_ . 1.: u- 5i...“ 495$.“ (‘1. I :M 1.] 01.1.1; V0.31”! fiat. ‘. VAC: u. .1rx11: .\._-i..1 .u v' c - .II .0. - wfiK’M‘Lot Ito! . . . L...- . u-n..uwu~...¢»u.zn.t-.::...8w .. . .. 7” . . . . . ._..%.u..r..n..u.-u.u.m.«.si1. o1. 4.. :- 4 7~.v .ll.n_¢£u 3%.... . $5- M?” L.. -H... w... adv-not . 1.161... n-walrbi < ....../A. H .a..: ‘ D O I 1 . «flank-napav .. . ,v I _ . . .. Kama . . o . I . n w H” .c .rc. . . . n. . 'v-HWM.-fl . Mv . In...“ . u .. . .. Swam- .Y. .Nra . . .. .23 I .IA.¢. .vu...v.aLIoumuv.-uu n».£.......n.. . $23. . 1 . v “I‘ll“. . ’1‘. . .2. an}. \. .- 1: .1 --- j ". II ‘-‘I ll" I‘lv‘- ..‘.-.v..H-.uI.-I--..-.- I}. . . O.‘.-....‘\Ib:."0wt I31 9’ ZKVKMCJAPWVFN .l livalr.» .VLI v «r... 1.. .... .. w a.\..-V. . g .Ioll’u- .54 ualdv I‘c}. , Y. tux-«470:; . LAD-«.4 .. v ‘ 3H,. $3.1“.- l 1.3.48 {.h‘clllr.‘ 3 n 1;!“ . < n .304..- dnh: \- WW ,"Hyh‘: T'- fr' '1 {H ‘| n' . ‘ .1. 2v 1: :3.» .. [tic-Nun 1..»- ....l.b1...9tl1 .4 . “HI-9.- . 4)... . . - .. . l! i I. - -. Q. m“:"’ DI .u -t'n‘u¢u.\.|.y-u. .xlv -|l|l¢s«;-a-s?b\.!l1. w .v‘ . 1|. (Kl->0. Ht! it‘lzwptl -. .u 'llll'. - .. ... v ..- 1-.. iwmmmzi... .- “-7 . {:1 2h, . 39?...3-JI. $493.31... 3- -.t . . I foo-$2- . .. b LIBRARY Michigan State University , This is to certify that the thesis entitled ESR STUDIES OF ALICYCLIC RADICALS IN ADAMANTANE presented by JOHN CHOO WANG SONG has been accepted towards fulfillment of the requirements for Ph: 0. degree in W ‘7762817’fi/624 Major ofe set ‘5‘ Date 9th September, 1977 0-7 639 ESR STUDIES OF ALICYCLIC RADICALS IN ADAMANTANE By John Choo Wang Song A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1977 n ABSTRACT ESR STUDIES OF ALICYCLIC RADICALS IN ADAMANTANE By John Choo Wang Song Solid solutions of a number of alicyclic compounds have been irradiated in solid solution in adamantane. The ESR spectra of the cyclic radicals produced are mostly isotropic and have been studied over as wide a temperature range as possible. These have been analyzed and the ESR parameters interpreted and compared with other information in the literature. The electronic and geometrical structures of the radicals have been deduced, as far as possible, from the nuclear hyperfine splittings and g values. Radicals studied include hydrocarbon radicals (cyclopentyl, cyclopentenyl), substituted hydrocarbon radicals (l-hydroxycyclopentyl, l-cyclo- butanecarboxylic acid, l-cyclopentanecarboxylic acid, l-cyclo- hexanecarboxylic acid, l-cycloheptanecarboxylic acid) and cyclic ketone radicals (2-cyclopentanony1, 2-cycloheptanonyl). Some heterocyclic oxygen and sulfur radicals (Z-tetrahydrofuranyl, 2-tetrahydrothiopheny1, Z-tetrahydropyranyl, 2-tetrahydrothiopyrany1, 1,4-dioxanyl) and nitrogen radicals (l—pyrrolidinyl, 2-piperidinyl) have also been investigated. Some radicals were also found to give isotrOpic ESR spectra in the pure crystalline precursor and were inCIUded in this work (l-cyclohexanecarboxylic acid radical in John Choo Wang Song cyclohexanecarboxylic acid, l-cyclopentanecarboxylic acid radical in cyclopentaneCarboxylic acid, l-cyclobutanecarboxylic acid radical in cyclobutanecarboxylic acid). Ring inversion in cyclic radicals can at present he studied ~on1y by ESR and one objective of this work was to obtain the barriers restricting ring inversion in cyclic radicals. Temperature dependent ESR spectra were observed for several of the radicals and values of 'Ea’ the energy barrier restricting inversion, were evaluated for 2-tetrahydrofUranyl, 2-tetrahydrothiopheny1, l-pyrrolidinyl, 2-tetra- hydropyranyl and 1,4-dioxanyl, among the heterocyclic radicals, and also for l-cyclopentanecarboxylic acid radical both in adamantane and in the self matrix, for l-cyclohexanecarboxylic acid radical in the self matrix, and for 2-cycloheptanony1 radical in adamantane. The values have been compared with values reported in the literature for radicals and with the barriers to inversion in comparable undamaged cyclic molecules as determined from NMR. The effects on ring in- version of introducing one and two sp2 centers into the rings have been determined. A computer program has been written for determining nuclear exchange rates from nuclear hyperfine splitting parameters and line- widths. The program requires the experimental relative intensities of second-derivative Gaussian lines within the region of temperature where the ring inversion rates are comparable to the inverse of the nuclear hyperfine splittings. TO My Mother, who has been critically ill during the course of this work ii ACKNOWLEDGEMENTS The author wishes to express his sincere appreciation to Professor M.T. Rogers for allowing the author freedom to pursue this subject, and for his infinite patience and guidance. The author wishes to thank sincerely Dr. R.V. Lloyd, without whose help this work could not have begun. Thanks are due to Dr. William Waller for his assistance in mathematics and computer programming. The author also wishes to extend thanks to Dr. J.Y. Lee for dis- cussion, Dr. Henry Luke for his assistance with the X—ray unit, and Kathy and Mark Uebersax in Food Science, Michigan State Univer- sity, for their assistance with v~irradiation. Mr. Wayne Burkhardt has been instrumental in keeping the ESR spectrometer operating throughout the course of this research. iii DEDICATION TABLE OF CONTENTS ACKNOWLEDGEMENTS LIST OF TABL ES LIST OF FIGURES LIST OF STRU INTRODUCTION HISTORICAL S I. II. THEORY III. EXPERIMENTAL I. II. III. IV. V. CTURAL FORMULAE URVEY Cyclic Organic Radicals Rates of Conformational Change from ESR Spectroscopy Analysis of ESR Spectra Interpretation of Hyperfine Interaction Constants (a) a—Proton Hyperfine Interactions (b) B-Proton Hyperfine Interactions (c) y-Protoanyperfine Splittings (d) Fluorine Hyperfine Interactions (e) Chlorine Hyperfine Interactions (f) Carbon-l3 Hyperfine Interactions (g) Nitrogen-l4 Hyperfine Interactions Exchange Rates from ESR Linewidths Instrumentation Chemicals Sample Preparation and Handling Irradiation Analysis of Data iv Page ii iii vii ix xiii 16 16 21 21 27 3O 3O 32 34 36 39 42 42 43 44 46 46 RESULTS II. III. IV. VI. DISCUSSION I. II. III. IV. V. VI. SUMMARY REFERENCES Cyclic Carboxylic Acids Cyclopentanecarboxylic Acid in Adamantane Cyclopentanecarboxylic Acid Cyclohexanecarboxylic Acid Cyclobutanecarboxylic Acid Cycloheptanecarboxylic Acid in Adamantane . Cyclopropanecarboxylic Acid in Adamantane yclic Ketones Cyclopentanone in Adamantane Cycloheptanone in Adamantane eterocyclic Oxygen Compounds Tetrahydrofuran in Adamantane Tetrahydropyran in Adamantane Dioxane in Adamantane eterocyclic Sulfur Compounds Tetrahydrothiophene in Adamantane Pentamethylene Sulfide in Adamantane eterocyclic Nitrogen Compounds Pyrrolidine in Adamantane Piperidine in Adamantane ther Alicyclic Compounds Cyclopentyl Chloride in Adamantane Cyclopentyl Bromide in Adamantane Cyclopentane in Adamantane Cyclopentanol in Adamantane l,2—Dichlorotetraf1uorocyclobutene in Adamantane monw>ow>zw>znw>zw>nmmcnw> Some Properties of the Adamantane Matrix ESR Parameters for Cyclic Radicals Alicyclic Carboxylic Acid Radicals Cyclic Ketone Radicals Heteroalicyclic Radicals (a) Five—membered Rings (b) Six-membered Rings Other Radicals (a) Cyclopentenyl and Cyclopentyl Radicals (b) l—Hydroxy-l—cyclopentyl Radical (c) The Radical from 1,2—Dichlorotetra- fluorocyclobutene Page 48 48 48 S8 61 67 67 71 71 71 73 79 79 83 87 92 92 99 99 99 102 107 107 109 109 109 109 116 116 118 132 137 139 139 143 145 145 147 147 149 151 Table 10 11 12 LIST OF TABLES Representative a— and B-proton hyperfine inter— actions for aliphatic radicals in the pure liquid phase or in solution Representative o- and B-proton hyperfine inter- actions for some alicyclic radicals in the pure liquid phase or in solution Representative 19F isotropic hyperfine splittings Representative chlorine isotrOpic hyperfine splittings Representative 13C isotropic hyperfine splittings . 14 . . . . . Representative N isotropic hyperfine splittings Calculated values of relative intensities F(w) versus log(l/r) for l—cyclopentanecarboxylic acid radical in adamantane Experimental relative intensities F(w) at various temperatures and the corresponding cal- culated values of log(l/r) for l-cyclo- pentanecarboxylic acid radical in adamantane Calculated values of relative intensities F(w) versus log(l/r) for l-cyclohexanecarboxylic acid radical in the self—matrix Experimental values of relative intensities F(w) at various temperatures and the corresponding cal- culated values of log(1/r) for l-cyclohexane- carboxylic acid radical in the self matrix Calculated values of relative intensities F(w) versus log(l/r) for 2-cycloheptanonyl radical in adamantane Experimental values of relative intensities F(w) at various temperatures and the corresponding cal- culated values of log(l/r) for 2-cycloheptanonyl radical in adamantane Vflvii Page 24 25 31 35 37 38 54 SS 64 64 76 76 Table 13 14 IS 16 17 18 19 20 21 22 23 Spin orientations for 2-tetrahydropyranyl radical under conditions of slow and fast ring inversion Calculated values of relative intensities F(w) versus log(l/r) for 2-tetrahydr0pyranyl radical in adamantane Experimental values of relative intensities F(w) at various temperatures and the corresponding cal- culated values of log(l/r) for Z-tetrahydropyranyl radical in adamantane ‘ Calculated values of relative intensities F(w) versus log(l/r) for l-pyrrolidinyl radical in adamantane Experimental values of relative intensities F(w) at various temperatures and the corresponding calculated values of log(l/r) for 2-pyrrolidinyl radical in adamantane ' Some sets of ESR parameters giving calculated spectra in good agreement with the observed spectra for the radical in irradiated c-C4F4Cl2 ESR parameters in solid matrices obtained in this work and in related literature studies Dihedral angles for B protons for radicals in some ideal conformations Conformation of cyclic radicals from proton hyperfine interactions Activation energies, E , and frequency factors, (l/t)0, for ring inversion in some alicyclic radicals Kinetic parameters for ring inversion in some alicyclic compounds obtained from dynamic NMR viii Page 84 88 88 103 103 115 119 123 124 127 129 Figure LIST OF FIGURES Page Potential energy relationships among the con- formations of cyclohexane ll Interaction of electron and nuclear dipoles in a magnetic field H, where r is the separation be- tween the dipoles and 9 is the angle between H and the electron-nucleus direction 18 Spin polarization in a n-electron radical with an a-hydrogen substituent 22 Dihedral angles for a n-electron radical with the odd electron in a 2pZ orbital on carbon, approximately planar arrangement of bonds to that carbon, and two 8 hydrogens (view along Ca - CB bond showing dihedral angles 61, 82 for H , H ) 28 B1 B2 Hyperconjugation (left) by overlap for a B-CH three-hydrogen atom MO of the same symmetry as the odd-electron orbital of a n-electron radical resulting in net positive spin density on the B hydrogens (right) 29 (3) Geometry of the trifluoromethyl radical from ESR data, a(Fa) = 144 G and ¢ = 19°, (b) Transfer of negative spin directly from fluorine to carbon in n-electron radicals leaving positive spin density on fluorine 33 ESR spectra of l-cyclopentanecarboxylic acid radical in adamantane 49 Nuclear spin states for the four protons in l-cyc10pentanecarboxylic acid radical 51 Field positions and intensities for the lines in the ESR spectrum of l-cyclopentanecarboxylic acid radical in adamantane (the numbers designate the spin states in Figure 8): (a) fast exchange region, (b) slow exchange region 52 ix Figure 10 11 12 13 14 15 16 17 18 19 20 21 Calculated values of relative intensity [F(w)] versus log(l/r) for l-cyclopentanecarboxylic acid radical (second line relative to outer in- variant line) in adamantane Plot of log(l/r) versus l/T°K for l-cyclo- pentanecarboxylic acid radical in adamantane ESR spectra of l-cyclopentanecarboxylic acid radical in the self matrix Field positions and intensities of the lines in the ESR spectrum of l-cyclopentanecarboxylic acid radical in the self matrix: (a) fast ex- change region, (b) slow exchange region ESR spectra of l-cyclohexanecarboxylic acid radical in the self matrix Calculated values of the relative intenstiy of the second line to the invariant outer line, [F(w)],versus log(l/r) for l-cyclohexane- carboxylic acid radical Plot of log(l/r) versus l/T°K for l-cyclo- hexanecarboxylic acid radical in the self matrix ESR spectra of l-cyclobutanecarboxylic acid radical: (a) in polycrystalline l-cyclo~ butanecarboxylic acid, (b) and (c) in adamantane ESR spectra of l-cycloheptanecarboxylic acid radical in adamantane ESR spectra of 2-cyclopentanonyl radical in adamantane ESR spectra of 2-cycloheptanonyl radical in adamantane (lines indicated by inverted tri- angles). The remaining lines arise from a second, unidentified radical Calculated relative intensities for the second line relative to the invariant outer line versus log(l/t) for 2-cycloheptanonyl radical in adamantane Page 56 57 S9 60 62 65 66 68 69 72 74 77 Figure 22 23 24 25 26 28 29 30 31 32 33 34 Plot of log(l/r) versus 1/T°K for 2-cycloheptanonyl radical in adamantane ESR spectra of 2-tetrahydrofuranyl radical in adamantane Calculated values of relative intensities [F(w)] versus log(l/r) for 2-tetrahydrofuranyl radical in adamantane Plot of log(l/r) versus l/T°K for 2-tetra- hydrofuranyl radical in adamantane ESR spectra of 2-tetrahydropyranyl radical in adamantane Field positions and intensities for lines in the spectrum of 2-tetrahydropyranyl radical in adamantane. The numbers refer to the nuclear spin states of Table 15 Calculated values of the relative intensities of the lines indicated by arrows in Figure 26, F(w), versus log(l/t) for Z-tetrahydropyranyl radical in adamantane Plot of lot(l/r) versus l/T°K for 2-tetra- hydropyranyl radical in adamantane ESR spectra of 1,4-dioxanyl radical in adamantane Calculated values of relative intensity, F(u), versus log(l/r) for 1,4-dioxanyl radical in adamantane Plot of log(l/r) versus l/T°K for l,4-dioxanyl radical in adamantane ESR spectra of 2-tetrahydrothiophenyl radical in adamantane Calculated values of the relative intensities, F(w), of the lines indicated by arrows in Figure 33 versus log(l/r) for 2—tetrahydrothiopheny1 radical in adamantane Plot of log(l/r) versus l/T°K for Z-tetra- hydrothiophenyl radical in adamantane xi Page 78 80 81 82 85 86 89 90 91 93 94 95 97 98 Figure 36 37 38 39 4O 41 42 43 44 45 ESR spectrum of 2-tetrahydrothiopyranyl radical in adamantane at —92°C ESR spectra of l-pyrrolidinyl radical in adamantane. Lines used for measurement of relative intensity are indicated by arrows Calculated values of the relative intensities, F(w), of lines indicated by arrows in Figure 37 versus log(l/r) for l-pyyrolidinyl radical in adamantane Plot of log(l/r) versus l/T°K for l-pyrrolidinyl radical in adamantane ESR spectrum of 2-piperidinyl radical in adamantane at 25°C ESR spectrum of cyclopentenyl radical from a sample of cyclopentyl chloride in adamantane irradiated at room temperature ESR spectrum of cyclopentenyl radical from a sample of cyclopentyl bromide in adamantane irradiated at room temperature ESR spectrum of cyc10pentyl radical in adamantane: (a) ~121°C, rapid ring inversion, (b) —187°C, slow ring inversion ESR spectrum of the l-hydroxycyc10penty1 radical in adamantane at 25°C. The line positions of the qunitet are indicated below while the remaining lines are attributed to other radicals (a) ESR spectrum at -60°C of radical produced by irradiation of l,2-dichlorotetraf1uorocyclobutene in adamantane, (b), (c), (d) simulated spectra using the parameters of Table 18(b), (c) and (d), respectively xii Page 100 101 104 105 106 108 108 110 111 112 Formula I II, III IV, V VI VII VIII IX, X XI XII XIII XIV XV XVI XVII XVIII XIX XX (50 (b) LIST OF STRUCTURAL FORMULAE Cyclopentenyl radical Cyclohexyl radical-equivalent chair forms 2—Cyclohexanonyl radical-equivalent half- chair forms 1,4-Dioxanyl radical 4-Piperidinyl radical (protonated form) Morpholinyl radical Cyclohexenyl radical-equivalent forms l-Cyclopentanecarboxylic acid radical- envelope forms and half-chair forms 1-Cyclohexanecarboxylic acid l-Cyclobutanecarboxylic acid radical: (a) planar, (b) puckered l-Cycloheptanecarboxylic acid radical: (a) chair, (b) twisted-chair l-Cycloheptanecarboxylic acid radical-un- symmetrical twisted-chair conformations 2-Cyclopentanonyl radical 2-Cycloheptanonyl radical-equivalent twist- chair forms 2-Tetrahydrofuranyl radical 2-Tetrahydropyranyl radical Z-Tetrahydrothiophenyl radical—equivalent half-chair forms 2-Tetrahydrothiopyranyl radical-equivalent twist-chair forms xiii Page 12 12 14 14 14 15 SO 62 71 71 71 74 76 84 84 97 100 Formula XXI XXII XXIII XXIV XXV xxvr xvaI XXVIII XXIX xxxrf XXXII l-Pyrrolidinyl radical-equivalent half-chair ferms 2—Piperidinyl radical-equivalent chair forms Cyclopentyl radical l-Hydroxycyclopentyl radical l,2—Dichlorotetraf1uorocyclobutene-anion or cation radical l,2-Dichlorotrifluorocyclobutene radical Adamantane 2-Methyltetrahydrofuranyl radical Z-Methyltetrahydrothiophenyl radical Cycloheptanone ketyl 1,2-Dichlorotetraf1uorocyclobutene radical anion 2,2,3,3,4,4-Hexafluorocyclobutanone ketyl xix Page 100 108 114 114 114 114 117 131 131 136 148 148 INTRODUCTION Chemists have been interested in magnetic resonance spectroscopy because of the variety of information of chemical interest which can be obtained from it. Nuclear magnetic resonance spectroscopy (NMR) has been particularly useful for the study of diamagnetic materials while electron spin resonance spectroscopy (ESR) requires the presence of unpaired electrons and so has been important in the study of paramagnetic substances. Since the discovery of ESR by Zavoisky,1 it has been applied by chemists to many areas of chemistry including the study of inorganic radicals,2 transition metal ions,3 organic radicals,4 and paramagnetic systems of biological origin.5 Although most organic and inorganic compounds are diamagnetic, a variety of free radicals may be obtained from them by chemical oxidation or reduction, or by irradiation with UV-visible light, v-rays, X-rays or electrons. Chemical information which may be obtained from a study of the ESR spectra of radicals includes identification of the radical species, electronic and geometrical structures of these species, and kinetic parameters such as rates of conformational changes. From rates of exchange measured as a function of temperature it is possible to obtain energy barriers restricting conformational changes and the accompanying entropy changes. Samples for ESR investigation may be single crystals, P01ycrystalline powders, solutions or gases. ESR spectra of gases have only been interpreted for diatomic radicals and are generally too complex to be useful. The most information is obtained from the anisotropic spectra of single crystals studied as a function of orientation since both isotropic and anisotropic nuclear hyperfine interactions are obtained.6.9 Polycrystalline spectra are a superposi— tion of single crystal spectra for all orientations; while they are often difficult to interpret, methods have been developed for obtaining ESR parameters from powder spectra. Although most information is obtained from single crystal ESR studies, single crystals are often hard to obtain and analysis of the spectra can be complicated. Lines in solids are usually broad and line shape analysis may be difficult where there are overlapping lines and unresolved hyperfine structure. As a result, there