augmwfiwéfiw . . r L 1.x: :1 .- ..t. . ‘ c. ., , . .55.? wmé h, . 2»? ., m. E... rfifinfii Lax. _ ,i . $33.5: 14. , if , a. “Jan. I. , u 5 3.3.: i433. .’ i... r... .n . a}; . . .- 2 . will a r . 3 u..— ., ,if. 55": . .3 .r» . . ~ . , . . .. . . . . KimFICV 5.5.; , . ._ u ,. . , ¥n~v.?. “Elimi— , . . , . . . , ., . . Va .. , . symmuhfimw. r. . , . , ., ._, , , . . 1 . 3., an} , _. , .. . “3.; w. .. ,. , , Maw. _ .u .dfiafirxéwa , utnn ., _ . .. ., 5.... :4. 433-“. , . . , ., ”Mfg.— h . ..x..1fl..«.hA , . . , , ‘ .msfir.aufi _ . . r2, , t . . fan. . , , . a .3392... vhf,» unfit»? . . , 3.435;. e. . x 3mm , J . .1, 3,, . Em” t. , “ENG... .. . xx.) , 1hr 3. . TIL, . , . , “may“! , . $5.13.. , Jill... , i , sunny}; a: 13!». 5.14. 2.: .. .. ‘ ‘ p I. amefingWmuu ,, . . ..i:..w.t...fi).~w?fl:2~h 1%.”..«5uai «3.) L , . . a . 1., :1 1.!rwwxi3 flueflh .flWA .r... .7); fl . . ... luau. .gglhqflz. .W! Hr... lr\ fun 1 . .fin, . . , a {Liam}??? i 51.1. , , A . 1.1.x.- . . . . 13.2.. . . gwwfin 1. 1, rum. 1 .9... , . . r 125.. ‘51 ‘x . a git??? rgn 2 .. H 3.2.: , its“ 1. :0. .131.“ 156395.12”? 5.1. :, Z5521. x. L , 1 . . .. , "Fan... ,. A... (#1., t, 57‘... , , ‘ . , . , A 1.- a; . __.J£_fl.u= . , £41, AH?!) . .a ‘9. . 3.... , . . a We? 8.... I; ., inmblfiasrflé I ... 1,.) . W‘s-.7304). S 4 . 1.3....) ‘ .. N3.5!3.H.V.J3l..rc.f.r?f$nl . 283.11....,.l.si, . ”9.2... $1er? I if. 32.9.. n I a ,V ..\3 .v...,:...... . 1 iv 3%.?(5...2(£5T. ._ 3.. . é} affix“. I. 3A).:e... Er, 5.. 3:111.) .. 12...)..1/ rnlsa: .11... .1 ...h ,1. r.¢I ..J|!tI/ufr:vi.l(l v 3., (4.. ...((,r1»:.., 2. .u... ”2.3% £1814! Y Daisy "1» n tn .rihh that ”if "was rrhun! PAH? 2: TN! HEEWUUC STMYURI 0V CmVLIITPENH an? H‘ 7"! IHUKS m .1004”?! 0‘ 7M 309397an 0‘ I m-UCL C \ZClELiit} CW1“ O 0 “Wu” "J ' \ arm-r '. 'meab ‘u‘u‘r'; r;!r'.2 u;'?\‘..‘ "r'! ' f' .- 0.2." ff".-"'!| 9 .- o- ‘s i ‘ - a) . a u .srcm - via—“Au -/ —) - ,' M {I’m Iqrv :r 4le hm “out ?, 177‘ c: M222 .1. 1 ‘9‘ 8.." '“V K,/\, ,’ ABSTRACT PART I: THE ELECTRONIC STRUCTURE OF CARBONYLNITRENES PART II: THE KINETICS AND MECHANISM OF THE FORMATION OF A NON-CYCLIC NICKEL(II) COMPLEX BY George M. Shalhoub Minimal basis ab initio SCF-CI calculations on a series of carbonylnitrenes, XC(O)N, where X = F, H, CH3, CH3O, predict the energy level pattern: E(3A") < E(1A') m E(1A"). In the parent compound, HC(O)N, the energy separations are 1 3 1.86 eV for A" + A" and 2.08 eV for 1A' + 3A". A geometry search for the low-lying states of HC(O)N predicts CON and HCN angles of 114° and 123°, respectively. The CO and CN bond lengths are computed to be 1.225 A and 1.504 A in the 3A" ground state. Substitution has little effect on the computed energy level pattern in the low-lying states of these simple carbonylnitrenes. An ab initio SCF-CI calculation in a [38,2p/28] basis 1 3 predicts a A" + 1A' + 3A" separation of 1.54 eV; these results demonstrate A" energy separation of 1.75 eV, and a George M. Shalhoub the efficacy of the minimal basis calculation. Ab initio calculations predict a 5A ground state for 1 a carbonyldinitrene, NC(O)N. The rate of reaction of the metal complex, [a,a"'- isopropylidenebis(azo)]di-a—stilbenolatonickel(II), NiMMK, with 1,3-propanediamine to give the product, [a'[[l—[2— [(3-aminopropyl)amino]-1,2—diphenylviny1]azo]—1—methy1- ethyl]azo]-a-sti1benolato]nickel(II), NiApSo, was followed spectrophotometrically in a solution of tetrahydrofuran and ethanol. The dependence of the rate of reaction on the concentration of each reactant was investigated, and the reaction was found to be hydroxide catalyzed. The experimental rate law is —d[NiMMK] = k NiMMK 1 3- n 2 OH— dt 1 + k'[OH'] * Activation parameters are AH338 = 7.81.3 kcal/mole and * AS338 = —48i2 eu. A mechanism which accounts for the observed rate law is postulated, and a short-lived adduct which involves the coordination of