PAP-MAGNETIC SE‘LEWMG 1N ELECT RON HE SOLUHONS HYPERFENE METALuAME RESONANCE STUDIES UP the Dogma a? M. 5. Thesis for SETY MlCH‘GN‘i STATE UNWER john E. Wraede 1963 MICHIGAN STATE UNIVERSITY DEPARTMENT OF CHEMISTRY EAST LANSING, MICHIGAN ’ , . ‘ o z ‘ -‘vv ABTTRiCT HYPERFTho SILITTTNE IN EL -Cwi’" FL.A‘vJucTIu mo3333133 STDJIE3 CF ICTh-Um'“ W“. L'I‘ICZIES by’John E. Hreede Solutions of lithium in ethyiar inc, cesium in ethyiamine, and rubidium in methylamine were studied by means of electron paramagnetic absorption using a Varian Modai LEO? Spectrometer with 100 KC field modulation. The primary purpose was the observation of hyperfine splitting but only rubidium in methyiamin' soiution evr tibited hyper- fine structure under the conditiono used. This solution gave an eight«1ino spectrum at room temperature but another lino in the cantor bocome apparent at low temperatures. The total hyperfine separation was plotted.as a function of temperature giving an essentially linear relationship in the range studied (-LCOC to #6006). Below the freezing point of the solution, the splitting reverted to neariy'its room teoperoturo value. A decomposed sample of Li-ethylamine solution had a broad line which may have been due to hyperfine interactions, since hyperfine Splitting has been reported for this solution. HYPER. HE SPLITTIEG IN ELECTRGN PflRANAGEETICIRESCiANCE STUDIES 0? rat-i111": summons; By John E. fireede A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of HASTER CF'SCIENCE Department of Chemistry 1963 To my wife Trudi ACKNCHLED7MENTS The author wishes to express his appreciation to Professor James L. Dye for his guidance, assistance, and encouragement during the course of this investigation and the preparation of this thesis and to Mr. Henry A. Kuska and Dr. Robert R. Dewald for their assistance in the investigation. iii I. II. III. IV. V. VII. VIII. TABLE CF‘CflfiT IBIECDMICN . . . . . . . . HSMIC“ O C I O O O O O I 1-: Eigi‘! O I O O O O O O O O O EXPERI A. B. C. D. 2" L.‘ O F. firffi‘l. ru‘atiAL o o o 0 o o a o Glassware Cleaning . Amine Purification . Sample Preparation . EPR Spectra . . . . . Temperature Determination Optical Spectra . . . . . up“? REJLnJ-Joooooo-ooo A. B. C. Lithium in Ethyiamine Cesium in Ethylanine Rubidium in Methylamine DISCUSSIC? . . . . . . . . . A. B. Lithium in Ethylamine Rubidium in Hethyianine FUTURE ”EEK o o o o o o e o‘ O REFEHXEQJCJD I o o o o o o o o 0 iv Page 1. 2. 3. ll. 12. .4 o. O O :1.“ \«. ifimriim splitting at C??33L13T:Li?§nji:fific fie 1d . . . lifeerfine sylittin; at cor. tant: freeman; . .~ . . . Amine put .i‘ication twin . . . . . . . . . . . . . . Li-wryie arr-2mm) vessel . . . . . . . . . . . . . . :‘bmmle mi-ae-up wen-3e}. . . . . . . . . . . . . . . fsrtzridrntlon of line wicith end :3 teams . . -. o . . Lii,hittn-»e.‘*,'ii'ine 3:73:32 9; retire . . . . . . . . . . . Line spacings in 2?.12.-::e&;.h;,r1*.iz::.~im spectre . . . . . . Te went m nde..ce of Apiitting in flame: miwim enrsfic*r3............... O D I iirzmrimntal vs. cairn,» ated mxrve of "Wm-y ant-w] mine 1- 33131103 0 o o o o o o o o o I a o e 0 o o o o fifffict Qf dgcan9afiitim on .;~flfltig-E.ira spec re. a Eiimjreeentzaiive tezmerewm spectre of Ethane:Wins-nine 3C1Utifin o u e o o e o . o o o o o o o o n o 0 Y‘., 12%;? 11.1 :# i} I. INTRODUCTION Hyperfine splitting in epr is important because of the large amount of date readily obtainable by simple measurement. In simple cases the nuclear Spin quantum number is determined merely by counting the number of peaks (given by 21 + 1). If one knows the nuclear mag- netic moment (obtainable for most nuclei from varian's MAR table)’, the electron density at the nucleus can be calculated from a simple form- ula (see section on theory) using he total spl tting. Of course, the data obtainable from just a single peak are also obtained when hyper- fine splitting is observed. magnetic susceptibility, gyronegnetic ratio, and an estimation of relaxation tines can be obtained juat as when no splitting is present. The present study gives us the electron density at the metal nucleus as a function of temperature. we can use the data to demon« strate the presence of certain Species in the solution and to test the various models for metal-amine solutinns. By measuring the spin cone centration.as a {Unction of temperature, it might be possible to esti- mate the equilibrium constants for the conversion of paranagnetic Species to diuzxagnetic ones. Other experimental procedures available for these determinations are not so specific. Absorption spectra, cenductivity and static sus- ceptibility, for example, involve assignment of pro rties to certain species by inference or comparison with other systems such as metel~ ammonia solutions. Epr gives a less ambiguous measure of the electron density at particular nuclei than do Knight shift measurements or pro- ton magnetic resonance studies. 