are relatively few studies of conformational changes, which depend on line shape analysis, in single crystals or powders. Solution ESR spectra are simpler since only isotropic hyperfine interactions are observed if the tumbling rate of the radicals is large enough to completely average out the anisotropic hyperfine interactions. Radicals are often too short-lived in solution for ESR study but a number of techniques have been developed for obtaining good spectra in solution. Fessenden and Schuler10 generated steady-state concentrations of radicals inside the ESR cavity with a pulsed high emrgy electron beam and obtained excellent isotropic Spectra of a variety of hydrocarbon radicals. Norman and co-workers11 used a rapid flow system in which two or more reactants are mixed in the ESR cavity to produce the desired radical. Kochi and Krusic12 generated the desired steady-state concentration of radicals by UV- irradiation of a mixture of tfbutyl peroxide with the precursor. It is also possible to obtain isotropic ESR spectra of radicals in solids by various matrix isolation techniques since radical recombination is often slow in solids, particularly at low temperatures. Thus, Bennett and co-workers13 devised an ingenious rotating cryostat in which free radicals, generated chemically, are deposited on a cold rotating drum with an inert matrix such as camphane or xenon. By mixing the vapor of an appropriate organic halide with sodium vapor and co-depositing the mixture on the drum, a variety of radicals were made and the ESR spectra obtained. Within a certain temperature range, characteristic of the matrix, it was found that the radicals tumbled freely enough to give isotropic spectra but recombined slowly enough that good ESR spectra could be obtained. Other solid matrices which have since been used to trap radicals include germanium tetrachloride,14 sulfur hexafluoride,15 and various substances, such as urea, which fbrm stable inclusion compounds.16 In the course of an attempt to prepare adamantyl radicals 7 discovered that adamantane by irradiation of adamantane, Wood and Lloyd1 provided a useful solid matrix for obtaining isotropic spectra of organic free radicals. Adamantane (C10H18) is a tricyclic hydrocarbon Which is not easily damaged by high energy radiation. It is a "plastic" crystal, sublimes even at room temperature, undergoes a change 18,19 and of crystal structure from tetragonal to cubic at 208°C, ferms solid solutions with many organic compounds. When a solid solution of an appropriate precursor in adamantane is irradiated the adamantane does not appear to suffer radiation damage and the radicals obtained are those from damage to the precursor molecule. Within a temperature range from that corresponding to onset of rapid tumbling of the radicals (usually ~60 to -80°C) to the temperature at which rapid decomposition of the radicals occurs, the ESR spectra of the radicals produced are isotropic. It is believed that the radicals are in substitutional rather than interstitial sites.17 Adamantane therefore provides an excellent medium for the study of conformational changes in radicals since a wide temperature range can be covered (as much as -160° to +80°C using the Varian Temperature Controller), line widths tend to be reasonably narrow (2-SG) and interactions between the radical and host crystal are small.17 Where the radicals are stable at room temperature the solid solution of precursor in adamantane can be irradiated at room temperature, otherwise irradiation is carried out at low temperatures. In either case the ESR spectra are followed as a function of temperature and at each temperature the microwave power and the modulation amplitude are varied to separate, as far as possible, lines arising from different radicals. Below the transition temperature of adamantane (208°K) the lines tend to broaden and the Spectra become those of a polycrystalline material; these are sometimes difficult to interpret and have been analyzed in this work only when 'the anisotropic splittings are small. Studies of conformational changes ida radicals trapped in an adamantane matrix have been reportedzo-22 from various laboratories but no systematic study of a series of compounds using exact lineshape analysis methods has been made. In the present investigation a series of alicyclic hydrocarbons and saturated heterocyclic compounds have been dissolved in adamantane, irradiated with X-rays or Y rays, and the ESR spectra obtained over a temperature range. Where the spectra show the effects of ring motions, detailed line shape analyses have been made and the barriers to ring inversion estimated. The results have been interpreted and compared with those for the diamagnetic precursors as well as with those for other radicals which are reported in the literature. HISTORICAL SURVEY I. Cyclic Organic Radicals The ESR spectra of the alicyclic radicals, such as the cyclopentyl and cyclohexyl radicals, obtained by loss of a hydrogen atom (or other group) from the corresponding saturated hydrocarbon (or substituted hydrocarbon), were first observed after high energy irrradiation of polycrystalline solidsz‘g"26 at 77°K but the resolution was not very good. The first high-resolution isotropic ESR spectra of the cyclic radicals from cyclopropyl through cycloheptyl were obtained by electron irradiation of both liquid and polycrystalline . 10,27 solid precursors. Cyclopentyl and cyclohexyl radicals have also been obtained by reaction of sodium with the corresponding halide28 and by irradiation of the iodides.29 Cyclopentyl radical has been prepared in a cyclopentane-neohexane glass30 and in a solid adamantane 20 Cyclohexyl radical was observed in an adamantane matrix21 matrix. and in a clathrate-hydrate matrix.31 A number of hydrocarbon radicals with one or more substi— tuents have been found to be stable and their ESR spectra observed. The series l-hydroxycyclobutyl, l-hydroxycyclopentyl , l-hydroxycyclohexyl and l-hydroxycyclooactyl was reported by Corvaja §t_al,32 and 3-amino- cyclopentyl and 4-aminocyclohexy1 by Norman §t_al,33 A particularly interesting series of l-carboxy radicals is obtained by irradiation 34,35 33-38 of the salts of the cyclic acids, the solid acids themselves, or the 1,1-dicarboxy1ic acids;39 many are stable even at room temperature. Irradiation of the cyclic halides appears to result in loss of the halogen except in the case of fluorocarbon halides which often give fluorine substituted cyclic radicals.10 A series of 1—substituted cyclopentyl radicals was prepared by U.V. irradiation of l,1'-azobis (l-X-cyclopentanes),40 where X = CN, CNO, CH OH and H. 3. The heterocyclic radicals obtained by loss of a hydrogen atom (or other group) from saturated heterocyclic compounds containing oxygen, nitrogen or sulfur atoms have been extensively studied by ESR and have been produced by a variety of techniques.41 The most stable radical is usually the one with the odd electron on the carbon adjacent to the heteroatom although other radicals are sometimes also observed. The 2-tetrahydrofuranyl radical has been obtained from tetrahydrofuran (THF) by reduction with Ti3+ in a chemical flow system}’33 by radiolysis with high energy electrons at low temperature,41 by UV photolysis in the presence of di-t;butylperoxide (DTBP),42’43 and by irradiation in an adamantane matrix.20 When 2-methyltetrahydro- furan is irradiated in an adamantane matrix more than one radical is obtained and their correct identification is still the subject of 44’45 Dioxane yields only the dioxanyl radica1.11’°2’33’42’43 argument. Tetrahydrothiophene gives the 2-tetrahydrothiophenyl radical43 and thiodioxane the corresponding thiodioxanyl radical.32 The nitrogen heterocycles give more complex products since there is a hydrogen, or other group, on nitrogen which may be lost or damaged in addition to the expected ring damage.41 The abstraction of a hydrogen atom with hydroxyl radical derived from the interaction of T13+ ion and hydrogen peroxide in aqueous solution in a flow system has been used to generate a large number of cyclic radicalsi1’33’46’47 Abstraction of a hydrogen atom with t-butoxyl radical derived from the photolysis of di-t-butyl . . . . . . 1 peroxide in a solution of the precursor alone,or mixed with isopentane, 32,42,43,48 has been widely used to produce cyclic radicals. Abstraction of a hydrogen atom from an alcohol or ether by photolysis in the 9 presence of a ketone has also been used.4 The cyclic radicals from CSHS- through C6H11. were prepared:58 by photooxidation of the carboxylic acids in the presence of Ce(IV). Irradiation of the precursor in a solid adamantane matrix has been employed recently in a number of 20,44,4s,so-52 a laboratories to produce both heterocyclic radicals nd radicals from cyclic ketones.22’53 In a number of cases more than one hydrogen atom is lost and a cyclic allylic—type radical results. Thus, irradiation of cyclopentane or a cyclopentyl halide in adamantane has been feund in the present work to give as one of the products an allylic radical which could be I or a more complex radical containing the grouping -CH2--CHLL$CHLL£CH-CH2- . An allylic radical formed as a radiation product of cycloheptane in thiourea:54 and the authors suggested that it was not cycloheptenylbut a complex radical resulting from reactions of the original radiation products. Allylic radicals have been observed as products of the irradiation of unsaturated compounds and 10,25,55—59 the proton hyperfine splittings for these, and the related cyclohexadienyl type radicals,61’61 are well characterized. A number of cycloalkenyl radicals were produced by Y irradiation of cyclic olefins in adamantane and the very well resolved ESR spectra provided proton hyperfine splittings for protons in various ring positions.62 A study of substituted cyclohexenyl radicals in adamantane has also been made.63 Analogous heterocyclic radicals have been identified in v-irradiated furoic acid and thiophene—Z-carboxylic acid.64 The oxidation of cyclic a-diketones has yielded a series of cycloalkane- semidiones,65 and the reduction of cyclic ketones a series of cyclic 66 ketyls, in both of which the odd electron is on oxygen, and ring inversion in these has been studied by ESR. II. Rates of Conformational Change from ESR Spectroscopy The different possible ring conformations for common ring systems, their relative energies and the rates of interconverSion between them have been extensively investigated and the subject is . . . . 67 68 dealt w1th in modern organic chemistry textbooks and monographs. Barriers hindering interconversion between conformers have been determined by NMR for a number of ring compounds and the subject has been reviewed.69 The NMR method does not give experimental barriers 10 when these are smaller than about 5 kcal/mole but microwave and infra- red spectroscopy are useful for the study of lower barriers.70 The ESR spectra of a number of cyclic radicals show lineshape changes with change of temperature and these can be used to determine rates of interconversion between conformers in the range 105 - 109 Hz and barriers below about 10 kcal/mole. Linewidth effects in ESR spectroscopy, and their application in studying conformation changes, have been reviewed.71-73 A detailed study of cyclohexyl radical includes a mathematical treatment for the case of Gaussian lineshapes?7 Ring motions in the 2-cyclohexanony1 radical have also been studied by ESR and details of the analysis of the spectra given.22 The most stable form of cyclohexane itself is the chair form in which the angle strains, torsional strains and Van der Waals strains are minima. Conversion from one chair form to theequivalent one requires that the molecule go through less stable conformations and the energy relationships among these are shown in Fig. 1. The activation parameters for the chair-chair interconversion have been measured by NMR spectroscopy69 and are AH* = 9.1 i 0.5 kcal/mole, AG* = 10.6 kcal/mole at -67°C, and AS* = —5.8 i 0.4 eu. In cyclohexyl radical the three atoms bonded to the ring carbon atom bearing the odd electron must be almost planar since the carbon-13 hyperfine splitting74 is only about 41 G, the same within experimental error as observed in the planar methyl radical. The inversion barrier determined from the ESR spectra of cyclohexyl radical in a solid ll l I kcal l’ulculiul cncrgy -—->- 5.5 kcal ##waw Chair Half— Twisr- Boar TwiSI- chair boat boat I 11 Conform" C on/nnm‘r . C onformer Figure 1. Potential energy relationships among the conformations of cyclohexane. 12 cyclohexane matrix27 is E‘ = 4.9 i 0.5 kcal/mole. This must correspond to a.chair-chair inversion (II-III) modified by the introduction-of 2 . . . . an sp center at one carbon in the ring and is con51derably smaller than II ‘ III the barrier in cyclohexane itself. The introduction of a second sp2 center into the ring in 2-cyclohexanony1 radical (IV,V) lowers the observed inversion barrier furtherzz’SS to 3.7 i 0.1 kcal/mole. The IV V barrier for cyclohexane determined from NMR data69 is 5.4 Real/mole, rather similar to that found for cyclohexyl radical. The situation is quite different in the five-membered saturated ring series. Cyclopentane undergoes rapid pseudorotation which is hindered by only 13 a small barrier.70’75’76 The barriers to interconversion between the different stable conformers of cyclopentane, tetrahydrofuran and tetrahydrothiophene are too small to measure by NMR. When the radicals obtained by abstraction of a hydrogen atom are studied by ESR, linewidth changes are observed and from these it is possible to estimate barriers hindering interconversion between ring conformers. For cyclopentyl radical the barrier in a clathrateehydrate matrix was estimated to be 2.7 i 0.2 kcal/mole while in adamantane it was 2.8 i 0.3 kcal/mole. The corresponding barriers fer the 2-tetrahydrofuranyl radical in a clathrate-hydrate and in adamantane are 2.1 i 0.2 kcal/mole and 2.4 i 0.4 kcal/mole, respect- 20,31 ively. Not many substituted radicals have been studied but l-cyclobutane-carboxylic acid radical shows a transformation between ring conformations hindered by a barrier of 1.5 kcal/mole in the v-irradiated sodium salt of cyclobutane-l-carboxylic acid35 whereas the same radical (but possibly with a proton on the carboxyl group) in 1,l-cyclobutanecarboxylic acid shows a barrier to ring inversion of only about 0.06 kcal/mole.39 The barrier to ring inversion for l-cyclopentane carboxylic acid radical in the sodium salt of cycle- pentaneel-carboxylic acid has been reported as 0.37 kcal/mole35 whereas fer the same radical (possibly with a proton on the carboxyl group) in 1,l-cyclopentanecarboxylic acid the value is £3 = 3.70 i 0.05 39 kcal/mole. Leung and Hunt mention in a footnote to their article37 on irradiated cyclohexanecarboxylic acid that the spectra indicated 14 an interconversion barrier of 4.9 kcal/mole for l-cyclohexane carboxylic acid radical. Gilbert gt 31.33 report barriers for ring inversion in three heterocyclic radicals obtained by T13+ reduction of precursors in aqueoushacid Solution, the values for the radicals from dioxane (VI), piperidine (VII) and morpholine (VIII) being 6.74 i 0.7 kcal/mole,. 7.27 i 0.6 kcal/mole and 9.25 i 0.9 kcal/mole, respectively. Pratt 22,53 and co-workers have measured the barrier to inversion in 2-cyclo- ' VI VII VIII hexanonyl radical (IV,V), AHI = 3.7 i 0.1 kcal/mole, and several other substituted cyclohexanonyl radicals in adamantane. Gilbert and Trenwith33 report a much higher value for the barrier in 2-cyclohexanonyl radical in acid solution (AHI = 5.44 i 0.5 kcal/mole). “Cyclohexenyl radic3163 in adamantane inverts between two equivalent conformation (IX, X) and the barrier is reported as 6.81 t 0.58 kcal/mole (C5 is above or below the plane formed by the other five carbon atoms but H,H' differ in IX and X).63 15 Ix " Although ring inversion barriers have now been measured for a dozen or so radicals, it is difficult to draw any conclusion from the results since some values are for the liquid phase while others are for radicals in a variety of solid matrices. Also, the analysis of the spectra to provide the inversion rates has differed and various approximation methods have been employed. 16 THEORY I. Analysis of ESR spectra77'79 The energy W of a magnetic dipole fi’in a magnetic field H- is given classically by the expression W = - EZH' = - uH cos 6 (l) where 9 is the angle between D and HI Electron and nuclear magnetic moments are, however, quantized and the allowed energies Wi are the solutions of the Schrodinger equation A w. = w.w. (2) where H is the spin Hamiltonian operator and wi are the spin eigenfunc- tions. For a single electron in a magnetic field H, the spin Hamiltonian may be written R = gBHSz, where 82 is the operator for the 2 component of the spin and z is the direction of quantization H. The spin eigenfunctions for the electron are Ia> and IB> and the energy of transition between these two allowed spin states Au = gBH . hv (3) is the general resonance condition. The spectroscopic splitting factor g is 2.0023 for the free electron but is altered when spin-orbit cO‘JPIing becomes significant; the Bohr magneton B = ‘he/ch = 9.273 x -21 - 10 erg G 1. 17 If a magnetic nucleus of spin 1 interacts with an electron in an oriented radical (in a single crystal, for example), a more general spin Hamiltonian R = BRIEIS + SITLI - gNB HZI- (4) may be necessary to give experimental energy states and transitions in agreement with observations. The first term is the Zeeman energy term describing the interaction of the odd electron of spin S with the applied magnetic field, E is a second rank tensor and the Zeeman energy in a field of 3000 G is about 9 GHz. The second term describes the interaction of the odd electron with a nucleus of spin I; The hyperfine interaction tensor T is a second rank tensor and hyperfine energies for protons in organic radicals usually fall in the range 0-50 G. This term may be written .I'= 35. 4H 9. N[—:-3—- '5' 1, (s) ”W I on U) 00 2 OD where the isotropic hyperfine interaction a arises from the Fermi contact interaction and may be written a = 8w/3lngNBleolzl . (6) W0 is the electron wavefunction evaluated at r = O, r is the electron- nucleus distance and gN’BN are the nuclear g factor and nuclear magneton, respectively. The second term on the right—hand side of Equation (5) is the electron dipole-nuclear dipole interaction term which may also be written as a second rank tensor T', which must be traceless. IfiT 18 is diagonalized, a is obtained from the sum of the diagonal elements a = 1/s(rXX + Tyy + T22) (7) and the diagonal elements of T' are given by TI. = T.. — a. (8) When the radicals are able to tumble freely the dipole—dipole interaction is averaged to zero. This may be seen by averaging the dipole-dipole energy .5 Id .9 -9 “e I“ 9 , ”\/ / / Fig. 2. Interaction of electron and nuclear dipoles in a magnetic field H, where r is the separation between the dipoles and 6 is the angle between H and the electron-nucleus direction. 