two amines in the axial positions of NiMMK is proposed. PART I: THE ELECTRONIC STRUCTURE OF CARBONYLNITRENES PART II: THE KINETICS AND MECHANISM OF THE FORMATION OF A NON-CYCLIC NICKEL(II) COMPLEX BY 1'. George M: Shalhoub A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1976 TO SUE ANNE ii ACKNOWLEDGMENTS I would like to express my appreciation to Dr. James F. Harrison and Dr. Gordon A. Melson for their guidance, understanding, and friendship during the course of my graduate studies. I also thank my friends and colleagues for making my stay at Michigan State University an enjoyable one. Special thanks are due to Larry, Dennis, Jim, Bob, Gary, Rich, and many others for the sometimes enjoyable after- noons we spent digging up the East Course. I would also like to thank Peri-Anne Warstler for her help in the preparation and typing of the final copy of this thesis. Finally, I think my wife, Sue Anne, who not only typed the rough draft of this thesis, but who also kept my spirits high during a rather tough term. iii Chapter TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . LIST OF FIGURES. . . . . . . . . . . . . PART I - The Electronic Structure of Carbonylnitrenes . . . . . . INTRODUCTION . . . . . . . . . . . . DETAILS OF CALCULATION . . . . . . . A. B. C. D. E. F. G. Formylnitrene. . . . . . . . Fluorocarbonylnitrene. . . . Methylcarbonylnitrene. . . Methoxycarbonylnitrene . . . Discussion of Minimal Basis Results 0 O O I O O O O O O O Formylnitrene in a More Flexible Basis . . . . . . . Carbonyldinitrene. . . . . . PART II - The Kinetics and Mechanism of the Formation of a Non-Cyclic Nickel(II) Complex . . . . . INTRODUCTION . . . . . . . . . . . . SCOPE OF THE THESIS WORK . . . . . . EXPERIMENTAL . . . . . . . . . . . . A. Preparation of Material. . . NiMMK. . . . . . . . . . . . Solvents . . . . . . . . . . Amines . . . . . . . . . . Sodium Hydroxide Solution. Dielectric Constant. . . . Procedure for Obtaining the Kinetics Data. . . . . Treatment of Data. . . . . iv Page Vi ix 10 11 26 37 40 42 47 55 Chapter Page RESULTS. . . . . . . . . . . . . . . . . . . 85 A. Hydroxide Dependence of the Rate . . . . . . . . . . . . . . . . 85 B. Influence of the Solvent on the Reaction Rate. . . . . . . . . . 92 C. 1,3-pn Dependence of the Rate. . . . 97 D. NiMMK Dependence of the Rate . . . . 104 E. Effect of Temperature on the Reaction Rate. . . . . . . . . . . . 104 DISCUSSION . . . . . . . . . . . . . . . . . 112 FUTURE WORK. . . . . . . . . . . . . . . . . 119 APPENDIX . . . . . . . . . . . . . . . . . . . . 121 A. STD-3G Basis Set . . . . . . . . . . . . 121 B. 3s,2p/Zs Basis Set . . . . . . . . . . . 122 C. The CI . . . . . . . . . . . . . . . . . 124 D. Syntheses. . . . . . . . . . . . . . . . 131 REFERENCES . . . . . . . . . . . . . . . . . . . 137 Table LIST OF TABLES Molecular Orbitals of Formyl— nitrene. . . . . . . . . . . . . Computed Energies of the Various States of Carbonylnitrenes . Natural Orbital Occupation Numbers in the Three Lowest States of XCON . . . . . . . . . Population Analysis of Formyl— nitrene. . . . . . . . . . . . AB: for the Reaction F + HCNO + FCNO + H. . . . . . . . . Population Analysis of Fluoro— carbonylnitrene. . . . . . . . . Population Analysis of Methyl- carbonylnitrene. . . . . . . . . Population Analysis of Methyoxy- carbonylnitrene. . . . . . . . . Energies of the Low Lying States of HCNO Computed with [3s,2p/Zs] Basis. Vertical Transition Energies of HCON in [3s,2p/Zs] Basis. . . . . . . . . . . . . . Population Analysis of HCNO in [3s,2p/2s]Basis. . . . . . . . . Vi Page 12 20 21 25 35 36 39 41 Table Page I-ll Molecular Orbitals of CNNO . . . . . . . 56 I-12 Carbonyldinitrene: Energies of States and Vertical Transition Energies . . . . . . . . . . . . . . . . 58 I-13 Population Analysis of Carbonyl- dinitrene. . . . . . . . . . . . . . . . 