2 The primary purpose of this research was to study the hyper- fine splitting of solutions of Rh in methylamine. These data com- bined with optical and electrochemical data should help us deduce the species present in metal-amine solutions. II. HISTORICAL The first hyperfine structure experiments using epr techniques were reported by Penrosel, who diluted CH(NH )zCSO.)3-6H30 by sub- stitution in an isomorphic crystal of dianagnetic Mg(MH.)Z(SO,)z-6HZO. By thus reducing the electron spin¢spin interaction between copper ions, the line width was reduced so that a quartet, characteristic of the nuclear Spin of Cu (I - 3/2), was resolved from the broad line previously described for this complex. Since Penrose's experiments in 1?h9 much work has been done with epr hyperfine splitting because of the vast amount of information obtainable from such studies. Although much work has been done on the ear spectra of metal- ammonia and natal-amine solutions, as outlined in the review articlc of Symon33, the first reports of hyperfine interaction in solutions of alkali metals in amino solvents were those of V6: and Dye“ and Bar-Eli and Tuttle5 in 1963. vos and Dye reported hyperfine splitting of cesium and rubidium in nethylanine, while Bar-Eli and TUttlc showed that solu- tions of lithiun in ethylanines, potascium in ethylanine‘, and cesium in cthylanine’ exhibit hyperfine splitting. The hyperfine Splitting in lithium solutions was attributed by these authors to nitrOQan nuclei rather than lithium. III . '31:? 2"! due: the ruclzuq cf 3 paramagneiic 3§QCiai has a magmatic moaent, hyperfine itieraction may he afiqervad. If t3: magnetic field is inrge relative to tfie field cofiiributed hy tHn nuciens at [ha ciecfrnn, the first order solution 0-! the Bimiilttrnian hac*~‘?a':“t in? - 3.3- 3‘! '0 .7153 when: 113 - tatal anfiuiar momentum g - spectroaccpic splitting factor 3;, - ckctrm magnetic mzrwehf... J H - ragaetic field 1‘31 I- me 1311' riffmtic (192.211:be mm‘xer and each ma;natic energy iavul will be split into 21 + 1 cowponznts (where I is the value of the nuclear spin in multiples of h’én) with equal spacing Jig. At least four canditiana must be satisfied for the observation of hyperfine atructurc’. These are: 1. The systcm must be sufficientiy dilute so that the hyperfinn interaction (ticctran-nuciéus nagfietic interactions) is 1ar~ get than either the exchange interaction or the dipole—dipole interacticn. 2. If an: reprssenzs the splitting interval in frequency uniis, the spaciaa' lifetime and the tharmal rsiaxation time T1 muat be greater than 1.4.30». 3. The microwave field axplitude must be small canpared to the splitting interval when bc.h are expres ed in gangs. S b. The field inhnmcgcneity over the sample should be smaller than the splitting interval. Splitting intcrvals of less than 130 nilligauas have been nbscrvcd in compounds containing many protnnm while for free hydrogen atoms, the splitting wcnld he in the neighbnrhoud sf 530 931.133. Since a prime detector is used for cbéervatinn cf the abnorptions derivative, the field modulation anplitudc must be much less than the splitting interval. Th. (21 + l) hyperfine lines will he of nearly equal intensity since the pocsiblc orientations of the nuclear spin vector are almost equally probable. However, when the electron moves in a dclocalizcd orbital and comes undcr the influence of scvural nuclear magnetic momentg, theta are at least two possibilitiea for the structure. when twc nnno equivalent nuclei exhibit interaction with the unpaired electron and one has a much stronger interaction, then the stronger interaction produces (21, + 1) well spaced lines, while thc other nuclcun splits each of these lines into 212 + l) canponents. When two equivalent nuclei interact equally with the electron, the effectiva nuclear spin is 21 and thus (LI 0 1) equally Spaced lines occur. Similarly for any n equivalent nuclei (ZnI + l) linen occur. Howevar, these lines are not equally intense, the intensity being directly propartional to the num- .ber of different combinations_of the Ni values which result in the elec~ tron spin magnetic mnment receiving an identical field. This situation would arise if, as proposed by By: and Dcmald10, an ”2+ ion were one of the paramagnetic species in netal-nmina solu- tions. For example, the Liz+ inn should give (h-3/3 + 1 ' ?) lines with relative intensities l—2~3—h-3-2-l ascuming just the Li7 isctnpe 6 (92% abunonnce). Four nitrogen nuclei, to which Ear—Eli and Tuttle attribute their observed hyperfine Splitting in lithium~ethylaminc solu- tions, would give (8-1 + l) - 9 lines with relative intensities l-h-lO- 16-19-16-10~b~1. Figure l is a diagram of the hyperfine splitting of energy levels at constant