19 W = [(1 - Seeszel/r3] uNzuez (9) over the surface of a sphere; then = l/3 and W = 0. Hence, in solution, or in a solid where the radicals have sufficient motional freedom, only the Fermi contact term a is found experimentally. In a single crystal the dipolar tensor T' is also found but ESR spectra 'muSt be obtained by rotating the crystal about three axes and taking spectra every few degrees. The Fermi contact term a is related to the 5 character of the odd electron orbital while the dipolar tensor T' depends on the magnitude and direction of the p (or d) character of the odd electron orbital. Values of Ti lead to estimates of the fraction of an odd electron in outer up (or nd) orbitals of atoms in molecules since Ti for occupancy by a single electron has been computed for various atoms from the wave functions. The final term in Equation (4) is the nuclear Zeeman term. The energy of interaction of a proton with a field of 3200 G is 5.26 G so this term is often small compared to the hyperfine interaction and may then be neglected. Solutions of the Schrodinger equation for radicals for which the spin Hamiltonian of Equation (4) is appropriate may be rather com- plex. However, the second and third terms are small compared to the first term in the magnetic fields commonly employed and we may then write : v H HO + H , (10) where AA H = gBHSz + aSZIZ+gNBNHIz (ll) 20 is the unperturbed Hamiltonian with z the direction of quantization, and H' =(a/2)(S+I_ + s_1+) =(sx1x + sny)a (12) is the perturbation term - the off~diagonal matrix elements. Second—order perturbation theory then gives the approximate wave functions wn and energy states En: W = ¢n - I :m 2 H) ¢m n .4 _ mfn m n 2 E-E E = c + - n m#n m n (13) where the basis functions mm are eigenfunctions of HO with zero-order energy values em. 21 II. Interpretation of Hyperfine Interaction Constants Hyperfine interaction constants are directly related to the electronic and geometric structures of radicals. ESR data for a variety of organic radicals have been reported and nuclear hyperfine splittings found to depend on the type of radical (U- or n-radical), the particular nucleus studied and its location relative to the odd-electron center(s)..77-79 In the following sections present knowledge concerning radicals of the type studied in this thesis, in which the odd electron is centered on a carbon atom, is reviewed and the interpretation of hyperfine inter- actions discussed for a-,- B- and v—hydrogens, a- and B-fluorines, a— and B-chlorines, carbon—13 when it is the odd electron site, and nitrogen-14. The calculation of radical structures and of hyperfine interactions by CNDO/INDO methods is also reviewed briefly. (a) a-Proton Hyperfine Interactions A detailed study of a-proton hyperfine interactions was first made fer the radical H-C(COOH)2 in irradiated single crystals of malonic acid.80 The diagonalized tensor for the a-proton was found to be T(Ha) =LTxx’Tyy’Tzz] = [-32.5, -21.7, -10.3]G, the isotropic interaction a(Ha) = -l/3[32.S + 21.7 + 10.3] = -21.5 G and the aniso- tropic portion of the interaction T'(Ha) = [-11.0, -0.2, +11.2]G. Some approximate calculationsBO-82 showed that the odd electron is almost .1008 in a 29% orbital on the central carbon atom (the z axisis taken normal to the plane of the radical) and that the radical is planar. This type of radical is often called a n-electron radical even when not strictly planar. Although it might be expected that the isotropic ur..._..‘.’a. 22 splitting would be zero, since the 2p orbital has a nodal plane through the proton, an appreciable spin density is observed and this must arise from odd electron density which has been transferred to the s orbitals of hydrogen. The sign of the interaction is negative which means that the net odd electron density on hydrogen has a spin opposite to that of the odd electron. This has been explained80 as resulting from spin polarization and the mechanism is illustrated in Fig. 3. The spin Figure 3. Spin polarization in a n—electron radical with an a-hydrogen substituent. orientation of the odd electron in the 2pz orbital on carbon is taken as positive. The electron of the bonding pair of the C-H bond which may be assigned to carbon tends to have the same spin by the Hund rule. The other bonding electron, which may be assigned to hydrogen, therefore tends to have a negative spin by the Pauli exclusion principle. The excess negative spin is small, the fraction of an odd electron on hydro- gen with negative spin usually being about 23/508, or 0.04, since the observed a proton hyperfine splitting is about -23 G for many n-electron radicals and the hydrogen atom itself shows a hfs of 508 G. The abbrevia tion hfs = hyperfine splitting is used extensively in the literature and occasionally in this work. 23 McConnell81 suggested that as the fraction of odd electron p_ in the sz orbital of a n-electron radical decreased, the a-proton hyperfine splitting a(Ha) would decrease according to the semi-empirical rule: atria) = QHo . (14) where QH is a constant for a series of related radicals and is usually assigned a value of about -23 G for simple aliphatic radicals. Values of QH for various series of organic radicals have been estimated by combining experimental observations with M0 theory.77'78 The observed a-proton hyperfine interactions for a number of cyclic radicals are given in Tables 1 and 2. The values are generally more positive than -23 G indicating that odd electron density has been transferred from carbon to the more electronegative atoms 0 or S. However, no single value of QH can be used for all these since radicals in which the :G-H group lies between two oxygen atoms may even show a positive a-hydrogen splitting. The dipolar contribution to the a-proton hyperfine splitting tensor of a :G-H radical from the interaction of an odd electron in a2pz orbital of the carbon atom with the electron distribution on hydrogen has been calculated82 and the tensor components feund to be [+15.3, -l.8, -l3.9] G along the C-H direction, the perpendicular to the radical plane, and the perpendicular to C-H in the plane, respectively. These values give good agreement with the experimental values for malonic acid radical80 and various from w-electron radicals of the same type. 24 Table 1. Representative a— and B-proton hyperfine interactions for aliphatic radicals in the pure liquid phase or in solution. Radical a(H )3 a(H ) a B . CHZCH3 -22.38 26.87 . C(CH3)3 - 22.72 . CHFCH3 -l7.3l 24.48 . CFZCH3 - 13.99 . CFZH (+)22.2 - . CHZCl -22 - . CCIZCH3 - 19.7 . CHZOH -l7.4 - . CH(OH)CH3 -15.4 22.2 . CH(COOH)CH3 -20.18 24.98 CZHS - N - CH - CH3 -l4.6 20.0 CZHS — N - CH2 - CH3 -l4.27 36.90 aValues are in Gauss. bAll values are from Reference 73 except the last two which are from References 17 and 113, respectively. 25 Table 2. Representative a- and B-proton hyperfine interactions for some alicyclic radicals in the pure liquid phase or in solution. Radical a(Ha)a a(HB) Referenceb C 21.20 36.66 10 .‘H 21.48 35.16 10 O—H 21.15 23 10 H ‘/ \/ 21.8 24.7 10 .H 18.0 38.1 33 I O QH 17.4 33,0 22(33) ( S 13.60 28.06 11(42,43) O H Q—H 16.07 24.41 101 Q 17.1 23.6 42(43) H 26.1 43 Z .3 16.9 5 H 26 Table 2 (continued) Radical a(Hh) a(HB) Reference OH 18.5 22.5 48 5 27.0 32(40) OH O—OH 27.6 32 16.3 29.1 47 :5 2! A a(H.) Values <>_H 1.12 .‘O 0.53 [>44 1.90 O.” 0.71 aValues are in Gauss. bWhere additional references are given in parentheses the values obtained are similar but are not given. cAll radicals are either planar or are inverting rapidly to give only an averaged value of a(HB). 27 (b) B-Proton Hyperfine Interactions Isotropic B-proton hyperfine interactions, a(HB), in n-electron radicals have been shown to depend on the dihedral angle 0 between the direction of the 2pZ orbital of the odd electron and the projection of the C-H bond direction onto a plane normal to the Ca-C bond 8 8 (Figure 4). Equation (15), where pa is the H 2 a(HB) = QCCH pa cos 0, (15) fraction of an odd electron in the sz orbital of the a carbon and H . . . 0 QCCH 1s a constant characteristic of the C-C—H system, has a quantum mechanical basis.78 The constant QCCH has been assigned the value 58.6 G from a study of CH C radicals.ll 3 The empirical relationship77 a(H ) = B + Bcosze (16) B o is also often used with B0 in the range 0-4 G and B in the range 40-50 G; the value of B depends on pg. In this work we will use Equation 16 . _ _ 2 _ H 2 along Wlth the value BO - 0 so that a(HB) — Bcos 0 - poQCCHCOS 0. When there is free rotation about the Ca—C bond, as is often the case B with CH C radicals, = 1/2 and a(HB) = B/2. 3 If the radical is not planar at the radical center, it has been found that B is altered and appears to decrease with increasing deviation from planarity.121 Ifigure 4. 28 )8 z 143} 91 111 92 1132 Dihedral angles for a n-electron radical with the odd electron in a sz orbital on carbon, approximately planar arrangement of bonds to that carbon, and two 8 hydrogens (view along Ca - CB bond showing dihedral angles 6 62 for HB , HB ; these are measured from the 1 2 1’ i z axes as shown). 29 The relationship of Equation (16) has been widely employed in interpreting the B—proton hyperfine splittings of simple n-electron radicals, and of cyclic radicals, in terms of structure (Tables 1 and 2). Thus, in cyclohexyl radical with a(HB ) = 39.4 G for two 8 protons l B ) 5.3 G for two 8 protons, the values 61 = -220, 62 = 82 2 and B0 = 4.5 G were obtained using Equation (16) along with the value %B+Bo = 24.7 G for the B-proton splitting in isopropyl radical. The radical is presumed to be nearly planar at the odd-electron carbon and a(H because of the large a-proton hyperfine interaction a(Ha) = 21.3 G and the small value of a(13C) (41.3 G)74 observed for that carbon atom. However, the values are consistent with either the chair or twisted boat forms for the radical.' It is surprising that a(HB), with maximum observed values as high as +50 G, may be much larger than a(Ha), which is typically Smaller than about —24 G (in absolute magnitude). Both the high values arui the positive sign have been attributed to hyperconjugation79-83 (Figure 5) which provides direct transfer of negative spin from the B Itydrogens into the 2pZ orbital of the odd electron; this leaves Figure 5. Hyperconjugation (left) by overlap for a B-CH3 three- hydrogen atom MO of the same symmetry as the odd-electron orbital of a n~electron radical resulting in net positive spin density on the B hydrogens (right). 30 net positive Spin density in the ls orbitals of the B hydrogens and should be a maximum for dihedral angle 0 = 0° and a minimum fer 0 = 90°, as observed. Anisotropic hyperfine interactions for B hydrogens are small since the electron nucleus distance r of Equations (5) and (9) is large. (c) v-Proton Hyperfine Splittings Hyperfine interactions with v protons of cyclic radicals are small but are sometimes observed and a few values are listed in Table 2. (d) Fluorine Hyperfine Interactions Information concerning u- and B-fluorine hyperfine interac- 84-86 10,87 tions has been obtained from both single crystal and solution Studies and some typical values are given in Table 3. It has been Shown that n-electron radicals with one or more 0: fluorine substituents 10’88 so that the odd electron is in an spn hybrid are not planar 011>ital rather than a pure p orbital. McConnell's rule, Equation (14), therefore does not apply for isotropic a-fluorine hyperfine interactions, but instead a large positive splitting is observed which increases with theiangle <1) by which the C-Fa bonds are bent from the plane normal to the odd electron orbital (Figure 6a). The angle of bending is known approximately for the fluorine-substituted methyl radicals from the observed values of a(lSC). (See Equation 17 in the next section). 31 Table 3. Representative 19F isotropic hyperfine splittings.a . b Radical a(Fa) a(FB) . c1=3 142.4 . CHF2 84.2 . CHZF 64.3 . CFZCF3 84.9 11.2 . crates)2 67.4 19.2 . C(CF3)3 -- 17.9 . CF012 84.6 10.5 . CH300820012 21.1 F2 F2<:::>>“C) 82.9 a(FY) = 37.33 F2 F2 F2 N02- 93.3 3.06 F2 H\ CH3 .._ 77.6 H = 15.6 H. F a( 8) CH3 CHiiH F : CH3 74.8 a(HB) 5.5 F2 _' F2<> F2 147.7 ___, F2 aAll values are from Reference 87 except the last three which are from References 116 and 117. bAll values are in Gauss. 32 10’84 as resulting from direct These results have been interpreted overlap of the p orbitals on fluorine with the odd-electron orbital and transfer of negative spin from fluorine to carbon by this direct mechanism. The interaction would be similar to that illustrated in Fig. 5 but with the lone pairs of the sz orbitals of fluorine, rather than the pseudo-n orbitals of the --CH2 or -CH3 groups, donating negative spin density into the odd-electron orbital (Figure 6b). Isotropic B-fluorine hyperfine splittings vary from 0-70 G land appear to depend on the dihedral angle between the odd-electron ‘ orbital and the C-FB bond according to an empirical relationship of the type of Equation (16). If B0 is taken to be zero and B = 76 G. then, for a freely rotating B-CF3 group,a(FB) = 38 G. Both a- and B-fluorine hyperfine interactions show consid- erable anisotropy; for a-fluorines85 T1[Fa] 3 [120, ~60, -6OG] and for B-fluorines86 it varies from zero to a maximum of T1[Fé) 3 [+54, -17, -17G]. The large net positive spin density in the 2p orbitals of the B fluorines leads to an appreciable anisotropic hyperfine 11“: eract ion . £2___Chlorine Hyperfine Interactions Chlorine hyperfine interactions behave qualitatively in a lmmnner similar to the corresponding fluorine interactions but are much . Smaller because of the smaller magnetic moment of chlorine. The ESR Spectra of chloro-radicals are more complex because two chlorine isotopes of different magnetic moment are present, each with spin I = 3/2 33 (b) c—— <— 0 Figure 6. (a) Geometry of the trifluoromethyl radical from ESR data, a(Fa) = 144 G and 0 5 19°, (b) Transfer of negative spin directly from fluorine to carbon in n-electron radicals leaving positive spin density on fluorine. 34 auui a large nuclear quadrupole moment. Some typical values of a(Cla) and a(ClB) are given in Table 4. f) Carbon-l3 Hyperfine Interactions If the radical precursor is enriched with 13C at the radical site, or if a sufficiently good signal-to-noise ratio can be obtained so that 13C—satellites are observable, the 13C hyperfine splittings may be obtained for the carbon bearing the odd electron. The isotrOpic 13 . . . . C interaction for a planar n-electron radical should be zero Since the carbon nucleus is at a node of the 2pZ electron orbital. However, even in the planar -CH3 radical it is found10 that a(13C) = 39 G. This is in excellent agreement with the value calculated by M0 theory for a vibrating ~CH3 radical.89 When the odd-electron orbital acquires some 5 character and the radical becomes non-planar the Karplus-Schrader . 89 equation a(lSC) = KO + K2[2tan2¢] (17) relates the observed splitting to the angle 6 (Figure 5) by which the 89 that for CH radical is bent from the plane. It is found theoretically 3 KO 3 4OG,whi1e Hartree-Fock theory gives K2 = ao(13C) = 11106 for the odd electron in a pure 25 orbital.90 The value a(lSC) = 2726 observed10 for 'CF3 thus leads to 6 = 17°, not far from the value expected fer sp3 hybridization of the orbitals of carbon. The frac- tional 5 character of the odd electron orbital may be estimated from the relationship ’Table 4. 35 Representative chlorine isotropic hyperfine splittings.a Radical a(Cla) a(ClB) . CHCl2 3.5 . CHZCI 2.8 . CHClCOOH 3.7 . CClFCONH2 3.0 . CCl3 6.25 . CCIZCFZOCH3 4.0 35 . ClCH2C(CH3)2 19°5(37C1) . 16.2( C1) . CFCl2 10.5 . CHZCOCHZCI a(Cly) = 0.42 . cnccuzcuz a = 14.2 HClC :7: CH 77: CH 1.2 2 3All values are from Reference 87. 36 025 = a(13C)/aO(ISC). (18) . . 13 . . . The anisotropic C interaction for the carbon bearing the odd electron is approximately of the form '(13C) : [289 _B) ‘8] (19) '4" with near axial symmetry. The value 280(13C) = 65 G has been calculated from Hartree-Fock theory for an electron in a pure 2p orbital on carbon.90 The observed maximum anisotropic hyperfine interaction ZBCISC) may thus be used to estimate the fraction of an odd electron in the 2pTr orbital from the relationship 02p = 23/230. (20) It is found, however, that an empirical value of 280(13C) = 45 G gives more reasonable p densities than the Hartree-Fock value which appears . . 13 . . . . to be too high. Some typical C hyperfine splittings are shown in Table 5. g) Nitrogen-l4 Hyperfine Interactions Isotropic hyperfine interactions for a nitrogen atom on which the odd electron is centered vary from 0-16 G. The nitrogen-14 nucleus with I = 1 has a rather small magnetic moment and an appreciable electric quadrupole moment. An a-nitrogen substituent gives a small isotropic splitting and 3(14Na) varies fumn 0-6 G for these. The anisotropic hyperfine interaction for a nitrogen bearing the odd electron is of the form of Equation (19), with approximate axial 37 Table 5. Representative 13C isotropic hyperfine splittings. Radical 3(13Ca)a a(ISCB) Reference . CH3 38.34 74 . CHFCONH2 40.0 85 . CHZF 54.8 10 . CFZCOO 79 85 . CFZCONH2 88.2 85 . CHF2 148.8 10 . CF3 271.6 10 <:::>>-f1 41.3 74 . CHZCH3 +39.07 —13.57 74 QH (+)26.6 a(13CH2) = (-)10.7 53 \O a(13c=0) = (-)14.0 PI (+)26.7 53 °O g (+)103.o 101 a O i Values in Gauss. 38 'Table 6. Representative 14N isotropic hyperfine splittings. Radical a(14N)a Reference CH3 - CH - NH2 4.9 17 CH3 — H - CH2 7.03 17 CH3 (>61 5.77 17 NH-CH2 QNCO 4.2 17 e ' 14.66 b e N me - 1.75 c OQKH H ' + (CH3)2NH 19.28 113 (CH3)2N 14.78 113 aAll values are in Gauss; me = methyl. bJ.R. Roberts and K.U. Ingold, J. Amer. Chem. Soc. 953 3228 (1973). cH. Taniguchi and Y. Kirino, J. Amer. Chem. Soc. 99, 3625 (1977). 39 14 . . . symmetry and 2B( N) = 35 6.90 Some typical observed nitrogen hyperfine splittings are shown in Table 6. III. Exchange Rates from ESR Line widths Rates of ring inversion motions may be obtained from a study of the ESR linewidths when the rate of exchange of nuclei with different hyperfine coupling constants Aa is of the same order of magnitude as l/Aa. The necessary relationship between exchange rate and linewidth is obtained by solution of the Bloch equations modified to include nuclear Spin exchange. This has been done for Lorentzian lineshapes and the theory is given in several textbooks-l7.79 In many cases the lines of the ESR spectrum are broadened by unresolved hyperfine inter- actions and this appears to be true for the cyclic radicals studied in this work. The lineshapes are then more nearly Gaussian and Ogawa and Fessenden27 modified the Bloch equations by substituting an empirical term ozt for l/Tz. Writing G(t) = Gl(t) + iGz(t) for the total magneti- zation, the rate of change of the in-phase and out-of—phase components of magnetization are . 2 [1(wo+wh1) - O t - l/T] G1(t) + G2(t)/T (21) d (E161 (t) (3E962(t) [1(w0+mh2) — Ozt - l/T] 62(t) + Gl(t)/T, (22) where T is the correlation time for the exchange, whl and wh2 are the isotropic hyperfine interactions for the spins at sites 1 and 2, respectively, and “0 = YeHO. The solutions to Equations (21) and (22) are 4O G1(t) = C121(t) : C222(t) (23) Gz(t) = 0121(t) + 0222(t), (24) where 21(t) = exp[—02t2/2 + (n1+y+iA/2)t] (25) 22(t) = exp[-02t2 + (nl—y+iA/2)t] (26) 21(t) = exp[-02t2/2 + (n2+y-iA/2)t] (27) 22(t) = exp[-02t2/2 + (nz-y—iA/2)t] (28) D1 = T(y+iA/2)C1 (29) D2 = T(-y+iA/2)C2 (30) HI = i(mo+ hl) - l/r (31) n2 = i(wo+ hz) L 1/1 (32) A = whz - whl (33) y = VQ’Z - (A/2)2 . (34) The line shape is given by91 1(w) = Rea1{f G(t)e-1wtdt} (35) o and the second derivative of the absorption, which is commonly recorded in this work, is then given by92 2 2 1.11;). = - 13 Real {;B+[-g++(1+gi)eEi/2((ZN);’[1-¢(E+)])}. (36) 3w 0 _ - — ~ — where (m-wm) gt = Y: + i O (37) Y+= [a I a2-1]a (38) a = Z/TA (39) a = A/Zo (40) A = whz — whl (41) mm = (wh1+wh2)/2 + mo . (42) 41 Certain lines of each spectrum are invariant to exchange of the nuclei and it is convenient to measure the intensities of the lines which are dependent on exchange rate relative to one of the invariant lines (for which T 4-a0. Thus, we measure F(w) = (321/3w2)/(32I/3w2) (43) T-HD and use these ratios, F(w), to obtain the correlation times, T , using either a graph27 or a computer program written by W. G. Waller.92 Details of the calculations and a listing of the computer program are given in Appendix I. A measure of the barrier restricting the exchange may be obtained from measurement of T at different temperatures. The Eyring equation r = (1/1) = (kT/h)exp(-AG*/RT) = (kT/h) $547R e-AH*/RT , or, £n(1/T) = £n(kT/h) + AS*/R - AH*/R(1/T)’ (44) permits the calculation of the enthalpy, entropy and free energy of activation fer exchange from a plot of 2n(1/T) versus 1/T°K. The rate r of a conformation change is inversely proportional to the 67,68 lifetime T of the conformation. The Arrhenius equation ln(1/T) = Ea/RT + inn/1)o (45) is also often used in discussing conformational changes and the energy barriers Ba and frequency factors (l/T)o are obtained from plots of £n(l/T) versus reciprocal absolute temperature. 