63 II-l Experimental Conditions for the Kinetics Experiments . . . . . . . . 79 II-2 Observed Rate Constants as a Function of Hydroxide Ion Concentration. . . . . . . . . . . . . . 93 II-3 Solvent Dependence of the Rate . . . . . . . . . . . . . . . . . . 98 II—4 Observed Rate Constants as a Function of the Square of the 1,3-propanediamine Concentra- tion . . . . . . . . . . . . . . . . . . 99 II-S Temperature Dependence of the Rate . . . . . . . . . . . . . . . . 108 Al Carbonylnitrene Basis Set: STO-3G Coefficients and Exponents. . . . . . . . . . . . . . . . 123 A2 [33,2p/2s]Basis Set for Formylnitrene. . . . . . . . . . . . . . 125 A3 Determinental Basis for Formyl— nitrene. . . . . . . . . . . . . . . . . 128 vii Table A4 A5 A6 Page Abbreviations. . . . . . . . . . . . . . 132 Results of Synthetic Work. . . . . . . . 133 Solvent Effects on Macrocycle Formation. . . . . . . . . . . . . . . . 135 viii Figure I—l LIST OF FIGURES Reaction Scheme for the Addition of Ethoxycarbonylnitrene to 4— methylpent-Z—ene . . . . . . Kinetics Scheme for the of Ethoxycarbonylnitrene . . . . Calculated Geometries of the Carbonylnitrenes . . . . . . . . Energy of the Lowest States of HC(O)N as a Function of the Out— of-Plane Bending Angle . . . . Contour Maps of the Electron Density Associated with the Natural Orbitals of HCON . . . Reference Configurations for Carbonylnitrenes . . . . . . . . Contour Maps of the Electron Density Associated with the Natural Orbitals of FCON . . . . Relative Energy Levels of Carbonylnitrenes . . . . . . . . Electron Distribution of the State of XCON. . . . . . . . . ix Page 16 19 24 29 31 34 45 II-l II-Z II-S II-6 II-7 II-8 II-9 Comparison of the Results of STO-3G Basis with the [3s,2p/Zs] Basis . . . . . . . . . . Orbital Contours for sz Orbital of CNNO. . . . . . . . . . . Orbital Contours for 8al Orbital of CNNO. . . . . . . . . . . Reaction of NiMMK with 1,3- propanediamine to Form NiApSo. . . . Absorption Spectra of NiMMK and NiApSo at Equal Concentra- tions. . . . . . . . . . . . . . . . Experimental Apparatus . . . . . . . Spectra Recorded During the Reaction of NiMMK with 1,3-pn. . . . Computer Fit of the First Order Decay at 390 nm. . . . . . . . . . Computer Fit of the First Order Decay at 340 nm. . . . . . . . . . . kObs as a Function of Hydroxide Ion Concentration. . . . . kobs as a Function of Inverse of the Dielectric Constant . . . . . k as a Function of Square of obs the 1,3-propanediamine Concentration Page 50 6O 62 68 75 82 87 89 91 95 100 103 Figure II-lO II-ll II-12 Page Ao - Aw . £n'——-—-— as a Function of A - Am Time for NiMMK O O O O O O O O O O O O O 106 in k as a Function of the Inverse of the Temperature . . . . . . . . . . . llO Schematic Representation of the Adduct O O O O O O O C O O O O O O O O O 11 7 xi PART I THE ELECTRONIC STRUCTURE OF CARBONYLNITRENES INTRODUCTION Nitrenes are electron deficient electronically neutral molecules in which the nitrogen atom possesses one single bond. These species are categorized according to the or- ganic substituent bonded to the nitrogen. Important classes of nitrenes are the arylnitrenes, alkylnitrenes, alkoxynitrenes, and carbonylnitrenes. All of the above nitrenes can be generated in three general ways. The first method is the photochemical decomposition of organic azides, which produces the nitrene and also nitrogen gas. Two other methods of forming nitrenes are thermolysis of azides and a-elimination of certain carbamates. In recent years a large amount of experimental data concerning the physical properties of some nitrenes has been obtained (1) and much of these data have been de- rived from electron spin resonance (esr) studies. Wasser- man (2) and coworkers have been able to observe the esr absorptions of various nitrenes at 4°K. The interpreta- tion of these data for a series of alkylnitrenes