magnetic field for a single nucleus with a spin of 1. How- ever, because it is difficult to obtain a linearly variable frequency aource in the microwave region, it is the usual practico in epr work to uso constant frequency and vary the magnetic field. with this in mind, we may illustrate the formation of hyperfine lines as in Figure 2. On. observes the expected number of lines in succession during an.in- crease or decrease in magnetic field as the separation of the two levels becomes equal to h . Also, as can be seen from the diagram, the separation in gang: must be the same for the vnrious levels because the separation of the energy of the various levels is the same. .,In contrast to the simple picture given above, hyperfine lines are sometimes found to have unequal splitting. There are at least three possible ¢Xplanations of this behavior: 1. If the external magnetic field is comparable to the magnetic field of the nucleus and not much larger as asouned above, the conpo- nents of the nuclear angular moment in the direction of the field are no longer perfectly defined and an unequal splitting of the levels results. 2. There may be an electrostatic interaction between the grad- ient of the crystalline electric field at the nucleus and tho nuclear quadrupole. Nuclear quadrupole moments exist when the Spin quantum +213. ,r' I f /‘ \s¥ A‘V’nr‘i‘g 6f // Trar‘m "7-01:. if A m 9 f. d \ C‘V‘ ‘ quA" ...3gnet1c 19.1. no Q, *tUML v \ ': .2 \— T .. \ + I-’OF‘ +1 number is greater than 1/2. .Such a nucleus has an asymmatry of electric charge. From observatiafi of this unequal hyperfine splitting, the quad» rapole interaction can be determined. This effect may alqo result in variation of line-width for the individual hyperfine components. 3. Non-equidistant lines may alga appear if two isotopes arc preaent in comparable amounts, as the case of rubidium. In this ca+e there are two spectra saperimpcsed and the resulting lines need not be equally spaced. The shape of lines reiniting from exchange interaction depends upon whether the exchange is between similar or disaimilar ions. With similar ions, the exchange effect narrows the lines at the center and broadens then in the wings. With dissimilar lynx, that is, iona having different Larmor frequencies, thr line will breaden provided the com— ponent lines are sufficiently close. Exchange effects with dissimilar ions are most commonly obsarved where the sums ion occura in different magnetic sites in the same unit cell. The width of n resonance line can givs fundamental information about the various interactions. The spin-«pin relaxation time; T2; may be calculated from the line width if the curve is Lorentzian. The spin- spin interaction depends on the reciprocal at the cube of the distance bétueen individual paramagnetic species. Dianagnctic dilution of the sample is urefui in increasing Tz, hug reducing 11nd uidt.. This might alga be accomplished.by increasing the spin—lattice relaxation time, Tl, by lowering the temperature. The width of the resonance line is influenced not only by relaxa- tion times but also by unresolved hyperfine splittinga and by local 'fialds from other species if they are sufficiently close. The lifetifle 1‘ x“, of tag paéaaagnetic 3935123 if 321’1c‘cn11g sntrt, can trun.»3 the resanancc through the opzratlen of tag Naiscrhsrg Uncertainty Princfipla. Resonance 11133 aid: '19 are vary 38313‘2113 3% tr?» 3.! . sit y“ 31' 13 3.2313331"? reports that thzre I! an optimut viscg3ity for the chsarvation cf hyper- fine structure. . there hypfarfim strut? 3m: «3', "we a 23m: (31‘ (we ciwwawet .3}: 23:2. tra is datum-3133M} by that {migrate of I‘ERfl‘o‘lij‘t-iaf‘a 323232.61, £33 turn, is governed by the ratio 9f the width of the 333 -r11na csvw~mrnts ta t2 3 2 3>rfiaz separation. The two extrena 63323 Ufifitd be distinct aparzte lines and a single broad line. The 0231;; term in the Eiareiltanirm far {:33 hatemction 0f the mclma with an electron to give hyperfine flplittivg which is of {apartuncz in our salutions is the Fetfl! cantact tarm. The Diners are ania: irfipic and are avuraged out in solutians.‘3 53 . 2 ,-\#~ :rb . 2 C a an- n i" ‘ '. ‘ 2 O _“ ‘ _ X his 3 ”5‘ 233151 J“ "' J' A f ' when: - electronic apectrvscopic splitting factor fi - Baht magneton g1 - nuclear spactroecnpic splitting fatter 531 ' We???“ 333-33222333 x 1Q T Kronecker delta 1535p- 1/21? {F+1)~I(I+1)~S(‘3 +1) 1...! s - . 1/2 .2:- - 1-12 mu 10 In the presence of a lax-3c extemzal mgrsetic field there is m I-‘i‘a cwpllng. Tm selection rule-s an 33-1 - O, aft, . *1 {at ahmrptim; Thus: 2.3 I 95:3!“ '0 PIA far transition 3*." - ~1~"2 m Fig - 9 1/2 since thz mclear nmmrco Inquiry < 4 elactrrm mmmmc fragwmncy 33 - 31% when: Pi . ~I, -I + 1, -I o 2, to *I ‘31:: 3 “I 91": " T ft: I’IIQT 3 ' ”I“ 7: ' T" F‘. We») total sfifilt'atlsn between $1 - --I to 3’1 - #1 13 F? g. 2 ” M 4. am) - 221(- -§- ' T a “ PM) 1.8!: total scysratian is than 4 L' 15” l an. 1‘ I “(3.3 . "' “'3‘ 41:15:, we?) 9 1 me Afar . 