42 EXPERIMENTAL I. Instrumentation Preliminary ESR spectra were recorded using a Varian E-4 X-band Spectrometer equipped with an E-257 Variable Temperature Accessory. More precise measurements were carried out with a Varian V-4502 X-band spectrometer system including V-4560 100kHz Field Modulation and Control Unit, Mark I Fieldial regulation, 12" electromagnet, V—4531 Multi-purpose rectangular cavity, and V—4540 Variable Temperature Accessory. Both first and second derivatives of the absorption were recorded, the latter by an arrangement for detecting at double the modulation frequency of 40 Hz. Second derivative curves are more convenient to analyze because of their resemblance to the absorption curve and were used in most cases. Sample temperatures were measured by placing a copper- constantan thermocouple in the cavity close to the sample and were read on a Trendicator 400 Electronic Temperature Meter. The error in temperatures measured in this way is estimated to be i2°C. Magnetic field calibration was carried out with a small NMR spectrometer using a doped water sample. The tunable marginal oscillator covered the range 12.2—21.2 MHz and proton markers were placed on the ESR spectra at various field intervals. The proton resonance frequencies were read directly on a Monsanto Model 1515A Counter-Timer to :1 kHz. All spectra were recorded on ll"xl7" chart paper using a Moseley Autograf Model 7000A X-Y recorder with the magne- tic field increasing from left to right. In this thesis all are second-derivative Spectra. Klystron frequencies were measured directly to 10.1 MHz with a Hewlett—Packard Model 5245L Electronic Counter equipped with a Hewlett—Packard Model 5255A Frequency Converter covering the range 3-12.4 GHz. Periodic checks on the performance of the spectrometer system and on the precision of measurement were made with standard Varian samples having the values g = 2.0028 (strong pitch) and g = 2.0037 (DPPH powder). 11. Chemicals Adamantane, C10Hl6’ purissimum (99+%) from Aldrich Chemical Co., Milwaukee, Wis., was used as a matrix in most of the ESR work and impurities which might give interfering radicals must be removed as far as possible.17 For this purpose the adamantane was dissolved in spectrograde n-heptane (Mallinckrodt Chemical Works, St. Louis, Mo.),treated with activated charcoal (Nuchar-C 190-N, West Virginia Pulp and Paper Co., New York, N. Y.), filtered,and n-heptane removed with a vacuum pump until most of the adamantane had crystallized out. The mother liquor was removed by filtration and the white crystalline powder dried. This process was repeated three or more times until a sample X-irradiated at room temperature gave very little background signal even at high microwave power. Such weak background signals become negligible relative to the stronger signals obtainable with most of the radicals studied even at much lower microwave power. Deuteroadamantane, C10D16 (Merck of Canada, Ltd., Montreal, Canada) 44 was used as a matrix in certain experiments to reduce line broadening. It was used as received. Most of the radical precursors used were obtained from the Aldrich Chemical Co. These include cycloheptanone, cyclobutanone, cyclobutanecarboxylic acid, cyclohexanecarboxylic acid, cyclopentane- carboxylic acid, cyclopentanol, cyclOpentylamine, cyclopentyl chloride, cyclopentyl bromide, tetrahydrothiophene, pentamethylene sulfide, tetrahydrOpyran, pyrrole and pyrrolidine. Cyclopropanecarboxylic acid was from Pfaltz and Bauer, Stamford, Conn.) and 1,2-dichloro- tetrafluorocyclobutene from PCR, Inc., Gainesville, Florida. 111. Sample Preparation and Handling Radical precursors were incorporated into adamantane by dissolving No.5 g purified adamantane in the liquid precursor, with the aid of heat where necessary, then removing excess liquid by evaporation. Where the liquids were not very volatile, the solution was rapidly frozen with liquid nitrOgen to incorporate as much pre- cursor as possible into the solid and the solid then allowed to warm to room temperature. Excess liquid was removed by suction filtration and the crystals dried under vacuum. The white, fluffy powders were made into a cylindrical pellet 3 mm in diameter and NI cm in length in a KBr pellet press using a Special tungsten carbide brushing (Strehlinger Co., Lansing, Mich.) andzihardened steel plunger which had been chrome plated. Where necessary the oxygen trapped in the pellet was removed by pressing the pellet while evacuating the die. 45 Solid precursors were cosublimed with adamantane, which sometimes results in the precursor molecules being incorporated into the adamantane,l7 or were melted together with adamantane. These methods did not lead to the observation of any suitable ESR signals with any of the solid precursors tried in this work and it is assumed that none were incorporated into the adamantane. In several cases pure crystalline solids were irradiated ("neat" samples). Where these were liquids at room temperature, they were sealed in quartz tubes (3 mm ID, 4 mm OD) after dissolved oxygen had been removed by repeated freeze—pump—thaw cycles. Samples which had been irradiated at room temperature were transferred directly to the ESR cavity and Spectra taken at room temperature. The sample temperature was then slowly lowered using the variable temperature control system and spectra were taken at intervals down to the lowest accessible temperature (about -190°C). Samples irradiated at 77°K were transferred to the cold ESR cavity as rapidly as possible to prevent warmup and initial spectra taken at as low a temperature as feasible, usually close to 77°K. The sample was then warmed slowly and spectral changes followed as a function of temperature. In most cases several cycles of warming and cooling were followed by ESR in order to separate reversible from irreversible changes and to aid in analyzing the often very complex changes occurring. When it appeared from the Spectra that more than one radical was present, power saturation studies were carried out. In many cases the lines of some radicals saturate more readily than 46 those of others and the changes in the spectrum with microwave power level assist in assigning the spectra. IV. Irradiation X—irradiations were carried out with a GE XRD-l X-ray generator Operating at 20-40 kV and 10 ma with a chromium target tube. y-Irradiations used the 60Co gamma irradiator housed in the Food and Human Nutrition Building. Where the radicals studied were stable in the solid matrix at room temperature, X-irradiation at room tempera- ture for 30-120 minutes was normally used. Where the radicals studied were not stable at room temperature, or where neat liquids were to be irradiated in the polycrystalline state, Y irradiation was carried out at 77°K. Small vials (l dram) or sealed quartz tubes (3 mm ID) were placed in a Dewar filled with liquid nitrogen and irradiated for 3-5 hours at a dose rate of about 0.3 Mrad/hr (3 x 107 erg gm-1 hr-1). V. gAnalysis of Data Radical g values were calculated from the measured magnetic field H0 at the center of the spectrum and the measured Klystron frequency v using the equation g = hv/BH0 with e = 9.27 x 10"21 erg gauss'1 and h = 6.6262 x 10‘27 erg sec. Magnetic fields were computed from the observed proton magnetic resonance frequencies using the relation 47 H(gauss) = va/yp = 2.348682 x 10‘4 v(Hz) based on the value yp = 2.6751965 x 104 rad sec.1 gaUSS-l for the magnetogyric ratio of the proton. I. Carrelation times for ring inversion I) were calculated from the intensities of those lines which varied with temperature, 'relative to those (usually the outer ones) which did not, by use of the computer program ESREX4 written for the CDC-6500 computer (see Section III of THEORY, above, and Appendix I of this thesis). The data input to this program include the measured Klystron frequency mo = 2flvo and the hyperfine coupling constants whl, mhz RAD/SEC.... fer the Spectral lines employed, along with their relative peak heights in second-derivative presentation. Measured hyperfine split- tings : Gwere converted to frequency units RAD/SEC for the program using the measured g value and the relationship A RAD/SEC = (ZNB/h) a (g/ge) . where ge = 2.00232 is the g value of the free electron. The program BSRCONl, which simulates isotr0pic ESR spectra, was also used in this work. Estimated hyperfine interactions for up to five different groups of nuclei may be input and the simulated ESR Spectrum displayed on a Calcomp plotter. Some preliminary Spectral simulations were done using the Tektronix 4010 during the early stages of this research. 48 RESULTS I. Cyclic Carboxylic Acids A. Cyclopentanecarboxylic Acid in Adamantane A solid solution of cyclopentanecarboxylic acid (CPCA) in adamantane was X-irradiated at room temperature. The ESR spectrum at 25°C [Figure 7(a)] is a quintet with relative line intensities approximately l:4:6:4 l and a(HB) = 31.5 G. On lowering the temperature the intensities of the inner lines change relative to the two outer lines, which are invariant in intensity. Selective broadening of lines 2 and 4 as the temperature is lowered [Fig. 7(b) - (e)] corresponds to intermediate rates of exchange. At -109°C the second and fourth lines have disappeared and a triplet with relative intensities 1:4:1 remains [Figure 7(e)]. On cooling further the lines broaden and seven-line spectra [Fig. 7(f) - (h)] appear with a(HBI) = 22.1 G and a(HBZ) = 41.0 G; these are presumably powder spectra since the lines are broad, as would be expected when the tumbling rate of the radicals becomes small and the adamantane matrix changes to its low-temperature structure. The resolution remains good even at the lowest temperatures Since B-proton splittings are nearly isotropic. These spectra are similar qualitatively to those observed for cyclohexyl radical, and first interpreted in detail by Ogawa and Fessenden,27 and are accounted for it the only stable radical prod- uced is the 1-cyclopentanecarboxylic acid radical (XI) with two "axial" and two "Equatorial" 8 protons and either the "envelope" or "half- chair" conformation (XI, top and bottom). Then, at low temperature where ring inversion is slow, the seven-line Spectrum corresponding 49 * in Adamantane +25°C COOH + i —109° C (b) If”) __ 0 70 C fl J, 452°C A l —97°c _ 1‘ i 4 493°C (d) - (h) lines used for intensity measurement. Figure 7. ESR Spectra of l-cyclopentanecarboxylic acid radical in adamantane. 50 "ENVELOPE ;;55 e R R=COOH a 9 "HALF-CHAIR" 0 e __3 R T’ e 0 e Cl XIa XIb to a partially overlapping triplet of triplets with a(HB ) = 22.1 G 1 and a(HB ) = 41.0 G,is obtained. On warming to room temperature, 2 ring inversion becomes rapid and the four 8 protons become equivalent giving a 1:4:6:4:1 quintet with = 1/2[a(H81) + a(H82)] = 31.5 G. At intermediate temperatures exchange of protons 81 and 82 by ring inversion leads to a mixing of spin states and changes in intensity of the three inner lines relative to the two outer lines. The nuclear spin orientations possible for four protons are shown in Figure 8 and the lines in the ESR spectrum when an odd electron interacts with two protons with a = 21 G and two with a = 42 G are shown in Figure 9. The spectrum of Figure 7(h) corresponds to the stick diagram of Figure 9(b) with seven lines with relative intensities l:2:3:4:3:2:1. The diagram of Figure 9(a), corresponding to rapid exchange of the two pairs of protons, shows how such rapid exchange leads to a narrow five-line quintet with = (21+42)/2 = 31.5 G and relative intensities 1:4:6:4:1. In the intermediate portion of Figure 9 the dotted lines 51 ”1 HR H p1 2R 2LCOOH * 1 LR LR LR LR if 1 2 3 4 H] -- --- —— -— H2 -- —+ +- ++ 5 6 7 8 H] —+ —+ —+ -+ H2 "‘— _+ +—’ ++ 9 10 H 12 ”I +— +- +— +- “2 -—— -—+ +— ++ 13 14 15 ‘6 H1 ++ ++ ++ ++ * L and R indicate left and right hand hydrogens on the diagram at the top. **There are sixteen combinations of the nuclear spin quantum numbers for the four 8 protons, labeled 1-16 above; mI = t1/2 are designated by + or -. Figure 8. Nuclear Spin orientations of the four 8 protons of l-cyc10pentanecarboxylic acid radical. 52 .cofimon omcmnoxo 30am any .conoH omcwcoxo ummm new "mm ousmfim aw mcowumucofluo :Hmm ecu oumcwwmop mpopfid: ocuv ocmucmsmpm :H Hmofivmu vwom oflfixxonuwoocmucomoHoxoufi mo sapwooam mmm can :« mocwa may now moflufimcoucfi can mcowuwmom vfioflm .m enamfim 0 3+ o_v+ o_~.+ o 2! o_..| o_ol .AA v.lllllwv. OmaIta «A: —.A.l0§ 0.. PIIW _ _ 2 n _ I 2 n a J I a. o x F _ / ax”. : ”I x r . / o. _ i \ k x / \ p - / \ / 0 _ \ a \ - _ / \ / _ \ , \ _ / \ x _ x , \ _ n \ / _ x , a _ — C z.\ / _ A3 _ 2 .AlllllumvmmU 111v >0: 2 IOOU 1 e a. 0:252:33 a 2 m. c. I . o. k 6 53 indicate that the line intensities are changing with exchange rate but with the two outer lines invariant in intensity. In this inter- *mediate region relative intensities of inner to outer lines may be used to obtain correlation times T according to Equation (43) using the theory of Section III. Values of log(1/T) as a function of the ratio of the intensity of line 2 to line 1 for l-cyclopentanecarboxylic acid radical (indicated by arrows in Figure 7) were computed by use of the computer program ESREX4 (Appendix B) and are given in Table 7. The experimental values of the relative intensity of line 2 to line 1 (F(w)) are shown in Table 8 at various temperatures and are plotted versus log(l/T) in Figure 10. Values of 10g(1/T) at various temperatures are plotted versus l/TOK in Figure 11 and the parameters of the Arrhenius rate equation, Equation 45, E8 = 1.43 kcal/mole and (l/r)o = 1.25 x 1010 sec-1, were obtained from the Slope and intercept of the best straight line through the experimental points. Using B0 = 0.0 G, B = 44 G and 01 + 62 = 600 in Equation 16, the experimental values of a(HB ) and a(HB ) for the l-cyclopentane- carboxylic acid radical lead tolthe valuesz45o and 15° for the dihedral angles between the 2p orbital of the odd electron and "axial" and "equatorial" B protons. These values are somewhat closer to the values of the dihedral angles (100 and 50°) for the half-chair conformation than they are to the "ideal" values for the envelope conformation (-100 and 70°). For this reason the conformations of the cyclic five-membered ring radicals have been drawn as half-chair forms in most of this work. 54 'Table 7. Calculated values of relative intensities F(w) versus log(1/I) for l-cyclopentanecarboxylic acid radical in adamantane.a,b log10(l/I) F(w) +11.0 1.99 10.5 1.97 10.0 1.87 9 5 1.68 9 0 1.17 8.5 , 0.42 8.0 0.00 7.5 0.28 7.0 0.66 6.5 0.88 6 0 0.96 5 5 0.99 aF(w) is defined by Equation 43 and is the ratio of the intensity of the given line to one of the outer invariant lines in the second- derivative spectrum. bThe rate of inversion is l/r where T is the correlation time for exchange (Equations 21 and 22). 55 Table 8. Experimental relative intensities F(w) at various temperatures and the corresponding calculated values of log (1/1) for l-cyclopentanecarboxylic acid radical in adamantanefi,b 3 T°C l/T°K(x 10 ) 8(6) 10g(1/T) +11 3.52 1.88 9.97 -11 3.82 2.00 10.80 -65 4.81 0.74 8.72 -81 5.21 0.29 8.40 -100 5.78 0.17 8.28 -121 6.58 0.00 8.00 -141 7.58 0.13 7.70 -152 8.26 0.26 7.53 -l6l 8.93 0.43 7.31 -171 9.80 0.63 7.04 8The values of l/r are taken from Figure 10. bThe lines used to obtain F(w) are indicated by arrows in Figure 7. 56 aha—v 00- Goo: .mcmucmemum :fi mocfia ucmfium>cfl amuse ou o>wumaou mafia vcooomv Hmuwvmu pace uwfixxonnmomceucomoHoxo-H How AP\HV moH mamno> Hmsvau xufimcoucw o>fiumaon mo mo=Hm> woustuHmu ll, “59— nag, 0N_ .oH mesmHe IOOQ occacoEovd‘ E “I o °|98 <31 Kigsualug 57 .ocmucmEmpm cw Hmofipma peom ofiaxxonamoocmpcomoHoxolfi pom on\H mamao> ne\vaoH mo uofia .HH opsmfim V3: 2 x 0.2 6.6 08 3 Q6 3 3. 3 :ml . w .1 m p P W) . . _ 03 law ado m 25. Z<<< = (28.3 + 34.0)/2 = 31.3 G, as observed. In the intermediate temperature range [Fig. 12(b) for —9°C] the outer lines remain invariant but the inner lines change in position and intensity as a result of the ring inversion process. The dotted lines of Figure 13 Show the results of averaging the two sets of B protons leading, at high temperatures, to the quintet of Figures 12(a) and 13(a). 59 200 "9‘" COOH O C _9°C (b) WM _36°C (C) ' ‘ Figure 12. ESR spectra of l-cyclopentanecarboxylic acid radical in the self matrix. w 60 .cofiwon omcmnuxo 30am may .cofiwoa owcmcoxo ummm new awakens mfiom ecu cw Hecapma vflom UHHxxonamoocaucomoHoxonfi we Edhuoomm mmm map :a mocfla ecu mo moflufimcoucw use m:0wuwmom macaw .mH onsmfim _ , _ _ Toma“ ILL T: 0:” _ , a i. a 5_ _ d \ i ‘ a _ _ _ ’_ /: :. n _ I E Z _ _ _ _ z _‘ x E _ _ Alloamlulw >om TinXU #00: ov+ 3:50 ou+ 0 Cal owl 61 The dihedral angles calculated from Equation 16 for the two sets of 8 protons of the CPCA radical in the pure CPCA matrix are 34° and 26°, respectively. The stable conformation in CPCA is thus different from that in adamantane. C. Cyclohexanecarboxylic Acid Pure solid polycrystalline cyclohexanecarboxylic acid (CHCA) was y-irradiated and the ESR spectra observed at various temperatures [Figure (14)]. At +9°C a quintet with a(H = 21.1 G and relative B) intensities 1:4:6:4:1 was observed [Figure l4(a)]. This quintet persists on cooling until about -50°C when it begins to change to a triplet. At —60°C a triplet with relative intensities 1:4:1 and a(HB) = 42.1 G is obtained [Figure l4(b)] and tins remains until about -90°C when a more complex spectrum appears. These spectra indicate that the only radical formed is cyclohexanecarboxylic acid radical (XII).. The low-temperature spectrum [Figure l4(c)] COOH ”IL 9 00011 XIIa XIIb is presumably a powder spectrum since the tumbling of impurity species in adamantane has usually slowed in this temperature range to the point where anisotropic interactions are broadening the lines. neat 20 G OCH 0 ::+9 C ~125°c “TWA/\f 0 “7] C (bWAh—W MN Figure 14. ESR spectra of l-cyclohexanecarboxylic acid radical in the self matrix. The spectra are thus qualitatively similar to those observed 1- for cyclohexyl radical~ and those described in detail above for l- cyclopentanecarboxyl radical and are consistent with a slow inter— conversion of radical XII between equivalent chair forms below -60° leading to hyperfine splittiNgs a(HB ) = 37 G for two axial B-protons l and a(H? ) 2 5.0 G for two equatorial B protons. Above ~50°C inter- conversion is rapid enougl so that only an average is observed with = 1 2[a(H8 ) + a(HB )] = 21.0 G and a five-line quintet of 1 , relative intensities 1:4:52421 is observed. At intermediate temperatures > i-J the spectral changes are as indicated in the correlation diagram of Figure 9. The same set of spin states as was used in discussing CPCA radical [Figure (8)] may be used also for CHCA radical. The intensity of the next outer lines (lines 2 and 4 of Figure 14(3)), which vary with temperature, were measured relative to the invariant outer lines and the ratios F(m) at various temperatures are given in Table 10. The computer program ESREX4 was used to obtain values of log (l/T ) as a function of F(w) and the calculated values for cyclohexanecarboxylic acid radical in CHCA matrix are given in Table 9 and also plotted graphically in Figure 15. Values of log (l/T ) corresponding to the observed values of F(m) at various tempera— tures are given in Table 10 and are plotted graphically versus reciprocal temperature in Figure 16. The slope and intercept of the best straight line through the points of Figure 16 lead to the v values E8 = 3.7 kcal/mole and (l/T )0 = 1.7 x 1017 for the barrier / to the ring inversion process and the frequency factor, reapectively. 