implies that these alkylnitrenes have a low-lying triplet state which may be the ground state. Also the values of the zero field splitting parameter, D (ml.6 cm-l), seem to indicate that the unpaired spins are localized in one :region of space, while the values of E (m.002 cm-l) in- delocalize into the n system of the arene since the nitrogen pN orbital can mix with the p1T orbitals of the aromatic system. The esr spectra of several arylnitrenes have been observed (3,4) and suggest that most of the arylnitrenes have a low lying triplet state. The estimated values for the zero field splitting parameter, D (~1.3 cm-l), are less than those D values for alkylnitrenes, and this fact may imply some mixing of the nitrogen pn orbital with the p1T orbitals of the phenyl group. In one case (3), the spectrum of the arylnitrene, m-phenyldinitrene, was interpreted in terms of a quintet ground state for the nitrene. The value of D for the dinitrene was estimated to be .156 cm"1 and the largest contribution to D was calculated to be from the interaction of the electrons on the same atom. Carbonylnitrenes have been proposed as intermediates in a number of reactions which involve the rearrangement of'azides or haloamides to form isocyanates. The debate on whether or not carbonylnitrenes participate in these rearrangements has continued since Tiemann first postulated the existence of carbonylnitrenes in 1891. In the Curtius, Ihoffmann, and Lossen rearrangements, a group migrates to an electron deficient nitrogen to form an isocyanate. R-CO-N-X + O=C=N-R Two mechanisms seem to account for the observed facts; the first involves a discrete carbonylnitrene intermediate, while in the second no discrete intermediate is formed. Recent experimental data (1) seem to indicate that carbonyl- nitrenes may not be involved in these reactions, but un- fortunately there is little positive data which supports the concerted mechanism. This is one instance where quan- tum chemical calculations may help clarify the situation by determining if the carbonylnitrene has suitable or- bitals which can bond to the migrating group. We begin our study of carbonylnitrenes by reviewing some of the data which have been obtained regarding these molecules. Most of the data on carbonylnitrenes concerns one species - ethoxycarbonylnitrene. If the photolysis of ethyl azidoformate is carried out in the presence of cyclohexene, one can isolate 7—carbethoxy-7-azabicyclo— [4.1.0] heptane, which is expected to result from the addition of ethoxycarbonylnitrene to the cyclohexene (5-8). A lower limit to the lifetime of the nitrene was esti- mated (6) as 3 x 10.7 seconds. This number is estimated by calculating the collision frequency of the nitrene with cyclohexene. Since the nitrene decomposes to C H 0° 2 S and NCO', an upper limit to the lifetime of the nitrene is estimated from the time of appearance of the intense NCO- bands in the UV spectrum, and the upper limit is 10'5 seconds. The experiments done by Lwowski suggest that C2H50C(O)N was formed in the photolysis of ethyl azidoformate. Mc- Conaghy and Lwowski (5,9,10) tried to determine the spin state of C2H50C(O)N using a model introduced by Skell (11) which correlates the mode of addition of carbenes to olefins with the spin state of the carbene. Skell's theory, applied to nitrene reactions, states that nitrenes in the singlet state add to olefins in a single fast step resulting in a stereospecific product. However, nitrenes in triplet states add in two steps and the extra step allows for rotation about the C-C bond of the intermediate and a nonstereospecific product is formed. Ethoxycarbonylnitrene was generated by photolysis of C2H50C(O)N3 and by a-elimination (5) of CGHSSOZONCOOCZHS, and it was trapped by the olefins, cis- and trans-4-methyl- pent—Z-ene, CH3CH=CHCH(CH3)2. This reaction