rotation spectra. (11, 12 and The angular notation on these figures gives the angle between the planes of polarization from the crossed position. Here one notices an interesting phenomenon. In the R branchy the line at I - 19/2 is missing in magnetic rotation and as the angle is increased in the positive or negative direction, the absorption for this line is greater than that of any other R branch line. If one examines the curves for the Zeeman splittings shown in Figure 4, it will be noted that the Zeeman splitting for j- s 19/2 in the state is almost 0. Since this -41line is m is si ng f r o m the magnetic rotation spectrum, we have here an experimental indication that the magnitude of the magnetic rotation depends upon the magnitude of the Zeeman splitting. If the R(19/2) line is missing then one would expect that a similar line, F ( 2l/2 ), should also he missing. This, however, is not the case. We have al­ ready seen that the magnetic rotation spectra depends upon the magnitude of the Zeeman splitting. It is also dependent upon the strength of the o r d i n a l absorption line since the change in the index of refraction is d e pe n­ dent upon the magnitude of the absorption. P(2l/2) Since the line is visible in magnetic rotation and the Zeeman splitting for this line is small, then the absorp­ tion for this line must be large. This, in fact, is true. Whi le studying the fundamental of N 150, Fletcher and B e g u n 17 observed the P(2l/2) line due to the presence of 6fo of N l40 and only the P(2l/2) line was observed. We have also noticed that the P(2l/2) line for the component is abnormally large as can be seen in Figure 10a. One can easily determine the direction of rotation for the sub-band components in the P branch since they are p well resolved. The rotation is negative for the Figure 10 Figure 11 QNVQ 0-£ Figure 12 ~45~ Figare 15 "*46* -47cojn.pon.ents and positive for the ^"77^3/g components• u s i n g the JHissing line at R = 19/2, By it is possible to deduce the direction of rotation for the sub-band com­ ponents in the R branch even though they are not resolved* F r o m the absorption spectrum it is known that the abp sorption for the component is m u c h stronger than P for the t T ryp component* Knowing the above facts, one can say that the direction of rotation which gives the maximum absorption for the R(19/2) line corresponds to the P branch line w h i c h becomes weaker in rotation for this same direction* g This component in the P branch turns out to be the TTl/p component* Thus, the R branch lines corresponding to 7 T -jyp transitions show negative rotations, while the R components for the 2 jj~ transitions show positive rotations• The Q, branch rotation seen for these wide slit openings and for the two directions of rotation of the polarizer indicate that the Q, component may show both positive and negative rotations. Since the Q, branch was m u c h stronger than any of the other rotations, it was possible to narrow the slits and examine its structure* The strength of the Q, branch in magnetic rotation may possibly be explained by its Zeeman splitting. The -48Q, branch is formed by transitions occurring for /\ J - 0* This then means that all the Zeeman components will have the same frequency shift. Thus, the Zeeman pattern for any one Q, branch line will appear to have only two components. Also, the splitting for the Q, branch lines is larger than that for any of the P or R branch lines. The next eight figures show the Q- branch in m a g ­ netic rotation w i t h narrow slit widths. Figure 14 is the normal absorption for the Q branches. seven figures The next (15 - 21) show the Q, branches in magnetic rotation for crossed polaroids and plus and minus direc­ tions of rotation of the polarizer. curves are composite curves. in the crossed position, by the analyzer. Thus, These magnetic rotation When the polaroids are not some radiation is transmitted there will be some absorption taking place along w ith the magnetic rotation that is observed. It was decided, therefore, to run absorption curves for the Q, branches at the same angles to be observed in magnetic rotation. If these absorption curves are then subtracted from the curves obtained with the field, what remains should be a true picture of the magnetic rotation in the Q, branches. These are the curves that are shown in Figures 15 through 21. -4 9~ CM a CM Q BRANCH FIGURE ABSORPTION 14 50» - 25 ° <2 0> FIGURE 15 -51 15 - ° CM