3 .fi{ ..&A;;Agrn5 gamma gsg’ ” . 15w 3 “Dion-31 ' - 3H1 wfo) The derivation followed 19 that suggested by Fingers.” a Cl: 1‘0. rt (1 it IV V””"’*-‘" :11. I ~18!» “\‘-\—kA Lot! A. glgzzgnqg_Clen§ing Th3 solvent purification train and the sample make-up vessel were cleaned with a hydrofluoric acid cleaner prepared according to the following formula: 2% acid stable detergent (Tide or equivalent) I33fi concentrated nitric acid 5% hydrofluoric acid 63% dixtilled water. The glassware was soaked in this golution for a few minutes, then rinsed in distilled water three or four times. The solvent purifica- tion apparatus was then washed with aqua regia, rineed with distilled watar about six times, demineralized water which had been radistilled in Pyrex, an additional three tIMBfi, and dried in an oven at 1LCOC. This procedure was considered inadequate for the sample makenup ‘vessel because of the difficulty in filling the 3 mm tubes. These tubes had to be alternately heated and cooled to be filled with liquid. and emptied. After soaking in the HF cleaner for a few minutes, the vessel was rinsed three times with distilled water, then filled with hot aqua regia to snak for at least four hours. It was rinsed three ‘ times more with distilled water and soaked in the dedineralized water for at least four hours. After this it wad rinsed at least six times ‘Hith demineralizdd water and four times with the redistilled water. Finally the veg—tel was dried in an 1.54308 raven. ll 12 w thout the aqua regia ooaking tho salpleo éacoqposod rapidl'. Possibly 530m catalytic mteriol ind been precipitated when the HF cleaner was heated to drive the liquid from the tubes. This may not have dissolved in tho ensuing'wator rinses but probably dissolved in the hot aqua regia. - Tap water was never uqad in the washing proccéure and only roa- gent grade acids were used to make up the cleaning solutions. 8. Mini: Purl f! mti on The mono-ethylanine was Eastman 586~X grade in 103 gm breakutlp bottles. It was purified in.an.apparatua as illustrated in Figure 3. Powdered BaO and cut piecao of sodiun were put into the pot to dry the amine. The bottle of ethylaninz was cooled to 0°C, the tip broken, and the contents of the bottle poured into the still-pot which was at dry-ica—alcohol temperature. Nitrogen, which was purified by passing over Cu.and C30 at LOCOS and through silica gel at'dry-icaualcohol temperatures, was bubbled through the cold liquid (at dry-ice tempera- ' cure) and the sample was refluxed. A polyethylene cup filled with powdered any-ice was fitted around the column to aid the cold finger (filled with a dny-lce-alcohol mixture) during refluxing. Eitrogan was bubbled through the sample placed in a dry-ice bath for 15 minutes, . in a 0°C bath for 1/? hour and exposed to room temporaturea for four hours. In.all; about 30 ml 333 distilled into the trap during reflux. when the refluxing was completod, the sample was warned, frozen, and degassed repeatedly until the pressure above the frozen sauple stabiliznd at less than 1 x lOM torr. At this preosure the sample was distilled. The first 15 ml: were condensed in the waste vessel and the remaining 13 was distilled onto a potassium film in the sample flask. The sample was kept cold enough to dietill at the rate of about one drop per sec— ond from the cold finger. The distilled sawple was then warmed, frozen, and degassed until the pressure etabilized at l x 10“6 torr above the frozen sample. It was then distilled lEMXESES onto another potassium film in a storage flask. Formation of the potassium solu- tion in the ethylanine was very slow. At first only a slight blue coloration was visible but after two or three days at dry-ice tempera- tures a dark blue stable solution formed. The mono-methylamine was obtained in a bomb from Enthieson Company, Inc. The purification in this case was similar to that of ethylamine. After about 150 ml: had been condensed in the pot at dry-ice temperature, nitrOQen was bubbled through the system for five hours. The methyl- amine, over Na metal, was then refluxed in an atmosphere of nitrogen at near atmospheric pressure for four hours. In these steps about 75 mls of the liquid was lost. The amine was then frozen and degassed and distilled in 33532 onto a potassium film. After standing in this vessel for 2h hours, the amine was warned, frozen and degassed repeat- edly until the pressure stabilized at 2 x 10"6 torr. It was then dis- tilled onto another potassium film in the otorage veesel. A: with ethylamine, potassium'was very slow to dissolve in the ‘purified methylamine. After much agitation a stable dark blue solution formed which had a more intenoe coloration than saturated solutions in ethylamine. 