64 Table 9. Calculated values of relative intensities F(w) versus 10g (l/T) for l-cyclohexanecarboxylic acid radical in the self—matrix. L0810(1/t) F(w) +ll.O 1.98 10.5 1.95 10.0 1.84 9.5 1.55 9.0 0.93 8.5 0.21 8.0 0.12 7.0 0.81 6.0 0.98 Table 10. Experimental values of relative intensities F(w) at various temperatures and the corresponding calculated values of log(1/r) for 1—cyclohexanecarboxylic acid radical in the self matrix.-a “ V H T°C l/T°K(Xl0 °) 5(6) 16g(1/1) +22 3.39 1.73 9.75 + 9 3.55 1.68 9.65 0 3.66 1.57 9.55 -11 3.82 1.17 9.18 -26 4.05 0.88 8.98 -39 4.27 0.38 8.65 -56 4.61 0.23 8.53 —71 4.95 0.0 3The values of log(l/t) are taken from Figure 15. 65 l 2.0 F0») 0 C neat 0 COOH [1’5 intensities [1.0 Relative o [.5 80 [We “00] 0L0 020 L0910( 4) Figure 15. Calculated values of the relative intensity of the second line to the invariant outer line [F(w)] versus log(1/t) for 1—cyclohexanecarboxylic acid radical. 66 .xflpqu mHom .62 mesmae ecu :fi Hmoflwma meow owflxxonamoocmxosoHuxosfi How on\H mamao> mp\avwofi mo uoflm Ex 0:: .nb n.V_ o.v_ m6. 18 r: 1. O .D 1: mm 6 0. 0 M I no r: ace: 67 D. Cyclobutanecarboxylic Acid A solid solution of cyclobutanecarboxylic acid in adamantane was y-irradiated at room temperature. The ESR spectrum consisted of a rather weak quintet with a(H = 31.8 G [Figure 17(3)]. 8) Better spectra were obtained when cyclobutanecarboxylic acid was irradiated as the pure crystalline solid at 77°K. The spectrum on warming to —15°C [Figure l7(b)] consists of a strong quintet with a(H = 32.9 G and relative intensities about l:2.5:3.8:2.5:l indicating B) that the l-cyclobutanecarboxyl radical XIII, with four equivalent 8 protons, has been produced. There are also two or three additional, weaker lines present of unknown origin. Radical XIII decomposes above about -8°C while below —40°C the lines broaden and the spectrum becomes characterstic of a powder [Figure 17(c)]. No temperature dependence of the isotropic spectrum, which Would be indicative of a measurable ring inversion process, is observed between -8° and -40°C. E. Cycloheptanecarboxylic Acid in Adamantane A solid solution of cycloheptanecarboxylic acid in adamantane was y-irradiated at 77°K. The spectrum at 25°C [Figure l8(a)] can be considered to be made up of the superposition of two multiplets - a quintet with relative intensities 1:4:6:4:1 and a(H) = 13.4 G, and a triplet of triplets with a(Hl) = 17.9 G, a(H = 22.5 G. Both 2) multiplets have the same g value but they appear to arise from different radicals. The quintet disappears on cooling to -56°C [Figure 18(b)]. The triplet of triplets persists over a wide temperature range (-600 to +600C) with relative intensities and splitting parameters 68 206 coon h L 1 +25°c \— m“. \ -.::_ ___._ M5. 1:) ”W! Z {NAAL 38M“) 1M Figure 17. ESR spectra of 1-cyclobutanecarboxylic acid radical: (a) in polycrystalline l—cyclobutanecarboxylic acid, (b) and (c) in adamantane. 69 Cyclohepione *‘ carboxyHC acud radicai f +25°c (a) if] —56°C Figure 18. ESR spectra of 1-cycloheptanecarboxylic acid racical in adamantane. 70 : g ((3 .. . x111 a Planar Puckered R O. 5 , ‘:_. Chair XIV(a) a __:5 e 'sr- R e a Twist-chair XIV(b) (symmetrical) .._.3 V'— Twist-chair XV (unsymmetrical) Some Structural Formulae for Cyclic Carboxylic Acid Radicals (R = COOH) Transformation among the twist-chair forms of the seven-membered ring (XIV(b),XV) can occur by a process of pseudorotation in which the C2 axis moves around the ring one carbon atom at a time. 71 almost unchanged and might be assigned to an 2—cyclopheptanecarboxylic acid radical, XIV (chair) or XV (symmetrical or unsymmetrical twist- chair), in a rigid conformation with two pairs of nonequivalent B protons. The quintet, however, is not spectrum which would result from rapid interconversion between two nonplanar forms such as XIV or XV, Since the hyperfine interaction of 13.4 G is much smaller than the average of the triplet splittings, and the triplet of triplets never coalesces into the quintet. The origin of these lines remains unexplained. Pure cycloheptanecarboxylic acid was also v irradiated at 77°K. In the range -90°C to -l20°C four broad lines were observed with a(Hl) = 25 G, a(HZ) = 61 G and g = 2.0036. Above -90°C the radical disappeared. There is not sufficient information available at present to interpret this spectrum but it is presumably a powder pattern since isotropic proton hyperfine splittings as large as 61 G are not observed in cyclic radicals. F. Cyclopropanecarboxylic Acid in Adamantane A sample y irradiated at 77°K showed no ESR spectrum and it is possible that the acid did not enter the adamantane matrix. When the pure acid was irradiated at 77°K complex ESR spectra were obtained which could not be interpreted. II. Cyclic Ketones A. Cyclopentanone in Adamantane The spectra of solid solutions of cyclopentanone in adamantane y-irradiated at 77°K are shown at two temperatures in Figure 19. The spectrum at —24°C is a doublet of triplets with each line further (a) AV (b) 72 .. In . Adamantane O —24 C O —l93 C Figure 19. ESR spectra of 2-cyclopentanonyl radical in adamantane. 73 split into a triplet. This spectrum persists to low temperatures although lines 1,3,5 become broadened relative to lines 2,4,6 at -193°C [Figure 19(b)]. The Spectra indicate that the only stable radical produced is the 2-cyc10pentanonyl radical XVI , and O 1L e XVI that it is either planar, or is undergoing rapid inversion even at the lowest temperature, Since the Spectrum at ~193°C is still a doublet of triplets. The small additional splitting of 2.2 G observed above -60° is apparently from two equivalent y protons. The ESR parameters observed are g = 2.0046, a(Ha) = 18.5 G, a(HB) = 36.0 G (two equivalent 8 protons), a(HY) = 2.2 G (two equivalent Y protons).° No changes were observed in the spectra which would indicate that ring inversion is occurring at a rate measurable by ESR. B. Cycloheptanone in Adamantane A sample was x-irradiated at room temperature for two hours and spectra recorded immediately following irradiation, first at room temperature then at a series of lower temperatures [Figure 20]. Although lines from a second radical are present at all temperatures, a doublet of triplets attributable to the 2-cycloheptanonyl radical 74 ‘4 ‘d 4 200 —68°C- .4 (b) i v V V A " _146°C (C) Figure 20. ESR spectra of 2-cycloheptanonyl radical in adamantane (lines indicated by inverted triangles). The remaining lines arise from a second, unidentified radical. 75 XVII can be identified at +3°C and these six lines are indicated by XVII small arrows in Figure 20(a) and (b). The ESR parameters at +3°C are a(Ha) = 18.5 G, a(HB) = 28.8 G (two equivalent 8 protons) and g = 2.0042 i 0.0002. As the temperature is lowered the two central lines broaden and decrease in intensity [Figure 20(b)] until the ’coalescence temperature of -88°C is reached. At lower temperatures the two 8 protons appear to become equivalent and the five line Spectrum at -146°C can be assigned ESR parameters a(Ha) = a(HBI) = 19.1 G and a(HB ) = 38.2 G. These values would predict a high-temperature 2 average splitting for the B protons after they become equivalent of 19.1+38.2 2 28.8 G. The radical is very stable at room temperature. = = 28.65 G in agreement with the observed value of Values of the relative intensities of one of the inner lines (lines 3 and 4) to one of the outer lines for 2-cycloheptanony1 radical were calculated as functions of log (l/T ) using the program ESREX4 and are given in Table 11 and plotted graphically in Figure 21. values of log (l/t ) at the nqunatures at which relative intensities could be measured are listed in Table 12 and plotted versus 1/T°K in Figure 22. From this Arrhenius plot the activation energy for the ring 76 Table 11. Calculated values of relative intensities F(m) versus log (l/r) for 2-cycloheptanonyl radical in adamantane. Log (l/t) F(w) +ll.0 1.99 10.5 1.97 10.0 1.91 9.5 1.73 9.0 1.28 8.5 0.54 .8.0 0.0 7.5 0.35 7.0 0.71 6.5 0.90 6.0 0.97 Table 12. Experimental values of relative intensities F(w) at various temperatures and the corresponding calculated values of log(l/r) for 2-cycloheptanonyl radical in adamantane. T°C 1/T°K(x 10 3) 5(6) 10g(1/t) +68 2.93 1.66 9.34 +40 3.19 1.62 9.30 + 3 3.62 1.53 9.21 ~16 3.89 1.47 9.18 ~37 4.24 1.26 8.99 ~45 4.39 0.79 8.66 ~55 4.59 0.52 8.48 ~68 4.88 0.32 8.34 ~79 5.15 0.27 8.30 ~88 5.41 0.0 aThe values of log(l/t) are from Figure 21. 77 .ocmHnmaacw cw Hecapmu axcocmpmonoHoxoiu Hem fip\vaoH msmno> mafia House uemwnm>cw on» on o>wumfion mafia vacuum one new mowuwmnouaw o>fiuemon woumuoommu .HN unawaa 5:32.33 o.o~ one can c.s o.o _ ._ a - _ O O .o O 0 o i “5% O 1- o o . o O O 0 tell 6 l.oN .1 o 2.35.53 5 0 F382 Sconeuumsopuzud 1 u... S o o o as: O o r17. 78 an O;- 2-Cycloheptanonyl radical in adamantane e 1:"... —4 as ‘79. c» 3 “2.. co 0 l l l I .T 3.0 3.5 4.0 4.5 _3 5.0 llT°K Figure 22. Plot of log(1/t) versus 1/T°K fer 2-cycloheptanonyl radical in adamantane. 79 inversion process is found to be Ea = 3.2 kcal/mole and the frequency factor (1/1‘ )0 = 5.5 x 1011. The coalescence temperature for 2~cycloheptanonyl radical (185°K) is slightly lower than that observed fer 2-cyclohexanonyl radical22 (198°K) which implies that the activation energy for ring inversion Should be lower for 2~cycloheptanony1 radical as is feund (E3 = 3.2 kcal/mole for 2-cycloheptanonyl radical and E3 = AH* + RT = 3.70 + 0.60 = 4.30 kcal/mole for 2-cyclohexanonyl radical.22) III. Heterocyclic Oxygen Compounds A. Tetrahydrofuran in Adamantane Tetrahydrofuran (THF) in adamantane was Y-irradiated at 77°K fer three hours and ESR spectra taken at a series of temperatures up to ~10°C where it becomes unstable (Figure 23). The spectrum is a doublet of triplets which indicates that the radical is the 2—tetra- hydrofuranyl radical XVIII (a) and the ESR parameters at ~71°C are a(Ha) = 13.6 G and a(HB) = 27.7 G (two equivalent 8 protons).' At lower temperatures the two 8 protons become nonequivalent with = 37.3 G. a(H ) = 18.0 G and a(H 82) The intensity of line 3 relative to line 2 was measured 81 at various temperatures down to ~193°C which is approximately the coalescence temperature. The values of log (l/t ) versus relative intensity are plotted in Figure 24 and values from this graph used for the Arrhenius plot of Figure 25. The Slope and intercept of the straight line of Figure 25 lead to the values Ea = 2.2 r 0.5 kcal/mole and (l/t‘)o = 6.9 x 1013 for the ring inversion process. 80 1 —71°C [0). IN A Adamantane MN (°) 20 G (b) I [ —185°C (C) 1 -193‘t , i (d) + lines used for intensity measurement. Figure 23. ESR spectra of 2-tetrahydrofuran y'l radical in adamantane. 81 .mqmucmamwm :fi Hmoflwmn Hacmnsmo9vxnmuuouum pom mp\vao~ m3mhm> mnsvmu mowufimcoucfi o>fipmfimh mo mosfim> woumfisoamu .vm onswwm Apazsofl . .8: of cm. ‘3. o O O O O 0 [m3 0 O O N. 3 o T.. u. .0 o m .1 O Ill .9 o O o a o ocOEOEov< E A v... o o o 82 “lion/1’ ) 10.0 3' I I I I I I I 1 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 X 10 Figure 25. Plot of 103(1/1) versus 1/T’K for 2-tetrahydrofuranyl radical in adamantane. 83 B. Tetrahydropyran in Adamantane A solid solution of tetrahydropyran in adamantane was y-irradiated at 77°K and the ESR spectra collected at various tempera- tures. At very low temperatures the spectrum consists of eight lines of equal intensity [Figure 26(c)]. This is accounted for if the radical produced is the expected 2-tetrahydropyranyl radical XVIII (b) with a rigid conformation so that the B protons are not equivalent. The ESR e 0 *1 __s a "’ e XVIII(a) Z-Tetrahydrofuranyl radical XVIII(b) 2-Tetrahydr0pyranyl radical .parameters are a(Ha) = 16.1 G, a(HB ) = 40.0 G and a(HB ) = 5.0 G. 1 2 0n warming, four of the lines broaden and their peak heights become smaller as shown in Figure 26(b) taken at -90°C. At -11°C they have disappeared and only four lines of equal intensity remain [Figure 26(a)] with a(H ) = 16.1 G and a(H ) + a(H ) = 45 G. a 81 82 The spectrum analysis may be carried out in the usual way by writing the nuclear spin orientations for three nonequivalent protons Ha’ H = H , and H = H where H , H indicate "axial" and "equatorial" 81 a 82 e a e type B protons [Table 13]. The corresponding stick plot for the spectra of 2-tetrahydr0pyranyl radical are shown in Figure 27. 84 Table 13. Spin orientations for 2-tetrahydropyranyl radical under ' conditions of slow and fast ring inversion. Spin * ** State Ha Ha He (H ' He Low Temp. (H - He)High Temp. 1 - - - -k(aa + a3 + ae) (-aa/2) ' 33v. 2 — - + ' -%(aa + aa - ae) -aa/2 3 . - + - -3(aa - 8a * ae) ‘3a/2 4 - + + -k(aa - aa - ae) C-aa/Z) + aav. 5 . - - +1.02%ll - aa - ae) (+aa/2) - aav. 5 + - + +¥(aa - aa + 383' +aa/2 7 + + - . +k(aa + 8a ' ae) +aa/2 8 + + + +3Caa + aa + ae) (+aa/2) + aaV- * H a He + maaa + meae to first order, where He 15 the magnetic field at the center of the spectrum in Gauss, a“ is the a-proton hyperfine splitting, aa and ae are axial and equatorial hfs, respectively, ma,m are the proton spin quantum numbers, and + in- e dicates m a +8, - indicates m = -k. ** aav. 3 (aa + ae)/2’ ydropyranyl radical in 86 .2 033. mo moumum 5% .3303: on» on women whoa—5c o5. .onmucmamvm 5 H3233 , Hanukkmonvxnmhpopnm mo abhuuomm one :« mocwu How mowuwmcoucw can mfiOwuwmom vHoflm .nm 0H:m«m .:<<: m h V n 0 n N _ .mimh >>O._ _ _ _ _ — — — _ n3 _ I . I \ _ I _ / \ u \\ u _ / / x \ q / _ / \ _ \ _ /_ /\ _\ _ — / \ I \_ _ . _ / \ / \ _ IN / \ _ _ :2: m V o M n _ E .mzmh 10.1 — — fi — — , 0 o_v+ ou+ o owl VI _ . a _ a 87 Exchange of Ha and He by ring inversion would interchange spin states 2 and 3 and 6 and 7 of Figure 27 and lead to the changes in spectra with temperature observed with the eight line calculated spectrum of Figure 27 corresponding to the observed spectrum at -l42°C of Figure 26. The high temperature triplet of triplets of Figure 27(a) should be observed when ring inversion is sufficiently rapid to make the two 8 protons equivalent. However, the radicals decay before a high enough temperature can be reached and the spectrum of Figure 26(a) taken at -11°C corresponds to the intermediate region of Figure 27, where lines 2, 3, 6, 7 are sufficiently broadened by inversion to be unobservable. Values of log (l/T ) versus relative peak intensity for 2-tetrahydropyranyl radical were obtained using program ESRBX4 (Table 14 and Figure 28). The values of log (l/T ) obtained using this graph plus the experimental intensities at different temperatures are given in Table 15 and are plotted graphically in Figure 29. The activation energy for ring inversion is E8 = 3.7 kcal/mole and the frequency factor (l/r )o = 3.9 x 1011 sec'1 from the slope and inter- cept of Figure 29. C. Dioxane in Adamantane A solid solution of dioxane in adamantane was Y irradiated at 77°K and the ESR spectra studied over a range of temperatures up to about 0°C where decay of the radical becomes rapid. At -82° [Figure 30(c)] eight lines of approximately equal intensity were observed characteristic of a radical with three different proton splittings.The 1,4-dioxanyl radical VI appears to be the only radical 88 Table 14. Calculated values of relative intensities F(w) versus log (l/T) for 2-tetrahydropyranyl radical in adamantane. L0810(1/T) FCw) +9.0 0.50 8.6 0.09 8.2 0.01 7.8 0.17 7.4 0.47 7.0 0.73 6.6 0.88 6.2 0.95 5.8 0.98 5.4 0.99 Table 15. Experimental values of relative intensities F(w) at various temperatures and the corresponding calculated values of log (l/T) for 2-tetrahydropyranyl radical in adamantane.3, T°C 1/T°K(x 10 3) ch) log(1/t) -58 4.65 0.00 8.20 -68 4.88 0.38 7.52 -75 5.05 0.43 7.46 -85 5.32 0.56 7.29 -94 5.59 0.69 7.09 -102 5.85 0.80 6.88 -109 6.10 0.82 6.66 -120 6.54 0.96 6.05 -131 7.04 1.00 5.70 aValues of F(m) are from Figure 26 plus some spectra not illustrated. bValues of log(1/I) are from Figure 28. 89 .onmucmemvm a“ Hmuwwmh fixcwaQOHvxzmuuouum Mom np\~von mnmuo> .fisvm . on ohswfim :« m3onum x9 woumUfivcfi mocwa on“ we mofluwmcounfi o>wumfloh one we mozfim> woumfinoamo .wN onsmflm t :3 . 6.6 96. d 6.8.. _ ok— 0 O o o O O O O o vnv . ms 0 O O O O O 0 Nu e o o e e e m” .0 0:35:53 5 o as: m. - 90 6:35:83 5 H323.“ Eggnonvxnmnuouum now you.) m—x n.0— 0.0— 335503 5 o 0 mamnm> mp\HUon mo uon .mm oaswflu n.n_ 06— 0.: ‘ 0'9l 021 (.111) 601 1 0'8 91 .uo~wu we may bum .uoHoI um may .uoov- we flew ocmucmamvm :fi Hmuwvmu axnmxofiwuv.fi mo mhuoomm mmm .om onsmfim U ONQII uoovI Uopol A8 3V n3 0 ON Q r mZ .nsvm .xpflmcoqu o>wumaon mo mosfim> woumHsonu .Hm oH:Mwm Ans: 00.. . 3.6 a... I as a 6 o. In: .0 0. .0 .0 0. .0 0. .0 0. I. a. mu 0:253:01; E v. nu AllSNBlNI'TBU 94 L7-5- m in Adamantane [20 L0 Gm (1/7") [6.5 [6.0 [4.0 1 5.0 In K Figure 32. Plot of log(l/T) versus l/TOK for 1,4-dioxanyl radical in adamantane. 95 +25°c (a) d I S. in Adamantane (b) . 206 1 + -Iaa'c A (c) I ‘ lines used for intensity measurement. Figure 33. ESR spectra of 2-tetrahydrothiophenyl radical in adamantane. 96 58 9 88-: XIX protons), are consistent with this assignment. 0n lowering the tempera- ture the two central lines decrease in intensity, vanishing at the coalescence temperature, -104°C. At lower temperatures the B protons become nonequivalent with a(HB ) = 37.4 G,a(HB ) = 19.0 G,a(HY) = 4.8 2 G and a(Ha) = 15.0 G at -17l°CT (Figure 33(c)). At intermediate temper- atures the ratio of intensities of one of the central lines to one of the outer lines was measured. Calculated values of this ratio were plotted versus log (l/t ) [Figure 34] and values of log (l/T ) corres- ponding to the observed relative intensities were then plotted veg§g§_ l/T°K [Figure 35]. The slope and intercept of the best straight line of the Arrhenius plot [Figure 35] leads to the activation energy for ring inversion, Ea = 2.7 i 0.5 kcal/mole, and the frequency factor, (1/1 )0 = 9.8 x 1011‘sec-1. A sample of tetrahydrothiophene was distilled three times, degassed and sealed under vacuum in a quartz tube. After y-irradiation at liquid nitrogen temperature, ESR spectra were taken at various higher temperatures. At -102°C only very broad lines were observed and above -80°C these vanished so no radical products could be identified. 97 .ocmucmamvm cw Hangman H»Cm:aownuouvxnmnumunm How m9\~umofi mamn0> mm ouswwm :w m3owum x2 vmumowvcfl mocfifi on» we ABVm .mowuwmaoucfi 0>Mumflou can we mosam> woumfisoamu .vm 662nma TSSSEZ 92 .m o.w ed 06 _ J _ _ o o o o o o u 0 0 1| 6 .N. o 6 .1 o . 952252 5 e o o o o as: o o o 2 o o [.0 98 .ucmucmemwm aw Hmowvmn fixcocmoflzuoncxnmwuounm How MoH\H m3mho> mp\Huwo~ mo uon .mm ousmflm xhxa To: 3 cg o... a...