is shown in Figure I-1 and the kinetics scheme used to interpret the data is presented in Figure I-2. Lwowski concludes that ethoxycarbonylnitrene is produced in the singlet state when the methods used to generate the nitrene are thermolysis and a-elimination. When this nitrene is generated by photolysis, approximately 30% of the nitrene is initially formed in the triplet state. The inter- preted data give a rough estimate of the relative energies FIGURE I-l Reaction Scheme for the Addition of Ethoxycarbonylnitrene with 4-Methylpent- 2-ene. HIH musmam I .omI [On MUIO, ...I foomImo + O/ o\ Aw _ \z wonImo NIIUII \ ..I NAMIoVIu\o...Io mm .. H .. m . ell $2:on oo I + .zwomIJ I I o FIGURE I-2 Kinetics Scheme for the Reaction of C2H50C(O)N. mnH magmas .uumum uoawcam may cw pmsuou xaamwuucw ocouuw: mo assess «nu aw m was .ouMum uoaaauu onu ca voEuou haamwuwaa «swung: mo ucsoam onu ma H «pupa qx\nx + Aawwoaovmx «x n m\H mcwvfiuuud camaoonmoouowmcoz . osakuan< camaumamoouuum J muusvoum «couuuz acouuaz onusooum u «3.... um :23... .4 2:33. .v. 8852 of the singlet and triplet states of ethoxycarbonylnitrene, but unfortunately the data cannot indicate the orbital distribution of the electrons in the molecule. Further evidence that ethoxycarbonylnitrene has a low-lying trip- let state is available from esr experiments (12). Since the carbonylnitrenes have such a transient existence, they are not easily studied by the usual spectroscopic techniques. It is in this instance that theoretical calculations can be of significant importance in elucidating the nature of these species. The work presented in this thesis is concerned with the structure, relative energy of states, and orbital populations of carbonylnitrenes and it is hoped that the results can supply some insight into the chemistry of nitrenes. lO DETAILS OF CALCULATION The calculations presented in this thesis were under- taken to determine the geometries, multiplicities, charge distributions, and relative energies of a class of carbonyl— nitrenes, XCON. These properties were determined for the low-lying electronic states of each molecule, and they were studied as a function of substituent, X, where X was chosen to be the fluorine, hydrogen, methyl and methoxy moieties. These groups span a range of characteristics from the highly electron withdrawing species, fluorine, to the electron releasing methyl group, and should give a complete picture of the structure of simple carbonyl- nitrenes. The initial geometry was chosen to be a trigonal planar structure possessing the symmetry characteristics of the CS point group. In this point group, states and orbitals are labelled (A') or (A"), depending on whether they are even or odd upon reflection in the molecular plane. The independent angles, a and 0, were set at 120° while the i) '1' /S\ H N CC’ and CN bond lengths were assumed to be 2.29 and 2.87 atcunic units (au) respectively. Note that throughout the 11 thesis, atomic units are used unless specified. Useful conversions are: .52917 Angstroms equals 1 an of length (bohr), and 627.5 kcal/mole equals 1 hartree (H), the atomic unit of energy. The CH bond length was fixed at 2.060 au. The other parameters were varied in the search for the minimum in the potential energy surface. A. Formylnitrene The Schroedinger equation has not yet been solved exactly for a system possessing more than one electron, so a number of methods for approximating a solution have been devised (13). The method of choice for molecular systems is the Hartree—Fock-Roothaan (14) (HFR) technique which is used in these calculations on nitrenes. This method, combined with configuration interaction (CI) (16), can lead to results of high accuracy. To use the HFR method a suitable basis set must be chosen. The basis used in these calculations is the STO-3G basis of Hehre, Stewart, and Pople (16) with their recommended values for the exponents of the Gaussian func- tions. Details of this basis are in the Appendix. When the HFR equations were solved for formylnitrene, a state of A' symmetry was obtained. The molecular orbitals for this closed shell SCF (CSSCF) solution are depicted in Table I—l. The core consists of eight a' orbitals which are: either 5 type nonbonding orbitals or sigma bonding 12 om.om| fill.mH mm.ma| .mN 5H.HH- .mM vm.al .MV vo.a| mmool .Mm mm>.