1h Cola Finger ’ to trap , and hood .4. - 00 V TLC QUE". TIO / nitrogen line \ receiving p\ j ' ves sel If! I Fig. 3. Amine Purification -rain 15 C. Sample Preparation The sample make-up vcscel used for lithium solutions is shown in Figure h. Lithiun.was obtained from the Lithium Corporation of America in the form of half inch rode stored under mineral oil. After cleen« ing in purified benzene, a small piece was cut from the center of a rod ueing a dry—box filled with argon. The small piece was put into the sample veseel while still in the dry#hox, and capped with I QL/LO standard joint using Apiezon T as lubricant. The vessel was then taken out, attached to the high-vacuum line 113 the 12/5 ball joint using Apiezon W wax, so that it could be evacuated immediately. Using this method, a bright piece of lithium could be obtained in the evacuated flask which showed no noticeable surface discoloration. The amine was then diotilled onto the lithium by cooling the flask with liquid nitrOQen. The stopcock was closed and the vessel was removed from the vacuum line and agitated to facilitate solution and the solution was poured into the sample tubes. The tubes had to be cooled slightly so that the vapor prescure of the amine would push the solution into the tubes. The height of liquid in each tube was measured and more amine was then distilled into the tubes from the other pert of the flask. The amine was frozen and the tubes were sealed and pulled off. By this procedure a different concentration could be Obtained in each ’tube and, while the absolute concentration wag unknown, relative con- centrations were known by the amount of dilution in each tube. For cesium and rubidium, which could be distilled in Pyrex, the sample make-up vessel was as shown in Figure 5. A tube of purified metal (taken from stock prepared as described by V031“) wa3 placed in 21 no ' lZ/S ‘/ \/ u of thcce tubes Fig. h. Lithium Sample Nakeup Vessel 17 6 Of these tubes J T31 0 CD 0 7 inches long *Fig. 5. Rnbidium Sample Makeup Vessel ~J 13 dry-ice to cool the natal. The top was broken off and the tube put into the side~arm on the sample oaks-up vugoei. The side-arm was capped and the vesrel immediately evacuated. After warming aad de- gassing until pressures of loom than 1 x 10'"5 torr were reached, the metal was melted and aliens-red to flow through the constriction. the metal oxide remaining on the walls. The tube was then Scaled off at the constriction and the top part pulled off. It was then necessary to warm and degas the metal again until a presoure of lens than 2 x 10-6 torr was reached. The metal was then distilled 13 31533 to form a mirror in the main part of the make-up vessol. The side-arm was sealed and pulled off so that thio part of the flask could be im— mersed in liquid nitrogen to enable the amino to be distilled onto the metal mirror. The following stopo were as described for lithium ex- cept that more agitation was required for cesium and rubidium than with lithium, in order to effect solution of these two metals by the amines. Also a solution concentrated enough for our work was oaoily obtained without saturation with lithium but even the essentially saturated solu- tions for the other metals lacked the desired sigoal strength but were apparently still concentrated enough to wash out the hyperfine split- ting. An attempt to remedy this was made using sample tubes of 5 mm outside diameter with a less concentrated solution. However, this was unsucceaoful because the dielectric losses of the extra glass and amine proved great enough to prevent the tuning of the hlystron to the cavity. In some cases it was necessary to dilute the solutions in the epr sampla tubes to obtain the splitting. As those tubeo were sealed under vacuum.and he contents would decompoae very rapidly if opened to the 1? air, dilution could not be accomplished by the usual means. Inetead, since the tubes were usually less than half full, it was passiblc to distill the solvent from one era cf tfie take t3 the other lzaving some of the metal deposited on the walls and giving a less concentrated solu- tlon. 'D. KPH Spectgg The epr spectrometer used was a varian Model LSOC x-Band Spectrom- eter with 100 K. C. field moéulation. The measurement of the magnetic field strength was done as described by V091“, by counting the frequency at which proton absorption occurs using a Hewlett~Packard Hadel SZhC Electronic Frequenqy Counter. Thia makes use of the relation3hip be» tween frequency and magnetic field at resonance: 2:! vp II {’11}; where: ‘5p - frequency in cycles/kzc (- t 1 Arif' lr‘. ‘1 -1 7p c (2.bffi23 i C.OOUVQ) x 1C 35c gwuss I (II ‘ a - "’ 4 H - magnetic field in gangs - 2.33631 x 13 '(p The nmr sample used was a 3 cm.length of 16 mm Pyrex tubing filled with fihaeffer's #316 black ink. This sample was fagtaned.betuecn the magnet's poles faces to the out31&3 of thw cavity and parallel to the apt sample. The magnetic field at the apt abaorption naximum for a sample was then determined from two prcton markers, one on each gida of the apt absorption. Th: nyparfine splitting can also be determined in this way. The method used 13 illustrated in Figure 6. .