“ _ _ _ _ mcoucmsmc< :. 1 0’6 99 B. Pentamethylene sulfide in adamantane A sample was y-irradiated at 77°K and spectra recorded \Nhile the sample was warming up. At -92°C an eight line spectrum characteristic of the 2-tetrahydrothiopyranyl radical XX was obtained [Figure 36] along with several unidentified lines. The ESR parameters 040.- r3 5 AW V0\a XX from this spectrum are a(Hsl) = 4.2 G, a(HBZ) = 42.8 G, a(Ha) = 18.1 G and g = 2.0042 1 0.0002. The radicals recombine rapidly above -80°C and have disappeared at -60°C. They are much less stable than the 2-tetrahydrothi0phenyl radicals (XIX) and spectra showing temperature dependence of line intensities could not be obtained. V. Heterocyclic Nitrogen Compounds f5. Pyrrolidine in adamantane A sample was y-irradiated at 77°K and ESR spectra recorded Inhile gradually warming to room temperature. The spectra at various 'temperatures between -190°C and -32°C are shown in Figure 37 and Tall appear to arise from the lOpyrrolidinyl radical XXI. At higher ‘temperatures (-32°) the four 8 protons are equivalent with #8? :7 370 XX 100 .06Nmu pm ocwucmamnm cw awofivmh axcmnxmoficuOHuxnmnuouum mo enhuoomm mmm Uomol ocean—«52 5. g . .om ousmfia 101 l S i" Adamantane N WW II»; II],,IM"‘3 [1 (a) d) ‘ [ —163C (8) l 490 c f) Figlare 37. ESR spectra of l—pyrrolidinyl radical in adamantane. Lines used for measurement of relative intensity are indicated by arrows. 102 a(HB) = 40.1 G, a(14N) = 14.5 G and g = 2.0048. At very low tempera- tures (below -173°K) the B protons appear to become nonequivalent. with two having a(HBI) = 27.0 G and two a(HBZ) = 54.2 G since these parameters, along with the value a(14N) = 13.5 G, account for the positions of the lines of Figure 37(f). The temperature dependence of the spectra has been employed to study the ring motion. The intensity of the central line of the next-to-outer triplet relative to the central line of the outer triplet was measured at each temperature (the mI = :1 lines of the nitrogen triplets change intensity relative to the mI = 0 line and so could' not be used) and values are given in Table 16 along with the calculated values of log (l/T ) corresponding to the measured relative intensities. 'The calculated values of log (l/T ) for various relative intensities Ivere calculated by use of ESREX4 and are given in Table 17 and plotted graphically in Figure 38. An Arrhenius plot log (l/tc) versus l/T°K i4; shown in Figure 39 and leads to an activation energy for ring iJrversion, Ea = 2.9 kcal/mole, and (1/1 )0 = 6.5 x 1012 sec-1. This radical (XXI) is stable to -10°C. A sample of pyrrolidine in adamantane irradiated at 25°C gives the spectrum of another radical. li;___JPiperidine in Adamantane A sample was X-irradiated at room temperature and the sP‘BCtrum recorded immediately [Figure 40]. The basic spectrum, neglecting some unidentified lines, is a doublet of doublets resulting fiT>fll two different splittings by single protons each line of which is fhxrther split by interaction with 14N into triplets. By analogy 103 Table 16. Calculated values of relative intensities F(w) versus log (l/r) for l-pyrrolidinyl radical in adamantane. Loglo(1/T) F(w) +11.0 1.98 10.5 1.94 10.0 1.82 9.5 1.49 9.0 0.83 8.5 0.16 8.0 0.04 7.5 0.37 7.0 0.72 6.5 0.90 6.0 0.97 5.5 0.99 Table 17. Experimental values of relative intensities F(m) at various temperatures and the corresponding calculated values of log (1/1) for 2-pyrrolidinyl radical in adamantanefi: T°C l/T°K(x 10 3) 5(6) 108(1/1) ~51 4.50 1.80 10.0 ~94 5.59 1.29 9.32 ”110 6.13 0.86 9.04 “120 6.54 0.38 8.70 ‘132 7.09 0.20 8.41 ”143 7.69 0.0 8.00 ‘163 9.10 0.35 7.41 ~173 10.00 0.55 7.21 \ aValues of F0») are from Figure 37. b Values of log(1/'r) are from Figure 38. 104 .ocmucmamwm :fi Hmowpmu axcprHOHhmeH how me\HVmoH mzmuo> um unawam cw mzohnm An woumoflwcw mocfia mo .navm .mofluwmcoucfi o>fipmaon 0:» mo mosfim> voumuaufimu .wm onswflm . o . A t: 53 o :_ o 0: od— od— OK. 0.0— m o o o o o w. o Imu mV 0 as mw M. o a o o .1...I .u... . a o nu “w m a s 0 ll 0 .9 0:252:on E o o as: O'z 105 no" x .o:mu:mamwm a“ Hmuwumu chfivwaonwmeH mom xob\H mamuo> ap\HumoH mo pon .mn unamfim . xh\fi OH”: o.oH_ o.mb o.m— o.n_ o.c_ o.m_ N .0 .n... r' 8 mu w 335.52 E 0 fl“ 1. . .m "U \l H .a _ l\ 106 .Uomm we. 0:35:83 5 302.2 icwvwnogmnm mo Efiuuomm «mm J. :3 00a 6.8.. t ocoEoEop< _ .ov unease 107 with the temperature-dependent spectra of the 2-tetrahydropyranyl radical [Figure 27 and Table 13] this spectrum can be assigned to the 2-piperidinyl radical XXIIin an intermediate temperature region XXII where the ring motion leads to low intensities for certain lines. The ESR parameters a(Ha) = 13.7 G, a(14N) = 3.4 G, a(HB ) + a(HBZ) = 41.9 G and g = 2.0030 were obtained on this assumption.1 Microwave saturation studies indicate that lines from a second radical are present and the activation energy for ring inversion could not be <>btained because of overlapping lines.‘ VI. Other Alicyclic Compounds .51: Cyclopentyl Chloride in Adamantane A sample of cydbpentyl chloride in adamantane was X-irradiated at room temperature and the spectrum recorded immediately [Figure 41]. It may be analyzed as a quintet of triplets each line of which shows a small doublet splitting and the ESR parameters a(Hl) = 2.7 G, E1012) = 14.3 G (two equivalent protons), a(HS) = 22.2 G (four equivalent PrOtons) and g = 2.0029 indicate that the radical is the cyclopentenyl rad ical I . 206 —>,.<— H “H 2.7G {—22.2911 II [I Figure 41. ESR spectrum of cyclopentenyl radical from a sample of cyclopentyl chloride in adamantane irradiated at room temperature . H2 2 O (3 Q l.____l H H H + 2 5° c Figure 42. ESR spectrum of cyclopentenyl radical from a sample of cyclopentyl bromide in adamantane irradiated at room temperature . 109 B. Cyclopentyl Bromide in Adamantane A sample X-irradiated at room temperature gave a spectrum [Figure 42] very similar to that of the cyclopentenyl radical in Figure 41, and the ESR parameters obtained are almost identical, with a(Hl) = 2.7 G, a(HZ) = 14.3 G (two equivalent protons), a(HS) = 22.2 G (four equivalent protons; and g = 2.0027. C. Cyclopentane in Adamantane A sample irradiated at 77°K gave the spectrum shown in Figure 43(3) at —121°C. This is a doublet of quintets characteristic of cyclopentane radical (XXIII) in the region of fast ring inversion with a(Ha) = 21.42 G, a(H ) = 35.6 G, and g = 2.0031. At -187°K the B spectrum changed to that of Figure 43(b) indicating that ring inversion is slow with Ea = 2.2 kcal/mole and (l/T)O = 1.46 x 1012 sec'1 found as . " j = : above with a(Ha) - 21.1 G, a(HBl) 25.2 G and a(HBZ) 45.6 G. D. Cyclopentanol in Adamantane A sample X—irradiated at room temperature gave a complex room temperature ESR spectrum which must arise from two or more radicals (Figure 44). The values a(HB) = 27.6 G (four equivalent protons) and g = 2.0033 indicates that one of the products is radical XXIV ‘undergoing rapid inversion at room temperature. The low-temperature spectra could not be interpreted so no information concerning the ring inversion process could be obtained for l-hydroxycyclopentyl radical. E. l,2-Dichlorotetraf1uorocyclobutene in Adamantane A sample was Y-irradiated at 77°K and the spectra observed at various temperatures. At -50°C the spectrum of Figure 45(3) was obtained and analysis of the spectrum by computer simulation using ESR was 110 .aoumumscm mama champ .U.HNH- new vauN .coflmuo>:w mcfin 30am .Uonwdu any "ocmucmamvm cw Hmofivmh qucomo~oxo mo asuuuomm «mm .mv ousmflm E 111 .mHmuwven guano on wousnwuuum ohm mocaa wnficamaou 0:» mafia: zones ecumUflvcw one peasant may we m:bwufimom mafia och .06mm we ocmunmauua ca fiaowvau “ApnomoHoxuxxonvxcua 0:» mo aunuoomm «mm .ve meanwm _ u all.“ 112 ‘a’ .70 G IdI Figure 45. (a) ESR spectrum at —60°C of radical produced by irradiation of 1,2—dichlorotetrafluorocyclobutene in adamantane, (b), (c), (d) simulated spectra using the parameters of Table 18(b), (c) and (d), respectively. 113 XXIII o e ._.L v. e 9 a H XXIV F F XXV c. 9‘. F Cl F XXVI Some Structural Formulae for Cyclic Radicals and Radical Ions 114 attempted. The observed relative intensities of the outer four low- field and high-field lines are nearly l:2:3:4 and suggest strongly that these are part of a septet arising from two equivalent chlorine atoms with a(Cl) = 4.1 G and predicted relative intensities 1:2:3z4: 3:2:1. The fluorine hyperfine splittings cannot all be equal, nor are there two pairs, since the overall pattern suggests a quartet rather than a quintet. We have therefore considered several cases for spectrum simulation. These include (a) three equivalent fluorines and two different chlorines, (b) two equivalent fluorines and two nonequivalent fluorines plus two equivalent chlorines, (c) four nonequivalent fluorines and two equivalent chlorines, (d) three fluorines of one kind and one of another plus two equivalent chlorines and (e) three different fluorines and two different chlorines. ESR parameters giving the best overall fit of calculated to observed spectra are given in Table 18 (b), (c) and (d) for each of the above cases (based on structure XXV) and simulated spectra obtained with these parameters are shown in Figure 45(b)-(d). A11 agree within experimental error with the experimental spectrum of Figure 45(a). The sets (a) and (e) of Table 18 correspond to a radical C-C4F3C12 (XXVI) formed by loss of one fluorine while the other sets are for a radical cation 0r anion c--C4F4Cl2 (XXV). Although the latter reproduce the experimental sPectrum better, it is possible to obtain a similar fit by allowing the relative intensities in cases (a) and (e) to deviate from the theoretical values. Since internal motions in the radical could cause SUCh deviations, XXV must be considered a possible structure. 115 Table 18. Some sets of ESR parameters giving calculated Spectra in good agreement with the observed spectra for the radical in irradiated c-C4F Cl . * Case a(Fl) a(Fz) a(F3) a(F4) a(Cll) a(Clz) (a) ' 36.9(3)+ 4.1(1) 8.20) (b) 40.0(1) 32.6(2) 15.8(1) 4.1(2) (c) 40.0(1) 36.6(1) 28.6(1) 15.8(1) 4.1(2) (d) 36.4(3) 11.7(1) 4-1(2) (e) 69.7(1) 28.7(1) 12.4(1) 4.1(1) 8.2(1) * The four cases correspond to those in the text and the first three to simulated spectra in Figure 45. 1.. All values in Gauss. Numbers in parenthesis after each hyperfine splitting are the numbers of equivalent atoms with that value of a. 116 DISCUSSION 1. Some Properties of the Adamantane Matrix Adamantane (XXVII)has been shown to provide a rigid, nonpolar matrix with spherical holes so the radicals, which substitute for} adamantane molecules, can rotate freely and give isotropic ESR spectra. The radicals tend to be stable for several hours at room temperature and spectra can be obtained over a wide temperature range. In order 1 r2 )ht XXVII to obtain isotropic spectra with good signal-to-noise ratio the precursor must dissolve sufficient adamantane and the radicals produced must be able to rotate freely in the holes. There is apparently a maximum size mOIecule that can be accommodated in the matrix. In this work it has beer! feund that l-cyc10pentanecarboxylic acid radical gives excellent 5Pe<:tra but 1-cyclohexanecarboxylic acid radical is apparently too 181‘ge to rotate freely and only anisotropic spectra are obtained. With straight chain radicals six or seven carbon atoms is oftZen the upper limit for radicals which can be accommodated.1 There 13 also a lower limit, since small radicals can diffuse and recombine, and this may explain our failure to observe radicals from 1,2-dichloroethane. :I“»Iother cases, particularly highly polar molecules such as Z-pyrrolidone, 117 piperazine, pyrrole and 2,4-pentanedione, the precursor does not appear to dissolve any adamantane. Adamantane itself yields 2-adamantyl radicals on irradiationgs’96 but not 1-adamantyl or any other radicals. If impure adamantane is used the impurities give ESR spectra and these have often been mistaken fer adamantyl radicals.97 Where the radical precursors employed yield radicals of greater stability than 2-adamanty1, the 2-adamantyl radicals disappear, probably by abstracting a hydrogen atom from the precursor, and only the spectrum of the radicals from the precursor are observed. However, 2-adamanty1 would not normally be observed in this research since it gives only a broad doublet which would be almost completely discriminated against in the second-derivative presentation used here. Hydrogen atoms appear to be produced from both adamantane and the precursor molecules on irradiation. They diffuse readily and ‘transfer the radiation damage. Larger groups do not appear to be lost Probably because of the cage effect. Thus, C-C and C-N bonds are not umullly broken and groups such as carboxyl (in cyclopentanecarboxylic 891d) are retained. While chlorine and bromine are lost from the cY'<:lopentyl halides, this may be the result of splitting out HCl or "Br from the radical first produced (by loss of one hydrogen) followed by diffusion of the small molecule out of the cage. The unusual stability of most radicals observed here must JPfisult from the slowness of diffusion out of the cage of neighboring aslamantane molecules and also stability of the radical with respect to 2~adamantyl radical which could always be produced by hydrogen abstraction 118 if it were favored. l-Cyclopentanecarboxylic acid radical was found to be particularly stable, disappearing only at 75°C in adamantane (but 5°C in the pure acid matrix). Several radicals could be studied up to 50-80°C and disappeared only slowly even at these elevated temperatures. 0n the other hand the radical from 2-pentamethylene sulfide disappears even at -60°C and may be an example of a radical which is less stable than 2-adamantyl and so is labile in the adamantane matrix. Where the odd electron can be delocalized over several atoms, as in the cyclic carboxylic acid radicals, the radicals may be more stable. When the precursor itself forms plastic crystals, freely rotating radicals may sometimes be obtained on irradiation of the pure solid at tmperatures above the transition point at which free rotation of the molecules occurs. In the present investigation isotropic spectra for the radicals from cyclobutanecarboxylic acid, cyclopentanecarboxylic acid and cyclohexanecarboxylic acid were obtained by irradiation of the pure solids. II. ESR Parameters for Cyclic Radicals Sixteen alicyclic radicals, which have been produced in thius investigation by y-irradiation of alicyclic compounds in an “Guantane matrix, are listed in Table 19. Where the B proton hyperfine interactions shew a temperature dependence, the values are given at both the lowest and highest temperatures attained. Dihedral angles expected for ‘3 Irrotons for various "ideal" conformations are given in Table 20 and the I)ll‘eferred conformations of the radicals studied, based on the observed values of a(HB), are listed in Table 21. Table 22 summarizes the experi- It‘ental values of the barriers to ring inversion determined in this work. 119 emua an .o .. .vfloa owaxonuou noceucomoAuxo mo uAmm ssfieom u mzo .I .oooo.eu ma HA ouocz Amuwemu ero owaxonumoocmpnnvoAouA pom amouxo mooo.o+ whouho oAnonoum o>mn m we mosAm>n .momoo umoa :A u m.QH machho oAnoAOHQ o>on use mmzmo :A one x903 mane :w m we mosAm>e c.4A . Ne < A.m 6 AN n.4A Ape: mAge < mNoo.N A.N AeovaN.NN AAcmocoavoAu Ago: mAce < Nnco.N A.NN Au.oo-ve.mm . - A.me AmNVmN A NN AeoomA a w.NN AAucmavoAu AA63 mAge < mmoo.N Aeovao.AN AAucoaoAuA6-A-AoneAm-A mAaermm 96:60 «.mnszAva ago: mAae a omoo.N A.mA a.AeuNm+Am oAc.oeV AAaAeAnoaAa-N a Na . Aa ooconomom xAuumz pm Aaosuovm AA mumv A :Vo A mum Aazva «manna Ammva Aaeram 8223 AA e3 123 voA Nem.oaa asAx «6 mu.vu MA com. com AAegu Annoucoo Nam 039v mmewm vouonaoaueo>om an: om mm co. co AAeoAuuoaaAmesu nweguuumwzp Nn omAI 6mm AAeoAuuoaaAmv Haezuuumwzp .oonn_ cam- «eon no uwenu Anoueoo New oeov munflm wouonaoaueo>om mm omA ome .AAeAAoIM Ae: Amuoucoo New ozuu owned monogaoauxflm o co peonuumfiza on- om peom on: om Hwenu Auouaoo New oaov madam mononaoauxfim oNA .on a c erce eeAuoxuaa o co Amos oevo>em cm on neceAm Amuouaou New ozpv mwcfim eohonaoauo>flm mm oA om ANuVmAAeao-pmAze mm .oA-..oA no>Ae mm.eom o co A no oevo>=m . on on neeeAm Ahoucoo New ozoV macaw vononaosuo>fim DNA oON u c oAmee meAAonsa .¢.oA oe.en eoauxusa ONA .oA u c eree eeAonosa om.AA om.Am ecaoxusa on.e~ on.v~ neeeAm AHoucoo New ocov muewm vononaoauhaom ooeouomom ucoaaou No Ac nowueahomeou meowueahomceu AeoeA oaom cw mAeufieem new meouenm m E 335. A323 .cN canes 1224 etc: «as» NN 966: maze - 4pc: ”Ash mm mN mA 906) .Age .661 Asap 0A ~62~g6-cA.z soptouon 625:6-CA8: oavo>co chuuo>cA er_amm unceAm umcgu z».0uon eccAACE AeckouOu omueA £uw3.u~ecu-umA)h pfiezu pompoAon useLUIwAe: vouoxuza chuuo>cA AApAnex oucocouoa coAueEuoucou eocuououn oAcenonm m.mv Ao.evv Am.AvV A.cm Am.cew Am.oma m.mw one .onA om.o omv .oAA o« o com .6. 1A.: .04 ..oN po.o 6cm .oen ea,c OCR .OCAI WHIU .AA .omA mm 9 .mcoduneeeucA ocemuonxz :ouowm ache maceweea uA.uxu mo ncoAueeACmcou A.44 .«.cN 0.0%. n Amv 6 AA - .1v Anec.-v A.0m .c or diam 0 an... L “I“ H AUV o.ov.o.m A.oo-v e.mA A.AeA-Vo.mA .A.mA co.om . Amy .Amxva. A A Axe. .A m:.. mm.3 mm.o mo.o vm.0 va.o ll) .mA .AN oAnek 125 nonhmlv to: :E. 2.5-2.: 3 .2 (AA 36 Ni. .92 x; “.2 . A227 < . \II/sn-. av RACLU fibHLOwVAC m.mm 00m .0" hN.O m.mn .m.0 Qh.O m.h~ :.~Om \AI HONOIV Atom ”Age uAazo o.mv .AA ..nA- oA.o e.Nv .A.v oe.o A.eA < . A.mNe m.mN to: fifi 3.6-2.: v.2 .2 .2 8.3.275 . S 6.2 :5 0.2 < / AONQIV \|.I is. 2.; 2.5: 8:685 99. 2.. .oN- 2.0 9% .58 2.0 AAA < O ..\ - . . N A $52.83...“ ALA g o me u m. u A65. .HAA e6 eAagu eoetonta N.Av .oA..oA- AA.o A.N.A-a c.04 .o.m NA.o A.oA a A.AA-. A.AN _\ Ate. «ALA AA.;6-8A.: o.mm .wv .oNA ev.o A.nmA-u A.Am .o.eA 36.9 o.mA < my, ea eAeeo .Am ..m an o No.A .me N c.Aom .v AoOmIv for .25. 353.313 notoAmA; v.2 6mm .6mm owe m.mm 5.: < noCOIv to: wit. .AAezu votoAmE m.A.m omo .17 1.0 0.5” .o.m 1: qum Ace: "Age AA.20-6.A:6 eocoAA.Aa ..A. .AA ..mN Ae.o o.~n .n.eN on» ecu euev woo mama: eucmeuno ouo) m use Nc.Ao mo noaAe> onrv .ONA oocouowom cm vocAwov me mA c mcwuoxusa we oAmc AAmzvev mo moaAe> mo zaewm n aowm coxeu moueeAumo Q . one a .~o .Ao mo moer> .oAneAAe>e WA AAmzvev AACO egos: .600 I mm + Am :oAumssmme ocu vce Amzve n onou m :OAuezvo ozu 90km N A m eca No .Ao AA e:V~\A axe. . N0N.66\AoNnou umNNAAaEa . .6 AAV ”aonAoe a. easemeno 66»: 6Anep wage as wauuuuceac noeeAsuAeu aged 7. v0.0 n .VAIN: I xuor «Ack ueceAm N.N~ I nmv NA.oI A.~IA: < v. expo: “Age eAazo N... .oa...A- mNo.o AN I .uv Ao.o A.AA < q. 00c0u0u08 gauEOMA—OU a 50v Aucxvd Ann—.OIV 00 H0700 Nun—uni "00:3“ uouuououn oAaenous no.~o AAozve ANuxce.AAo:ue AeoseAucooe AN .Aaap 127 mm - me.m =.Aom NN - A+==A mafia How .oAp\Av .mnouoem Aocoseonm woe .em .mowmnoco :oAue>Auo< .NN oAneh .ere owaxonueuAeoceusnvoAo u = §-= 26.87 G B) or B = 53.74 G. It has been shown11 that delocalization of the odd electron, which decreases the odd electron density in the 2pTr orbital of the trivalent carbon, leads to lower values of B. Thus, for the radicals XXVIII and XXIX the B-methyl hyperfine splittings44 are . N XXVIII XXIX 19.0-19.5 G and B = 38-39. In the radicals studied here values of B are generally in the range 40-48 G or, if Equation 15 is employed along with the value B = 58 Gll,pa z 0.7 - 0.8 for these radicals. a(14N) is smaller, usually in the range 1.7—5.7 G (Table 6). Nitrogen hyperfine splittings, when they can be resolved, are useful in assigning radical structures. When the odd electron is centered on a nitrogen atom in a fl-electron radical, a(14N) = 14.0-15.0 G, ‘whereas if the odd electron is centered on an adjacent carbon atom a(MN) is smaller, usually in the range 1.7-5.7 G (Tab1e6 ). 