| .Mm omm.- .ms mamf [Imam mmv.| 00: =MH svm.- ob _Mm hmmf mom CH poemsoooa zm =mN manuanuo m3. « mom E 338098 20 .23. ..w>fl..om.. Ham. ow: .mm «mm. zwo .mHH mam. mamuflnuo Hmsunfi> towns: mwo .mNH mam. owe .MMH Admv msam>cm Hm mmmmmmmmmmm coaumucmmmu mm Hmuflnno msnfiosemuuH wcmuuwcamauom mo mamuflnuo H-H magma HMHSOOHOE l3 orbitals. The 9 a' orbital is a lone pair localized on oxygen and is denoted do. This orbital is perpendicular to the carbon-oxygen bond. The same nomenclature is used for the 10 a' orbital and since it is localized on nitrogen, perpendicular to the CN bond, it is labelled ON. The 2a" orbital is essentially localized on nitrogen and it is designated as pN, as suggested by Hoffmann (17). The fact that the ON orbital is unoccupied in the CSSCF function is a shortcoming of the single determinant repre— sentation of the wavefunction. This deficiency was removed in the configuration interaction phase of the calculation where a linear combination of Slater determinants is used as a representation of the wavefunction. The assumption which is used to generate the CI func- tion is that an adequate representation of the low—lying states of formylnitrene can be obtained by maintaining the lowest eight 0 orbitals of the SCF solution doubly occupied and allotting the remaining six electrons among the five "chemically significant" orbitals: "CO’pN’NCO’ OO’ON' These orbitals are either n types or lone pairs and they are not involved in the sigma bonding framework of the molecule. It is hoped that an adequate represen— tation of the low-lying states can be obtained by using a CI comprised of excitations within this orbital set. The CI wavefunction is comprised of 52 determinants of A' symmetry and 48 determinants of A" symmetry. This 100 (ieterminant CI function was used throughout the 14 computations on formylnitrene and it is listed in the Ap- pendix. This CI function is not expected to account for much of the correlation energy. The search for the optimum geometry was completed in two phases. In the first phase the angles 6 and w were varied over a span of 40°. The potential energy surface generated by this technique is checked to ensure that it encloses the minimum; the surface is then fit to a parabola since the potential energy, in the vicinity of a minimum, can be approximated very well by a parabolic function. The interpolated values for the energy and bond angles are obtained from this function. After obtaining the equilibrium bond angles, the search for the equilibrium bond lengths was begun. The bond length search was attempted at the best com- puted angular geometry for each species. A set of reason- able’ bond lengths was assumed for the CO and CN moieties and then these lengths were varied in the same way as was done for the bond angles. The same parabolic approxima- tion was used to compute the equilibrium bond lengths and energies. A CI calculation was then undertaken at the predicted minimum geometry to check the accuracy of the approximated energy and the results were found to be in good agreement with the interpolated values. The final values for the bond lengths and bond angles as a function of electronic state for various carbonylnitrenes are shown in Figure I—3. 15 FIGURE I-3 Calculated Geometries of Carbonylnitrenes 16 9“. 