- ' "" I d awn. his... 20 .WO.,._. 0 n. n. (w , .0 fi.. . . Aumy#r._J.LHv-dfsiHumfl I wmdLfiw.HomCflfl C 70 o . w I"... OIIODI'I '1.-.‘Llub.'|.¢rll I I l .I Y - D ‘ O. ' II ‘l.l‘u.'..|1i “H t my. age Ur ..V_ mp ._,....OV I {\‘l .4.- M...\? 6.10% .... ro m._.a u 59:4: mcwq .y4. I a 4 «I. J ... .. 5-. u. .1! .A. .u 17.4... .1. . ‘.4 \.. . MOP)? t (F .f..n..fll“fifl PVC .Hmfinohf m:.:HN.O msamp a dam... Ho :.0 . C 4.44.”..Ju rv ..'.rs.. w magma Deon—u. m. 9 .n«1 :H 00mm fl.Nmflfiu a? o. of Nao:..:d % nw_0: .NH .1 ~ poxpwfl cepcpw AN .._......_..-_L P CO moNH ‘..\r_‘.......... 02 momc :ozmzcchp coapwmam h‘ - ~ ‘ r . .9 n 1v .I- ' r “ tam «w '7 v- " I . ' -v ' " .-- U To determine: (~le g Viz“. 1m: are: mm} {.0 man... tne 13.1 yanrom f" £333.13:ng as wall as the magnetic field. This frcrzwnqy could E~e d..tr .iseJ from toe micrometer r23-i:; on a naveiine a}? wavanetar which had pre» viousxy been callkratsd in terms af Frzruc IJ. The pracicion wag limited by the rggfiing of the micrameter. ?fia accuracy in kiystron frequency uaq at least i 1 ms. Apparatus HES 8120 availahia far beatirg tile freann y oft eklystron again.m;ttE13 stawflird frequency EH ofa ”"icrounaw" Vodel 1C1 Eicrovavn Frequency Calibratmr and ccunting the frequency with the He.lott-Pat. rd ceuifl er after auplification. 1 E4 l-t This procedure can give accuraqy in determining tne klys tron frequency ;J to t 0.02 ms, but it wag félt that the errer in data mining the center of the splitting was great enough ta make this meaxurawznt unnecessary. E. .?fintrature {vfarfiinItin Low temperatures were obtained by p aging a stream of dry nitro- gen gas through a liquid nitrogen bath and into the dewar holding the sample tube in the cavity to get the desired temperature. with tha proper flaw rates, a reaso..a:1y stea:.iy temnerature could be obtained. A stream of nitrogen.at room temperature flowed betuean the outside of the deuar and the inside walls of the cavity to prevent water con- densation. It was desired t0 maintain the temperature to within 1 1° of the starting tempe1%tut3 and temperature st:E ilities of tiis order were obtained. It is easier to obtain stability uaing a dry-ice- acetona bath in place of the liquid nitrogen, butcf ccursz the range of tamparatures possible is much smaller. A heater was incorparated in the flow Ii: 3 of the cold ni trogen but thiq was only needed far temperatures above -ICGC, loser tax nattxres being obtained merel3 by 22 odjuoting the flux rate: of the two streams of nitrogen. orpcraturcs were moo: “ed with a coppcrwconotantan thermocouple and a Honeywell 2?L6 potantionctcr. At first the thermocouple was. placed ncar the bottom of the dcwar holding the sample tuhe in the cavity so that it would not interfere with the microwave field; how» evcr this proved unsatisfactory becauoe of a teopcrature differential between the thermocouple and the sample tube. Aloe, tho thermocouplo apocarcd to stabilize before the temperature at the sample became stable enough for operation. There is a null plane in the cavity at which the field is so low that absorption by the thermocouple wire causes no difficulty. This plan: is easily found by fastening the thermocouple to the sample tUhc, placing it in the cavity, and rotating the tube until the klystron can be easily tuned (tuning dip at a maximum). Placing he tip of the thermocouple next to the sample in the cantor of tho cavity docs not interfere then with the signal and gives better temperature valuoa than are determined with the earlier procedure. The optical absorption spectrum of one of the solutions of rubid* ium lo methylaminz was qualitatively measured using the 3 mm sample tubes instead of special cells. A piece of black cardboard hold the sample in the cell holder of the Bockmon DK«2 spectrophotometer with black tape covering the edges so that light passed only hrough the solution. No reference cell was u1 be J: .4 Camdmfiw .L/ 4,:vgraumamxrhw .rvfi 0% mprpmpmgzop Econ pm wrcc m .Hmpmm mgomon .959 .I/ (u q p Q .Hgo ezHflzwlc;_ 1 L. 2C :C .w 5‘ _ , r . K . f .. 32 O ~125 C. T1 7 ‘ A ‘15. 