132 The measured isotropic g values given in Table 19 are generally somewhat larger than the free spin value (2.0023). This is attri- buted to spin-orbit interaction and values of Ag = (gobs - 2.0023) become larger when delocalization of the odd electron onto neighboring atoms, particularly heavier atoms, becomes more important. Thus, for planar n-radicals the succesSive substitution of hydrogen by an atom with an unshared pair of electrons, at positions where the spin density 94 The is appreciable, cause a progressive increase in the g-factor. g-factors are often useful in the identification of radicals. For example, the values Ag = 0.0025 for the radical from irradiated pyrrolidine and Ag = 0.0007 for the radical from piperdine indicate that the odd electron is centered on nitrogen in the former case but carbon in the latter case. The very large value Ag = 0.0025 fer 2- cyclohexanonyl radical has been shown53 to result from extensive delocalization of the odd electron onto oxygen - 25% as estimated from the 170 hyperfine interaction. In otherwise similar radicals Ag should be larger when delocalization is onto a heavier atom; Ag = 0.0021 for 2-tetrahydrothiophenyl radical compared to Ag = 0.0011 for 2-tetrahydro- furanyl radical. However, an Opposing decrease in g occurs when the odd electron orbital acquires 5 character and the radical becomes nonplanar.11 III. Alicyclic Carboxylic Acid Radicals l-Cyclopentanecarboxylic acid radical (XI) at ~193°C gives an ESR spectrum (Figure 7(h)) which can be analyzed as a triplet of triplets with hyperfine splittings of 22 and 41 G. It is reasonable to assign these to two nonequivalent pairs of B protons corresponding 133 to a cyclOpcntane ring frozen in a single conformation on the ESR time scale. The room temperature quintet can then be assigned to four equivalent 8 protons resulting from rapid exchange between two equivalent conformations and the fact that the quintet splitting (31.3 G) is just the average of the two low~temperature splittings tends to confirm this intepretation. It will be assumed here that the radical at low temperature has either the "envelope" or "half-chair" conformation68 and that exchange between the two equivalent conformers X18 and XIb occurs at room temperature. The dihedral angles 01, 02 between the direction of the odd electron 2p orbital on carbon and the B—C-H bonds may then be calculated from Equation 16. The simultaneous equations IQ l\.) Bcosze1 = Bcosze7 = 41.0 G .0 G 0 + 6 = 60° lead to 01 = 45°, 62 = 15°, B = 44.0 ( Figure 4). Gilbert and Trenwith33 have shown that for cyclic five—membered ring radicals in solution the ESR data are in better agreement with the C2 ("half-chair") conformations than with the CS ("envelope") conformations. Assuming no distortion from tetrahedral angles or from the ideal symmetry, are 100 and 500 for the C radical and 2 2 -100 and 700 for the CS radical. Our data agree better with the C2 con- the calculated values of 01 and 0 formations for the 1-cyclopentanecarboxylic acid radical although the probable errors in such a calculation are rather large. 134 When pure cyclopentanecarboxylic acid was irradiated at 77°K the overall behavior of the spectra was similar but the hyperfine splitting constants at low temperature were more nearly equal with a(H = 28.3 G, a(H = 34.0 G, and (at 0°C) = 31.1 G. ) ) B1 B2 0 From these one obtains 01 = 34 , 02 tion 16. These values are much closer to the values 01 = 02 = 30 = 260 and B = 41 G by use of Equa- 0 expected for a planar radical than are the parameters for the same radical in adamantane. The molecule is thus less twisted from planar 1 + 02 = 600 may not be correct for cyclopentane derivatives but the H-C-H angles do not in the self matrix. Also, the assumption that 0 appear to have been determined precisely, the usual assumption being that they are tetrahedral. The energy barrier restricting ring inversion in l-cyc10pentane— carboxylic acid radical (XIa Z XIb) is found to be E3 = 1.43 kcal/mole in adamantane. This is less than that found for cyclpentyl radical itself,31 also in adamantane, E3 = 2.7 kcal/mole. In cyclopentane it has been estimated that the planar and puckered forms differ in energy by 5-6 kcal/mole7s’76 but there are no definite energy minima and maxima (such as XIa, XIb) since the puckering can move around the ring by pseudorotation with a change in internal energy of the molecule of the order of RT (0.6 kcal/mole at 25°C). The l-cyclohexanecarboxylic acid radical appears to not be pro- duced in adamantane, possibly because of the steric factor. However, it is observed in the self matrix with low-temperature splittings a(HB ) = 37 G, a(HB ) = 5 G and high-temperature (+90C) splitting 1 2 135 = 21.1 G. These values lead to dihedral angles (Figure 4) 01 = 69°, 02 = -9° and B = 39 G. The conformational change is presumably a chair I chair inversion as in cyclohexane itself (XIIa I XIIb). The energy barrier restricting ring inversion is found to be Ea = 3.7 kcal/mole which is somewhat less than in cyclohexyl radical where E3 = 4.9 kcal/mole.27 The "coalescence” temperature (at which the high—temperature quintet Splits into the low—temperature triplet of triplets) is also lower (-600) for l-cyclohexanecarboxylic acid radical than for cyclohexyl radical (225°C). The decrease in inversion barrier in going from cyclopentyl to l-cyclopentanecarboxyl and from cyclohexyl to l-cyclohexanecarboxyl may result from delocalization of the odd electron onto the carboxyl groups of the acid radicals with resultant partial double bond character in the C-COOH carbon-carbon bond. This would tend to make the acid radicals planar at the trivalent carbon whereas the hydrocarbon radicals are believed to be nonplanar. Thus, the chair : chair interconversion in cyclohexyl radical would be accompaned by an inversion whereas in 1-cyclohexanecarboxyl it would not. At present the only carbon-l3 splitting which has been reported for any of these radicals is the value a(13C) = 41 G for cyclohexyl74 and this suggests that the radical is bent a few degrees from planarity. If K0 is taken to have a value midway between that found for planar methyl radical (38 G)10 and that calculated for a hypothetical planar tertiary radical (27 6)}03 the use of Equation 17 would lead toEl= 3.4° for the bending from planarity in cyclohexyl. 136 For l-cyclobutanecarboxylic acid radical (XIII) only an average B-hydrogen splitting, = 31.8.6, could be observed and no conditions could be found which would give further splitting of the lines corresponding to freezing out one ring conformation. The same radical in the single crystal matrixss’39 l,l-cyclobutanedi- carboxylic acid gave = 32 G at 300°C and at low temperature one ring confbrmation was frozen out so that a(HB ) = 36.4 and a(HB ) = 27.8 G at 77°K. Apparently even at the lowest :emperatures accesgible in adamantane the ring inversion remains rapid on the ESR time scale. The results for y-irradiated cycloheptanecarboxylic acid in adamantane are rather difficult to interpret since several radicals are produced but it does appear from the spectra that the l-cycloheptane- carboxylic acid radical (XIV,XV) is one of the products and that it is in [a rigid confermation at all temperatures employed in this work. Ohmae gt_§l,25 found a similar result for the cycloheptyl radical in irradiated solid cycloheptane with two different 8 proton hyperfine 81) = 24 and a(HBZ) the liquid at -6°C only a single 8 hyperfine interaction, a(Hb) = 10 splittings, a(H = 20 G, at -l96°C. However, in 24.7 G, is found. Also, the cycloheptanone ketyl (XXX ) is observed66 in a single confbrmation between —40° and +88°C in dimethoxy- methane solution with a(HB ) = 7.02 G and a(HB ) = 2.08 C. These 1 2 results are surprising since cycloheptyl derivatives appear to inter- convert rapidly on the NMR time Scale with values of Ba < S kcal/mole.104 The radical structures are undoubtedly more rigid and show that introduction “‘0’ w XXX 137 of sp2 centers makes 5- and 7-membered rings less flexible. The rather low values of the B proton hyperfine splittings found for 1-cycloheptanecarboxylic acid radical, as well as for the other cycloheptyl radicals reported in the literature, have not been explained satisfactorily. It appears that for these radicals either the value of B is smaller than in other cyclic radicals or the spin density on the a-carbon is smaller. If we assume the former,and use 8 = 27.4 along with the observed 8 splittings of 17.9 and 22.5 G for l-cyclo- heptanecarboxyl, the values 01 = 25°, 02 = 35° are obtained. These are approximately the values to be expected if the radical has the chair conformation XIV with the radical center in the l-position. The twist-chair forms (XV) which are favored in theoretical calculations on cycloheptane itself105 lead to quite different angles with the radical center in the l-position and do not appear to be the stable 0 ferms for the radicals (01 = 75 , 0 = -15° for symmetrical twist-chair). 2 IV. Cyclic Ketone Radicals The 2-cyclohexanony1 radical (IV, V) in adamantane has . 22 53 . . . . . been studied ’ and the activation parameters for ring inver51on between two half-chair forms found to be AH* = 3.70 kcal/mole, E = AH+ + a + RT = 4.3 kcal/mole and AG + = 4.39 kcal/mole. NMR studies = 5.2 kcal/mole for the ring inversion + of cyclohexanone106 lead to AG process in that molecuk§;compared to AG = 10.6 kcal/mole in cyclohexane. Pratt gt_§132 conclude that the introduction of one sp2 center into cyclohexane, as in cyclohexanone or cyclohexyl radical (Ea = 4.9 kcal/ molelo) lowers the barrier by about one-half. Introduction of a second sp2 center in 2-cyclohexanonyl radical produces an additional lowering 138 of only about 1 kcal/mole and this is attributed53 to the allylic-type resonance in the radical which gives the C1-C2 bond about 15% double- bond character and so tends to make atoms 1-2-3—6 of the ring coplanar and give a structure already halfway through the conformational change of cyclohexanone. It is not surprising then that 2-cyclopentanonyl radical XVI interconverts rapidly on the ESR time scale even at the lowest temperatures studied here and that the two 8 protons remain equivalent. Cyclopentanone has been shown107 by refined force field calculations to exist in the half-chair form with an angle of twist of the C -C 3 4 bond out of the C5-C1-C2 plane of about 24°. The barrier to inversion must be less than 5 kcal/mole since it has not been detected by NMR. An additional sp2 center, plus the contribution of allylic resonance (XVIa‘S XVIb), will tend to make C -C -C -C coplanar and so give a 3 2 l S H O -—A 1‘ <7“ cyfi";’llll\ XVIa XVIb ground state confbrmation already halfway through the conformational change accompanying the inversion of cyc10pentanone. The radical would then either be planar or have the envelope conformation XI with C4 at the "tip of the flap" and a low barrier to inversion of C4. The B proton hyperfine interaction = 36.0 G is comparable to the averaged values found in the rapidly interconverting cyclobutyl (36.6 G), cyclopentyl (35.2 G) and 2-cyclohexanonyl (33.0) radicals. 139 The ESR spectra of 2-cycloheptanonyl radical in adamantane show temperature dependence which indicates a ring inversion process. The coalescence temperature for transformation from two nonequivalent B protons to two equivalent ones occurs at a lower temperature (-99°C) than the corresponding change in 2-cyclohexanonyl (-72°) indicating a lower barrier to inversion (Ea = 3.2 kcal/mole versus Ea = 4.3 kcal/ mole) for 2-cycloheptanonyl. The most probable structure for cycloheptanone is a half-twist-chair form with the carbonyl group at C2 and the odd electron on C1. The carbon atoms carrying the odd electron and the carbonyl group, and those on either side (7,1,2,3, of XVII) will tend to be coplanar. The remaining three can then take either equivalent conformation of XVII and the inversion barrier 108 However, because of the observed will correspond to this change. flexibility of the cycloheptane ring it may be possible to go from one form to the other via low-energy paths of the type giving rise to pseudorotation and so have a barrier lower than that in 2-cyclohexanonyl but averaging the H8 splittings in a complex manner. The observed values of a(HB) lead to the values 01 = 46°, 02 = 14° and B = 39 G (using Equation 16) for the dihedral angles of the rigid conformation at low temperatures. Lown66 reported 01 = 57°, 62 = 3° for cycloheptyl ketyl but no inversion even at +88°C. The 2-cycloalkanonyl radicals with 5, 6 and 7-membered rings are all very stable in adamantane even at room temperature. V. Heteroalicyclic Radicals (a) Five-membered Ring§_ The ESR spectra of the 2-tetrahydrofuranyl and 2-tetrahydro- thiophenyl radicals, XVIII(a) and XIX, both show temperature dependence 140 indicative of ring inversion processes. The barriers hindering inversion are 2.2 kcal/mole and 2.7 kcal/mole, respectively. Previous measurements of the barrier for 2-tetrahydrofuranyl gave 2.1 kcal/mole in a clathrate—hydrate matrix31 and 2.4 kcal/mole in adamantane20 in good agreement with the results of the present study. Tetrahydrofuran, like cyclopentane, undergoes pseudorotation with only a small barrier; both are able to invert by a low-energy path which does not involve going through the high central barrier of the planar configuration.1°9 However, with the introduction of a radical center apparently two equivalent conformations exist which interconvert only by passage over a substantial barrier (2.1 kcal/mole). It seems likely that these are the two half-chair conformations XVIII(a) suggested for cyc10pentanone110 which also has one sp2 carbon in the ring. The barrier in cyc10penty1 radical (2.7 kcal/molezo) is slightly higher than that in 2-tetrahydrofuranyl as might be expected from the larger repulsions when oxygen is replaced by a methylene group. Inversion in 2-tetrahydrothiophenyl radical XX would similarly involve the transformation by passage through an intermediate form of 055%? 5* )vs H 141 higher energy. In view of the expected greater delocalization of the odd electron onto sulfur, it is surprising that the barrier in 2-tetrahydrothiophenyl (2.7 kcal/mole) is larger than those fer 2-tetrahydrofuranyl or cyc10penty1; however, it must be noted that the errors are of the same order as the differences being considered. Irradiation of pyrrolidine in adamantane leads to formation of the radical XXI by loss of the NH proton rather than to the expected radical by loss of an a proton. With the odd electron centered on nitrogen, the l-pyrrolidinyl radical might have either the envelope or the twist-chair structure but the height of the barrier to inversion (2.0 kcal/mole) is only slightly higher than the values for the other five-membered ring radicals discussed above so the structures are probably similar. Inversion could then be attributed to the transfbrmation - 0 by analogy with cyclpentanone.110 The a proton hyperfine interactions in the 2-tetranhydrofuranyl and 2-tetrahydrothi0phenyl radicals, 13.6 G and 16.0 G, are much smaller than observed20 for cyclopentyl (23 G in a clathrate-hydrate matrix). This indicates that the odd electron population of the 0 carbon pa is smaller in the heterocyclic radicals. The latter may be estimated, 142 along with the 8 splittings, using Equations 15 and 16 with the empirical constants feund by Shiga et_al, for cyclic radicals: pa = a(Ha)/22.S and a(HB) = 50 pacosze. These lead to the values pa = 0.60 and 0.71 and = 50 °o(°/4) = 22.5 and 26.6 (if 6 is taken as 30°) for 2—tetrahydrofuranyl and 2-tetrahydrothiophenyl, respectively, to compare with the observed = 27.7 and 25.5, respectively. The higher apparent spin density on the o-carbon of 2-tetra- hydrothiophenyl, despite the general belief that delocalization onto sulfur should be greater than onto oxygen, has been explained by Biddles gt_gl,43 They suggest that the oxygen radicals deviate more from planarity at the trivalent carbon and so have smaller o-proton splittings despite a larger spin density on the radical center. It has been shownll’111 that increased bending at the radical center, and the accompanying increase in s character of the odd electron orbital, leads to more positive values of a(Ha). Since the values of a(Ha)observed for the cyclic radicals studied in this work are all presumably negative, the result would be smaller numerical values for oxygen radicals than for sulfur radicals with equal spin density on Co‘ This would explain the discrepancy between the calculated and observed values of noted above since, if pa for the oxygen radical is actually larger than calculated from the a(Ha) value, the calculated value of would also be larger. It is found for nitrogen containing radicals that a relation- ship of the type of Equation 14 holds so that a(14N) 3 QgpN with q: 219.0312’113 The observed value a(14N) = 14.5 G for l-pyrrolidinyl 143 radical thus indicates an odd electron spin density of about 0.75 on nitrogen. The B proton splittings appear to depend on the dihedral angle 6 through a relationship of the type of Equation 16 but with a larger value of the constant B. Thus, if we choose B0 = 0 and B = 57 the observed values a(HB) = 27.0 and 54.2 at low temperature give 2 corresponding values for cyclopentyl and for l-cyclopentanecarboxyl the values 01 = 13°, 0 = 47°. These are reasonable angles since the radicals are essentially the same. For the aliphatic radical (CHS-CH2)2N the values 3(14N) = 14.3 and a(HB) = 36.9 were observed113 with the latter value attributed to a conformational preference corresponding to <0>= 35°. (b) Six—membered Ring§_ The four heteroalicyclic compounds with six-membered rings studied were tetrahydrOpyran, 1,4-dioxane, pentamethylene sulfide and piperidine. When irradiated in the adamantane matrix all gave cyclic radicals in which an a proton had been abstracted. The stable confbrma- tion for all these radicals is assumed to be the chair form and ring inversion involving a chair : chair transformation is expected and is observed for the oxygen radicals. The radical from tetrahydro— thiopyran is unstable above -80°C so temperature-dependent ESR spectra in the region of coalescence could not be obtained. In the case of piperidine, more than one radical was produced and overlapping lines from the different spectra made it impossible to obtain the barrier fer ring inversion. 144 The barrier restricting ring inversion in 2-tetrahydropyranyl radical was fOund to be 3.7 kcal/mole and in 1,4-dioxanyl radical to also be 3.7 kcal/mole. These values are somewhat lower than the value Ea = 4.9 kcal/mole reported for cyclohexyl radical10 as might be expected from the replacement of a methylene group in the ring by one or two oxygens. The higher barrier AF+ = 10.5 kcal/mole found by NMR for ring inversion in cyclohexane compared to the values * i= AF = 10.3 kcal/mole and AF = 9.7 kcal/mole for tetrahydropyran and 1,4-dioxane are in qualitative agreement with this explanation and there is various evidence that a lone pair is smaller than a hydrogen.114 Recently Gaze and Gilbert.