13“ szuwl m“ x MN? IA" Iaal m’ 1.5an “6' IA! Figure 3A" T '0 3 mg I' 1.904 '23. c I23. '38.: Lad r/M\N IA" 123' I 123' F/m’\N IA! 124. I 120' F/HG'\N a! 17 Since it is known that some of the excited states of formaldehyde are not planar, a few calculations at non— planar geometries were attempted for formylnitrene. The non—planarity was induced in the molecule by bending the hydrogen atom above the plane of the CON group. This results in a molecule with no symmetry elements except for the identity so there is no factoring of any matrices into symmetry blocks to facilitate the calculation of the SCF or CI wavefunction. Angles of up to 20° were included in the calculation but no significant non—planar geometry was detected in the three lowest electronic states of formylnitrene. The results of this calculation for the three lowest states of formylnitrene are shown in Figure I—4. The energies of the various low—lying states of some carbonylnitrenes are presented in Table I—2 and the natural orbital occupation numbers are listed in Table I-3. The natural orbitals, wi, are obtained by diagonalizing the first order density matrix (15) for each CI wavefunc- tion. In the natural orbital representation the electron density at a point, R is given by _ 2* o<§) — Zniwi(R) The natural orbitals give a concise description of the electronic distribution of these molecules. Since the first eight sigma molecular orbitals of formylnitrene are 18 FIGURE I-4 Energy of the Lowest States of HC(O)N as a Function of the out-of-plane Bending Angle. Energy+165 (au) ‘2455 1 I A f465 ]N A q “2475 <2 —.535 3/33 C —.545 0° 10° 200 Bending Angle Figure I-4 ___l_ .muumfiomo =¢m 020m um omusmsoo mum mmwmumcm 20 Hwnuo .mwwmuwcm Edflunfiafisvm man» was 020m .ozom mo mmumum .«a .=¢H .=¢m How mmmewGM% omamm.ss~n msumm.mo~n smmmm.~mmu mmmmm.mman .. may NIH OHQMB MO mowmumcm wmusmfioo 21 ooo.o moo.o soH.o ama.o oao.o oma.o nN~.o sod.o mao.o HNH.o aaa.o Hn~.o on. i cam.” SHI.H n~a.~ III.H noa.H NII.H HAI.H InI.H ~oa.a III.H HII.H hoa.H on: III.H aao.H nnm.a mam.~ ooo.~ ooo.~ ooo.~ ooo.~ ooo.~ oco.~ coc.~ III.” In enn.fl a~n.~ ~ms.~ on~.H ~oo.~ soo.a moo.H Hoo.H ~oo.~ Hoo.H noo.H ooo.~ Ia mao.o hoe.o Ion.o och.o ooo.~ ooo.H ooo.~ ooo.~ ooo.~ ooo.a ooo.~ ooo.a Io onIo nIu I I «Ioo nIo I I nIoo «Io I I o‘H 3‘." onVOx mo wouwum umm3oq mouse on» :w mumnasz coflpmmsooo Hmuwnuo Hausumz MIH manna 22 kept double occupied, they are not modified when the density matrix is diagonalized. They are regarded as natural orbitals with occupation number, ni, equal to two. In Figure I-5 the contour plots of several of the natural orbitals are displayed. The contour values are from .005 to .075 electrons/bohrB. The most striking feature of these plots is the degree of localization of the pN on nitrogen and the n and n* in the carbonyl region. CO CO Note that the largest nonoxygen contribution to 00 is the nitrogen lone pair which is localized directly on the CN bond axis. From Table I-3 it is noted that * n(nCO) + n(wCO) = 2.00 and n(oN) + n(pN) = 2.00 3 1 l for the A", A", and A' states of the carbonylnitrenes which indicates little charge transfer between the nitro- gen atom and the carbonyl group. The Mulliken population analysis results are pre- sented in Table I—4. na., the gross atomic population due to in-plane electrons, and na"’ the gross atomic population due to out-of-plane electrons, are listed for each atom in the molecule. 