12. RubidiIm-Yethylaztine :4- :i Speczra VI. DISCUSSICN A. Lithium in Singinring Since the broad peak was not visible in the purer solutions of lithium in cthylamine, it was decided to determine whether this broad peak could be intensified relative to the narrow one and perhaps rc- solvcd into hyperfine structure by using impure cthylaminc. It was believod that this absorption could be due to some minor species which obtained suificicnt concentration relative to other paramagnetic species only in rapidly decomposing solutions. A possible species is the Li2* ion, which if present should give seven hyperfine lines (neglecting tho contribution of Li‘) with relative intensities 1-2~3~ h-3n2~l. Bar-Eli and Tuttlc report nine hyperfine lines which they believe to be due to splitting by four N14 nuclei using a modified form of the Becker-lindquist-Aldcr monomer model for metal-nminc solu~ tions. That a rapidly decomposed sample still gave a resonance sig- nal after couplets disappearance of the blue color would indicate that the paramagnetic species responsible has been included in the precipi— tate, possibly as a salt of the ion. Since the narrow line and the broad line decay at different rates, at least two paramagnetic species are indicated. Bar-Eli and Tuttlc also report the presence or two species and attribute the narrow line to tho solvated electron. B. lfiubidium in Ecthylnnine V0315 did not detect the two outside peaks in the hyperfine struc- ture of rubidium in methylnninc'uhich.arc clearly visible in our 33 3h spectra. Since the Rb”? is split twice as much as the Rb55, it is reasonable to asoune that the outside peaks of the Rh57 should be sep- arate from the six-line Rb55 spectrum and eaoily resolved from it if the six lines themselves are resolved. Indeed, in certain spectra, only the too outside lines (distinouishnble by their greater width) are visible. the rest being masked out by the single center peak. Since the separation of the two outside peaks has the same temperature depend- ence as the total separation of the six inner peaks, the two sets of peaks appear to involve only an isotope effect. Assuming these two outside peaks to be due to Rb57, we can calcu- late from measurements on.only these two.the position of the Rb57 peaks, the position of the Rb”! peaks, the center of the spectrum, the line width of the Rb°7 peaks, the relative intensity of the Rb57 absorption, and the total area under the Rb55 integrated peeks, since we know that the ratio of the total peak areas must be in the ratio of isotopic abundances, and that the ratio of the total separations is given by the ratio of the nuclear magnetic moments. Using these calculated data, an attempt was made to reconstruct graphically, assuming Lorentzion line shape, the observed room tempera- ture spectrum. It became apparent that at least one other line was also present. The peak positions were reproduced quite well, but the relative heights were not. We can assume that the brood single line is also present at room temperature and should be included as an extra line. With this extra line the relative heights are much better. The data indicate that at low-temperatures the line width of the center line in Rb-nethylamine solutions follows the behavior of the center line in Cs-methylamine solutions on reported by Vos. If this ‘Y'. If). ‘ 35 agreement extends to higher temperatures, the line width at the temper~ ature of curve-fitting (36°C) should be too broad to detect with the other lines present. For this reason we believe that the line widths of the center lines for the two solutions are not the same at these higher temperatures. 3e cannot determine the line width in Rb solu- tions directly at these temperatures to see if this is correct. Using the formula for the electron density as a function of the total splitting previously derived, we get a value for 03 of 0.165210-3 which is 7.0% of the value for the free atom at 25°C. This is twice the value obtained by Vos since he did not detect the two outer peaks, but rather attributed the two outer peaks of the six-line.spectra to Rb57 rather than to Hbas. The hyperfine separation seems to be a linear function of tempera- ture in the region investigated. However, it would.probah1y not follow the linear relationship at lower temperatures since Vbs's curve of cesium in methylamine also appears linear in this region but not at the lower temperatures. The optical spectra obtained correspond fairly well with the absorptions reported by Dewaldl6 fOr solutions of rubidium in ethyl‘ enediamine. Although the relative intensities of the various peaks are different in the two solvents, it indicates that the same types of absorbing species are present. There are at least three possible explanations for the dependence of the hyperfine separation on temperature. Bar-Eli and Thttle attrib- ute this to antactual increase in the size of the framework over which the electron is spread as the temperature is decreased. This would re- sult in a lower electron density at the nucleus at low temperatures. 