°° have measured the barrier in 1,4-dioxanyl radical prepared by photolytic decomposition of di-t-butyl peroxide in the presence of dioxane. They obtain a value Ea = 7.64 kcal/mole, which is about double that reported here and represents an unusually large effect of the matrix. The introduction of an sp2 center into ”5111* E 11.4 cyclohexane reduces the inversion barrier from Ea = AH kcal/mole to Ea = 4.9 kcal/mole in cyclohexyl whereas the reduction appears to be somewhat larger in the two heterocyclic radicals but any differences are not larger than the combined experimental errors of the two numbers (:1 kcal/mole). The degree of bending at the radical center was found to be small (zero i10°) in an INDO study of l,4-dioxanyl.°° It is probable that the lowered barriers on introd- ucing an sp2 center are a result of the lower torsional energy of sp2 -sp3 relative to sps-sp3 bonds and that the changes on introducing a hetero-atom into the ring are related to the changes in these torsional . 107 energies. 145 VI. Other Radicals (a) Cyclopentenyl and Cyclopentyl Radicals The cyclopentenyl radical (I), produced by irradiation of both cyclopentyl chloride and cyclopentyl bromide, is very stable in adamantane at room temperature. The hyperfine splitting parameters fur the radical (a(Hal) = 14.3 G, a(HaZ) = 2.7 G and = 22.2 G) are similar to those observed for the same radical produced by irradiation of cyclopentene in adamantane62 (a(Ha ) = l4.0,'a(Ha ) = 3.1 and = 21.6 G) and are typical of :llylic radical:115 When cyclopentyl bromide is Y irradiated at -196°C the product is the cyc10pentyl radical29 (XXIII) which is frozen into a single confbrmation at that temperature with two 8 protons and one a proton giving a(Ha) = a(HBI) = 22.8 G and a(HBZ) = 45.7 G. In the adamantane matrix the cyclopentyl radical does not appear to be a stable product. Irradiation of cyclopentane itself at -l96°C, however, was found to give only the cyclopentyl radical.25 In this investigation it was found that cyclopentane in adamantane gave the cyclopentyl radical when Y irradiated at 77°K and allowed to warm to -60°C. The spectrum at this temperature was a doublet of quintets indicating that rapid ring inversion is averaging the B splittings and = 35.6 G, a(H“) = 22.1 G. When the matrix warms to room temperature the spectrum changes to that of cyclopentenyl radical. The mechanism of this change is not clear but it seems unlikely that either hydrogen atoms from the initial radiation damage or the adamantyl radicals of the matrix 146 abstract the additional two protons from the ring since this would require two additional hydrogen abstraction reactions; it also seems unlikely that H2 is lost from cyc10pentyl. Helcke and FantechiS4 suggested that cycloheptenyl radical was formed in irradiated cycloheptane (in thiourea matrix) as a result of abstraction of a hydrogen from cycloheptene by the first-formed cycloheptyl radical; the cycloheptene would be one radiolysis product of cycloheptane at low temperature and would not be detected by ESR. By analogy the reaction + fiH.H H H might be the source of cyclopentenyl radical at higher temperatures in irradiated cyclopentane in adamantane. The cyclopentenyl radical is probably planar since the 0 system extending over three of the ring carbons would tend to favor the conformation in which the other two carbons are in the same plane. However, the observed average value of a(HB) could arise from a planar radical for which a(HB) = 0.6(50)(3/4) = 22.5 G, since pa 2 14.3/23 O _ and 0 = 30 , or from a rapidly inverting radical with 01 = 20°, 62 - 40° and = (26.4 + l7.6)/2 = 22.0 0. 147 (b) l-Hydroxy-l-cyclopentyl Radical Irradiation of cyclopentanol in adamantane leads to formation of the l-OH-cyclopentyl radical for which the fourfiprotons are equivalent and = 27.6 G. The same radical in toluene solution was produced by UV irradiation of l,1'-azobis (l-hydroxycyclo- pentane) and = 27.0 observed?0 and also by UV irradiation of di-t-butyl peroxide, cyclopentanol and iso-octane mixtures with = 28.0 reported.32 The radical appears to be rapidly inter- converting between half-chair forms at room temperature. (c) The Radical from 1,2-Dich1orotetrafluorocyclobutene While the structure of this radical cannot be deduced with certainty from the ESR spectrum, the loss of a fluorine to give radical XXVI appears to be most probable. Irradiation of l-fluoro-l-bromo- 2,3-dimethylcyclopropane gives the l-fluoro-Z,S-gisrdimethylcyclopropyl radical116 with a(Fa) = 77.6 G which is comparable to other a fluorine splittings in aliphatic radicals (Table 3). Also, the splittings for B protons on the same side of the ring as the odd electron are larger than those on the opposite side. Since typical B-fluorine splittings fbr such radicals are in the range 20-40 6118 and a-chlorine splittings in the range 4-10 6118 the hyperfine splittings estimated for structure XXVI are reasonable. On the other hand, formation of a radical anion with the odd electron in a higher molecular orbital would probably lead to larger values of a(F) than observed since the radical anion of perfluorocyclobutane shows eight equivalent fluorines with a(F) = 147.7 6. Addition of the odd electron at the double bond to give 148 radical XXXIII is also not likely since the ketyl of perfluorocyclo- (a M» F/ xxxxn CI xxxxv F butanone (XXXIV) shows large B and y hyperfine splittings118 a(FB) = 82 G, a(Fy) = 35 G, and two fluorines with a(F) = 82 G would give an ESR spectrum with much greater spread than observed. 149 SUMMARY Solid solutions of sixteen cyclic compounds in adamantane have been irradiated and the ESR spectra studied over the temperature .range 77°X-300°K. The cyclic radicals produced have been identified from the ESR spectra and the nuclear hyperfine interaction constants and g values measured. In the case of ten radicals the temperature dependence of the spectra has permitted the rate of ring inversion to be studied as a function of temperature and the constants Ba and (1/7 )0 of the Arrhenius equation determined. The significance of the hyperfine interactions, g values and energy barriers restricting ring inversion, has been studied. Barriers to ring inversion in six-membered rings are reduced by about 59% in going from the values measured by NMR in the undamaged molecule to the radical. Energy barriers in five- membered cyclic radicals are smaller than in the correSponding six-membered ring radicals but no NMR results are available for the undamaged molecules. The only seven-membered ring radical which could be studied, 2-cycloheptanony1, shows a barrier to inversion slightly smaller than the six-membered radical 2-cyclohexanony1 but much larger than 2-cyclopentanonyl, which is too small to be measured by ESR in the adamantane matrix. Various characteristics of the radiation chemistry of cyclic compounds in adamantane have been discussed along with the effects of the matrix on hyperfine splittings and inversion barriers. 150 A computer program ESREX4 has been written for calculating the rates of exchange of equivalent nuclei in radicals from the ESR parameters and the relative intensities of second-derivative Gaussian lines in the ESR spectrum. 1. 4. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 151 References - E. Zavoisky, Doctoral Dissertation, Moscow, FIAN, 1944; J. Phys. U.S.S.R. g, 245 (1945). P. W. Atkins and M. C. R. Symons, "The Structure of Inorganic Radicals", Elsevier, New York (1967). A. Abragam and B. Bleaney, "Electron Paramagnetic Resonance of Transition Ions", Oxfbrd University Press, New York (1970). J; K. Kochi, editor, "Free Radicals", Vols. 1 and 2, Wiley-Interscience, New York (1972); S. Y. Psezhetskii, §£_al,, "EPR of Free Radicals in Radiation Chemistry", Wiley, New York (1974). H. M. Swartz, J. R.Bolton and D. C. Borg, editors, "Biological Applications of Electron Spin Resonance", Wiley-Interscience (1972). R. F. Picone, Ph. D. Thesis, Michigan State University (1973). R. C. Schoening, Ph.D. Thesis, Michigan State University (1973). J. R. Shock, Ph.D. Thesis, Michigan State University (1973). W. G. Waller, Ph.D. Thesis, Michigan State University (1973). R. W. Fessenden and R. H. Schuler, J. Chem. Phys. 59, 2147 (1963); J. Chem. Phys. 45, 2704 (1965). W. T. Dixon and R. O. C. Norman, J. Chem. Soc. 4850 (1964); A. J. Dobbs, B. C. Gilbert and R. O. C. Norman, Chem. Commun. 25, 1353 (1969); J. Chem. Soc. A 124 (1971). J. K. Kochi and P. J. Krusic, Chemical Society Special Publication No. 24, London (1970); P. J. Krusic and J. K. Kochi, J. Amer. Cham. Soc. 29, 7155 (1968). J. E. Bennett, "Molecular Spectroscopy", Proceedings of the 4th Conf- ference on the Spectroscopy, Institute of Petroleum Chemistry, London (1968). J. Roncin and R. Debuyst, J. Chem. Phys. 51, 577 (1969). R. W. Fessenden and R. H. Schuler, J. Chem. Phys. 45, 1845 (1966). O. H. Griffith, J. Chem. Phys. 41, 1093 (1964); 42, 2644, 2651 (1965). D. E. Wood and R. V. Lloyd, J. Chem. Phys. 53, 3840 (1970); 55, 5932 (1970). C. E. Nordman and D. L. Schmitkons, Acta Cryst. 18, 764 (1965). J. Donohue and S. H. Goodman, Acta Cryst. 23, 352 (1967). A. P. Kuleshov and V. I. Trofimov, Zh. Strukt. Khim. 14, 866 (1973). 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 152 L. Bonazzola, N. Leray and R. Marx, Chem. Phys. Lett. 24, 88 (1974). H. F. Walter, Ph. D. Thesis, University of Pittsburgh (1975); D. M. Camaioni, H. F. Walter, J. E. Jordan and D. W. Pratt, J. Amer. Chem. Soc. 95, 7978 (1973). B. Smaller and M. S. Matheson, J. Chem. Phys. 28, 1169 (1958). Ia. Cherniak, N. N. Bubnov, L. S. Poliak, In. D. Tsvetkov and V. Voevodski, Opt. i Spektr. 6, 360 (1959); G. M. Zhidomirov and N. Bubnov, Opt. 1 Spektr. 12,245 (1962). Z8 nlo 2 2 NIQ t2 + (01 - Y + t + (n1 + y + 2 R2 («gin 2 iw)t (t) J im)t dt} dt + (A-30) (A-31) (A-32) (A-33) (A-34) (A-35) (A-36) (A-37) (A-38) 02 . m - 5— t + (02 + y - 2 - 1w)t I2(w) = Ream)1 f0 e dt + 02 2 1A . (A-39) m ‘ 2"t + (”2 ‘ Y ' 2_" 1w)t D2 foe dt} Now (.0 ‘(D n + ié-- 1w - 1(w + w ) - T-1 + 1( hz hl) ‘ iw 1 2 O h1 2 wh1 + whz -1 = 1(00O + 2 - w) - T (A‘4O) = i(wm - w) - 1'1 where w = %(w + w ) + w defined by Fessenden (A-41) h1 h2 O m - w n - lé-- 1w - 1(w + w ) - 1-1 - 1( hz hl) - iw 2 2 O 1'12 2 whl + th -l = 1(m0 + 2 -- w) - T (A-42) = 1(wm - w) - T-l Because Eq. (A-39) is the same as Eq. (A-4l) the integrals of I1(w) are the same as those in 12(w). .So we combine (A-38) and (A-39) into V V co + co _ I(w):=.Rea1{[C1 + Dljfoe dt + [C2 + DZJIOe dt}, (A-43) where 02 2 -1 V = - E—-t + {-1 :_y + 1(wm - w)}t (A-44) Here y=/—17-(%)2=(%) /(§52-1=(% aZ-l (A-45) T 164 where a = Z/TA as defined by Fessenden. (A-46) . -l A . Since T = (5)0 we can wr1te V+ as 02 2 A 2 ' v1 = - z—t +{(-2-)[-01 1/01 - 1] + 1mm - w)}t (A-47) 2 {w - w) = - ngl— + {E-a :_/a2 - 113 + i——E—E———J(ot) (A-48) where a = A/Zo as defined in Fessenden. (A-49) So 2 -— / 2 (wm - w) V+ = -12[ (0t) 4’ 2{[+a + O. - 1:13 - 1——o——— }(Ot)] (A-SO) 1 2 I 2 =-2[(ot) + 5,] + 15, W51) where (w - w ) —- 2 . m €:-- [a + /a - 133 + r—-7;———— (A-SZ) So a, V+ ‘16: a, -‘z[ (0t)+£:+12 J+ = foe —dt = e ‘ foe * dt . (A-53) 1 co on Let t' =ot+g+,then dt' =odt= dt=3 dt' and t |O=ot'|E — 2‘. 2 2 18a, ,,2 {—142 _ _°°-ét .1. __21 i- - J:-- e f€+e 0 dt' - o e Ll - ¢(Ei)] , (A 54) 1 x -‘u2 where 0(x) is the standard error function -——-f m e 1 du. (A-SS) V20 - [Note that the error function has a complex argument.] To evaluate C C D and D 1, 2, 1 2 we set the following initial conditions: which from Eq. or' 01‘ and So C— —1+1-E. 1 - 21v - 217' 2 4y But from Eqs. (A-45), (A-46), 2 TA 2 - 1 TY=(§—)G'1=aa 24__ 1A i 47 ' - 4(% 012-1 202-1 _ 1 a 1 1 1 _ __ C1-2( 2 )+2-2(-'----—------2 ) a — l a — 1 Thus, 61(0) 62(0) 1 =201 + 165 l =*C1 + C = 1 2 D2 ll p—a (A-25) gives 'A 'A m + 3—1c1 + rm + §—)(1 - c1) = 1 C1{T(Y + $4) - rm + g—An 1+T(Y-%‘A—) + 2 2 C II lr—I f——fi I Q Q I . 2 . 1A _ :[a -1 .2 1(Y + 2—JC1 - ( a + a)C1 in/az - 1 + 11%)(1 + a- 'A T(-Y + é—l = 1 a 1 (A-56) (A-57) (A-58) (A-59) (A-60) (A-61) (A-62) (A-63) (A-64) (A-65) 166 01‘ 0 =1—( 62-1+i+(a-i)+i———(°"i)) 1 20 l./az - 1 =%a-(/01:-1+01+———— i—°‘———) 01-66) ./a2 a - 1 N = A-67 20 /—————- F———_‘) ( ) So, 1 a2 D1 + C1 = Ea.(f + a + a + (12 .— -\//_____.) (A-68) or D + Hl- (A-69) Since (D1 + C1) + (D2 C2) =2 we have a D2 + C2 = l - -————-——— (A 70) / 2 a - 1 Using Eqs. (A-69), (A-70), (A-54) in (A-43) gives 1(0)) = Rea1{(1 + + (1 - ———°‘————)J_} (A-71) 2 . /a2 . /a - 1 =-ZZEhRe a1{2 B e1- [1 - ¢(g _)]} (A-72) O + — when B+ = 1 4;-——E-———- as defined by Fessenden ; (A-73) - 2 a - 1 and from (A-52) 167 €:-= Y:-+ 1 (w ; mm) 7 where Y+ = [a :- az - 1]a as defined by Fessenden and - w , h2 h1 w = go» +4» ) + w m h1 h2 O . Let 2 ~ _—— %E r 1(w) = “2" {2 8+6 1 L1 - ¢(£+)1} . + _ _ Then I(w) = Rea1{T(m)}, 31(m) 81(w) 32I(w) 821(w) ..____.. = Real{ } = Real{——-—-} - 3m 3m 8w2 2 3w We see from (A-74) that 35 '0' . " l. 3w 0 and from (A-SS) that 3¢(x) = 1 -%x2 3X m e 80 (A-74) (A-75) (A-76) (A-77) (A-78) (A-79) (A-80) 168 (A-81) (A-82) (A-83) (A-84) (A-85) (A-86) (A-88) (A-89) 2 ~ Lg , 3I(w) _ J2; 2 i- .i _ T — —O_ {E B:{3 (5:) (0.)[1 ¢(§:)J _' 2 2 kg -%6 ~ + e i [- —l—-e t (%J]}} /2—1r L52 =5? 1 1E1-¢(£)]-1} + 0 I. 1- 2n ii 1 2 1 6E = __“. 2" {2 B e + ' ¢§_- )} 0 O i i i 1 2n 0 or N 16+ /_ 31(wl._ " T(w) .-—gl 3 B ( 1 ). 3(1) 0 C + __ 2.". 0' Thus, 2~ . 8 1(w) 1 2 = ( )(;)I() 3w 1; “L. .5; 1 +__t{ 1(w) - z ( )} a O O :_ 2n 0 £2 -£ + + — (- l§-- —%§I(w) - 12E:Z+B _( ) o o + 2 -— 2n 0 /__ €+ e52 = l: .21 {2 3+[ —'- (1 + £3)e i [1 - ¢(E+)]]}.(A'87) o o :h._ /5; ._ ._ Finally’ 1452 3 21(;)- 3 Real{2 B :[- -g: + (1 + g :)e —{/E;t1 - ¢(€ +)]}]}. 8w 0 __ \______.~ ___=____J Fessenden's erf(€+) We have 2 -1/ ¢(z) = if?” e at dt J2? 169 From the "Handbook of Mathematical Functions")edited by M. Abramowitz and I.A. Stejun (1967)/referred to as [A3,there is defined (p. 297) erf(z) = —2-f8 e-tzdt ; (A-90) /; . let t=1 then dt=£1:—' and tug->3; S+tlé§z so J? /2 /2 , .2 erf(z) = 2-—l—.fg§‘ e-gt dt' = 2{¢(/§ z) - ¢(0)} /if or erf(z) = 2¢(/§ z) - 1 . (A-91) Also [A] defines erfc(z) = 1 - erf(z) (A-92) and _22 w(z) = e erfc(-iz). (A-93) We then have erfc(z) = 2 - 2®(/§ z) (A-94) and -22 w(z) = 2e [1 - ¢(—/§ 12)]. (A-95) So 22 ¢(—/§ 12) = 1 - ge w(z) (A-96) or ¢(2) - 1 1e‘l’zzzmciz) (A 97) - -6 — ‘ - /5 Thus, %22 12 e (1 - ¢(Z)) = l/2w(—) . (A-98) 170 Using (A—98) in (A-88) gives ii 2 ___ + é—l121-= :l-Rea1{2 B [-E + (1 + €2)/n/2 w( _311, (A-99) 3 + + + 2 o + —- —- —- /2 3w —- According to formula 7.1.8 on p. 297 of [A] °° (' )“ w(z) = z —l§———-—- (A-lOO) n=0 F('§+ 1) Since F(%) = /;, we can expand this as w 2 k . 2 k w(2) = z {1:E71_.. (lz)('z ) 3 1 }. (A-lOl) k=0 1 (K + aux - a) - (‘2')(31/71 On the bottom of p. 328, we have the approximations : For 2 = x + iy, if x > 3.9 or y > 3 then . 0.4613135 0.09999216 0.002883894 w(z) = 12( 2 + 2 + 2 ) + 5(2) Z - 0.1901635 Z - 1.7844927 Z - 5.5253437 (A-IOZ) -6 where Ic(z)| < 2 x 10 . If x > 6 or y > 6 then w(Z) = 12(0é5124242 + 0&05776536 ) + "(2) (A-103) z - 0.2752551 z - 2.724745 where |n(z)l < 1 x 10-6. Also we have the identity (from bottom of p. 325 of [AJ) w(-x + iy) = w(x + iy)w(x - iy) 2 2 = zey -X (costy + isin2xy) - w(x + iy) (A-104) 171 Eq. (A-104) also gives 2 2 m(-x - iy) = w(x - iy) = Zey -x (costy - isin2xy) - w(x + iy), (A-lOS) If Im(€+) = O and Re(£+) > 6, then i 15 5i— . 1- .5124242 .05176536 w( ) = 1 2 + 2 /2 /2 15+ 15+ 4] - .02752551 [2421 - 2.724745 /2 /2 (A-106) +g+ _ -— .5124242 + .05176536 5 £2 2 .1 .1 (7) + .02752551 G.) + 2.724745 :3, 9 ) TDU INTENSITY. II x o ’0 'v15v ANGE ’ / I FOP JUMPINGTSEC/RAU) 0F TAU T N0 EXCH 08 f H olEXo‘NOEXCH(THE INTENSITY AT TAU I OUTPUT) .15X.F10.3,’(AT HI)’. AU.SIGHA) NTvkELINT oRELINT 15X.515.7.15X,F10.3.‘(AT H2)‘./// 1 172 5.0) 60 T0 200 APPENDIX B 55.140ee vHHIeHHZoTAUQSIGHL’ Xy‘CORPELATION T .27Xo£15.7¢15X,FIC.3.’(AT RP)‘. PROGRAM ESREX4 2.TAU.SIGHL) C 2vTAU.SlGHAD OUTPUT l‘ALOG 1E7! 121ET F. 9HH10HH29TLUOSIGHA’ 36HN1OHHN1X 509/3Xi9/“0 DrEUTPZUICTAUI DENNIIHHNI 9076009 SH1733E E9921£9 [TC/ECZLUI “Ni/S 0061" T3370 .1 OI ’/ 13‘ 01 NNuCZOEHH:6(Ol-PEI (bits-H. up.“ :ICnll|H : E‘HH : 6‘LGIU Aloo7o:H(H(T3.2XTT . (TRHZXTTTTITTTTHTTTTLO R #8oleFCEAE::EEAQEA.§:EFNEAOFNEAOFNEA:LOI GL::S:H3:XTHOGU:THoTHoCG::ITH0:ITHC=ITNGDTT OA12:UGHTEIDoOOUIROIRXHOUTLIEHTLIPHTLIPO! NDAPOI LP: F PEHFDLI : NCF01LLLF01FC5 : LANEFC : NED C : NEFCLFOONEHSP : EL HI PPHHHTSHINHF:AETHFEHF1HETIRHFHIRHFHIRHFAIGCERIZZ DAAHD C ZPIZ ’ A 1’0 LAN L192Ah EHO ADO. Z7. DCIhtwooOOo O IZHXOOA‘) USHGLQ‘HUU PHA'OOO) TSHH EC L 1C UU UU Z? O. IT D‘Ao :AAD ::==: LLRZSTH : ZZZ: ZZZZZZZZZZ ‘ OI!) PM 1130 DECO O on!“ 0022 V 0’ . II’IIOrLFUv XXpHooII 226611XXH FETI’t‘OHIPAXpZZU’1FF.OPQ‘TZZT LX3OHTO§SH‘LL. .ZOOIIZZAAl‘t. .0 NPLSIH‘Ol/CXPAAA/ooDDTTZlc.(TT 0H952./ZH’ LHZHHAOCZZRR((220.219 IOHZ/ZO‘HH2Pc1PPHOOOOOQHHIIPHOHH TCCOIHo/(H H TLLDOODHSSZZDPaBPZ 1PH2AO.‘C=R‘AL00YYZZ OZZA1122 T::Lo(P=FS‘TZ‘( ZZZZT II. ‘0 OL/OoOoH :0, ‘ SIGPA.HN.CELTAoOIF 18.109/. p '- — H0.HH1.HH2,TAU=',5 20.10) ZARMKZWR. .“E .ghElaotcOl’. O O PleP V.) 1. 0. 0:2 ! .u“..../ 1.85 I .52 00 In ’0 .1 Ir 0 0 6:0 22‘ 2 UAHE U2 '1‘ Z P O '0 T.. U'x II}... 9 32:...‘1A‘1‘1‘ O 0000 ZET 0'. DIP PHHHPHPH PABABAB PHGGZZUQTT: PHUPHIIRQ UYYUBBXXA‘ HIHIH? =REQEC AATTH CkChOo ZZZZZFF F F 100 11‘ 102 1C1 173 111vatL 1X,Y1 I T P 1 0 F C L A 3 5: w ‘1 0 ON CT 0F 1.1NTEG 7 3 D A F O X U1L C8 FPADI 9» XchTD HX NEGAAAN .0 E.HA JH E:TNE 1" H NOH7 2H: 111119 1.1 ZOYIAHZ HA1 5‘11" 21. EH1h o X T IIFD NYC. A‘pco 050 LS1£ o ILN UHTDK’ TDI/PVOR 0“ CH CLLOEOb NC/IALSHBQ; UCHXCO/131 F IE F0 N1 1’06 0 ‘ «AI NNsv1oHN k r.CNI‘.A(”U F L10 T 2 G U PLHAUN:ONC TDHPHTOO CIT NCPCruRI‘IT S PECICD TZ.0 SC11N0 INFPU‘ HUFXOO 1FC.EFI CCCCCCC Z A ' D N 0 U o F G O c. S N o v1 o o \- S P I 0 '1‘. O O X CH2 A .I N'l/ 0 U 2 1 9 FCuo E. 0 ‘0 NIHU N o 01. fi S IV." 0 T P Z p U’x 'I 315 I Y. 111 H 9 RPN 0 o .1. o H 5 $01 A E 1 IoTNF N 11 DZCIB 1|! 1’5 (N A x II) ,0 O LTUN 1 CO C 00 AG???» 0 o O 00 9 1 "0 V" v 90 097 0 9551 _L 0 ! 9,3 0 O/HGY 1 V 3 15 0 52b 0 2 N01 8 o I In 5“ 6 393 O O O s 1 P) T Q 57 I 6“.) o O leI‘ 11 O I 0 2h 3 1QZ 1 o 1 R10 5 o P S 3 52 5 085 C 1 D & NN 0 0 77 97. Z c 1 D 05200 19 Y P 1 20 1 105 J 0 O A NI II 00 9 1 02 1 0111 1 9 2 0| 5 1 1 A OTT O o 1 o i C :1 C 01.3 1 5 K G 6 H 2 T1TCC 02 Z r. F o 1. o 1. OA 5 o A 9 Z 1 52 NN '1 1 X YO & 6 .01 3 .016 5 0 = H 0 1 L 1YUU 06 G . .P1 5 o 012 o 01/2 9 1 0 I 10 G A ELTFF ofi L YZXT (IA T 1/1A 0 .2 A H1 0 OJ N1HAI 1 1 H 0=C:X1 56 I10 6 I100 7 : 1K 1 E N1 1 EN H DZTHNFL 0 1: .1 1 2 1 .000 o 101 o 10 .1 6 . Z R1R1 x1110 Z 0. RIRA o A X Y GNTNY Ooh Y CoCA 112 S o X6Uv .YYoC . NASOFdC 0 :H C C (EA I1 .01 1 .0 v1. 0.1K : o 1 9215 X. 0 5. 1 C1ENN 1. z 9 : TI.9IQ. N C o 0 9. o 3 91». 7T. 1 12x . ED .xXiL’ I TCIET O 6 HS F XG1R 962 Q 9692 393 X0 ONAER1 YuolI X NXA HA 2“ Y ClOZIY 00 23. 0 5180 81 No 10550 000 X . OEL 5TH 07 ,1 6660 IH13 0 “51 o 3231 1211 "0 OArhx 0111 V. o 01 O 9 ILU 0U .C 12 a s G 01.1001 260 1 1980 01R0 X9 H1X1101H 1222R R TPCSNRH 06 1o 1A1A$GO¢00 Q7. 9 398. .00. 012 PL1loooz P110X X CHLTIOT 02 00 1 OLOL DEBOQ o 211 I 1921 0150 s QKKZDEAT vAT-OOI‘ XSNO‘ 1 NOANHFA 010 O D Z oF.F2N Ev6 159 3 6909 09.92 1/lTDTEFALEEG 5015! X UCCE H ob 00 1001019 T00. 5000. “0000 0050T111ZOAIR0H000J OCSoL L F "NH 050 o-N L O o 1XNX61 o o 00.1 o 0 o 007. o o o 06 01.110 S‘ZGGN 0. 0 GD- P TEUOOF’ as 01/ A:E:E:911516200015200001o10001:10¢TH:= HAAO oXXAH H XINGRIOBC. o1U E G G 0 00 .0111 001111 211111 KOC Z :=ELLCN2XXLCNCN ECICFS 612 000 PGoceG1vPN0XZ OXZ 0. / H 3 H2 RFF R F R R LITA NK9.1U1NO AXLYA 17.1 = : :11 .. = : =1 :=Z=:L..=:: ..RHHXXII=U===1:U=U PLU LL01 / :L1L1L:1S1 1: 1: s o E D EEEA111 T 1 T TDHPOXAP01..Z= I: x FFFFF OIFOF 12HOF 123H0 1211HT1120H1FPHFFrusanIFHEHENOpr-sfixe .- XIIIIIGGHIGIOAAZGIQAAAZGOTT KZIKYTAZIXXXIIIZRxxx-IAZFZRECI Lcem7oUSNNU AFT : SPE 19 0 LN 0 50 00 0 5 fl 5 0 5 0 INTEAZ 0 1 F T? 23 “5 6 2 6 0 o 1 1 Z HONHHO 1 IVA T 1 1 1 1 1 TCITrP CCC c CCCCCC 15.10 oLT.1.CE-261 L171"2‘AIHAG1T1"21 5.100500001 HTAL