36 This theory requires drastic changes in electron distribution with temperature. Perhaps a better explanation would involve the admixture of several s-type wave functions with the relative contributions vary- ing with temperature. For example, the 23 nave function has a radial node which is not present in the ls wave function. Therefore, for the wave functions to add constructively at large values of r, the 23 must have an opposite sign to the is at the nucleus, thereby decreasing the density at the nucleus without a large effect on the electron distribu- tion. This explanation would be valid for higher 3 orbitals as well because each succeeding orbital has one more node than the previous one. A third possible explanation would attribute the hyperfine split- ting to dissolved atoms whose population relative to other paramagnetic centers would be strongly temperature dependent. To observe variation of the coupling constant with temperature rather than merely total collapse of the hyperfine structure would require that electron ex- change be very rapid compared with the spin-spin relaxation time. For each of these cases, possible explanations for only a very slight dependence on temperature of the splitting in the frozen sample might be either that the lattice binds the species more tightly in position, thus allowing little change in orbitals or that the relative concentration of the species giving hyperfine splitting is independent of temperature in the solid. As was indicated, the splitting decreases as the temperature is lowered until the freezing point of the solution is reached. The separation then increases to approximately that at room temperature and remains nearly constant as the temperature is lowered. A similar 37 type of phenomenon was observed by Bar~£11 and Tuttlea for lithium in cthylamine. In this case there was an Increase in line width as the temperature decreaaed down to the freezing point of the ethylamine, when the line width suddenly decreased. They offer no explanation for this and at present we have no adequate explanation for the sudden change in electron density at the nucleus when the solution freezes. VII. FUTURE WORK If some Rh“? could be obtained and an epr spectrum in methylmaine solution were determined, it could he proved conclusively that the two outside peaks were due to this isotope. However, due to the difficulty in separation of these isotopes, it is not anticipated that this will be done. This additional step is probably unnecessary since the spec- trum can be so nearly duplicated by calculated values. In fact it is felt that the most desirable work for the rubidium-methylaminc system would be to find the Spin concentrations as a {Unction of temperature which would give an estimate of the effect of temperature on theequi- librium between the species giving the hyperfine splitting and that giving the extra line. Some work planned for the future involves trying to reproduce the spectra of lithium in ethylamino reported by Bar-Eli and Tuttle ueing L15 isotOpe. If the hyperfine lines reported were actually due to splitting by Liz+, only five lines would appear for L15 instead of seven for L17 (92% natural abundance). If Bar~Eli and Tuttlo are correct in their report of nine lines and assignment of these nine lines to the four nitrogens, then substitution of the Li5 isotope should produce no change in the pattern. Attempts should also be made to study in more detail the ear spec- trum of the decomposition products of a solution of lithium in ethyl- amine. The precipitate which forms in a very rapidly decomposing solu- tion should be separated from the liquid to determine which phase in responsible for the epr signal. 3:3 39 New solvents could also be used in attempts to observe splitting due to the Hz+ ion. ‘ “-5- ‘t VIII. REFEREEZCE 1. Penrose, R. P., Nature léi: 992 (19W). 2. BER 'i‘rlble, Varian Asaociates, Palo Alto, California. 3. Symons, E. c. P", Quart. Revs. (London) 13, 99 (195:? . h. Vos, K. D. and 0,72, J. 1..., J. Chem. Phys. 3:, 2033 (1962). 5. Bar-Eli, x. and Tuttle, r. R., Bun. Am. Phys. Soc. 33, 352 (1963). 6. Bar-Eli, K.and Tattle, T. R., to be published. 7. Bar-Ell, K. and Tuttle, T. R., private commnlcatlon. 8. Bleanzy, 8., Phil. fihg. 1.3, M41 (1951). 9. Fake, G. 5., Heisman, S. I., and Tcmscnd, J., Disc. Far. Soc. 33,, 12.7 (1955). 10. Dzmld, R. R. and Ely-"e, J. L., J. Chem. Phys., to be publixhed. 11. Il'yasou, A. V., J. Struct. Chm. (English trawl.) 3, 81; (1962). 12. Hausser, V. K. H., Zeitschrift far Electrochemie fig, 636 (1961). 13. Fake, 6. 3., Pammggetic Resonance, w. A. Benjamin Inc., New York (1952). 1h. Regers, M. 1., privata comnication. 15. V03, K. D., Ph.D. Dissertation, Michigan State University, East Lansing (1953). 16. Dew-aid, R. R., Ph.D. Diascrtation, BZicixi=_fian State University, East Lansing (1963). Lo km: E!- GH'LMS‘L-J upward