5, the charge on the atom 23 FIGURE I-5 Contour Maps of the Electron Density Associated with the Natural Orbitals of HC(O)N. 25 mo.l om.H om.m mo.l mm. mo.o no.1 mm. mo.m 2 0H.1 HH.H mm.o mo.l oo.a no.5 oH.I no.H mo.h O OH. mm. mm.v mo. om. oh.v HH. mm. vm.v 0 no. IIII mm. ho. IIII mm. mo. IIII om. m 0 =mc cm“ 0 :wg .MC 0 3mg pm: .da =¢H :d Scum mcmnuflqamauom mo mamaamcd cowumasmom oIH OHQMB 26 is also presented. The results show hydrogen and carbon to be positively charged while oxygen and nitrogen are negatively disposed in the three lowest electronic states. These results also imply that little charge transfer has occurred. B. Fluorocarbonylnitrene Other carbonylnitrenes were studied to determine the effect of substituents on the energy levels and orbital occupancies of these species. The fluorine derivative was the most extensively studied molecule of this set. The STO—3G basis was used and the CSSCF solution for the 24 2 2 2 (core) "COOOpN resultant molecular orbitals resemble the orbitals of configuration was constructed. The formylnitrene and these orbitals were used in the CI cal— culation. 3A" of formylnitrene was The computed geometry of the used as an initial approximation for the geometry of FCON. The CF bond length was set at 2.54 bohr. The first CI function was constructed with the hope that all single excitations from a selected set of occupied orbitals into the virtual orbital space would produce an adequate representation of the wavefunction. The results of this attempt demonstrated that single excitations from the fluorine o lone pair was of little consequence. This CI also showed that there was very little tendency for 27 highly antibonding orbitals to be populated in the lower electronic states. The CI calculation did show that certain configurations, which were very important in describing HCON, were also important in describing the states of FCON. The second CI wavefunction was based on the premise that the four reference configurations shown in Figure I-6 are the basis for a good representation of the low-lying states of XCON. To improve the wavefunction all single and double excitations are taken from these configurations. '1 The orbitals of interest are NF, flco, 00’ pN which are occupied in the HFR solution and n20 and ON which are unoccupied in the CSSCF function. There are 225 deter- minants that can be generated from this orbital set and these determinants were combined to form structures which are space and spin adapted combinations of Slater deter- minants. The structures were separated into singlet and triplet manifolds and the CI calculation was performed on each spin level. The geometry of the molecule was optimized and the results are presented in Figure I-3 and Tables I-2 and I-3. The optimum CF bond length is 2.53 bohr in the 3 A" ground state. This bond length was not optimized in the excited states of FCON. The computed angular geometries for the states of FCON are similar to the geometries of the corresponding states of HCON. Contour maps of the electron density associated with the natural orbitals of FCON are displayed in Figure I-7. Because each of these orbitals qualify as CO orbitals 28 FIGURE I-6 Reference Configurations for Carbonyl— nitrenes 30 FIGURE I-7 Contour Maps of the Electron Density Associated with the Natural Orbitals of FC(O)N 32 (they are in-phase and out-of-phase linear combinations of GO and CF) they are both used as active orbitals in the CI. The orbital contour maps show that the electron density in FCON is localized in specific regions of space, and this localization of electron density is similar to the localization of electron density which occurs in HCON. The relative energies of the electronic states of fluorocarbonylnitrene follows the pattern set by formyl- nitrene. The 3 l A" is the ground state, and there are two singlets, A" and 1A', separated by 5 kcal/mole, which are approximately 45 kcal/mole above the ground state. Figure I-8 depicts the energies of the excited states relative to the ground state energies for various car- bonylnitrenes. To further indicate that substitution of fluorine for hydrogen has little effect on the relative energies of the states of these carbonylnitrenes, we computed AB: for the reaction F(2P) + HCON + H(23) + FCON The electronic state of the carbonylnitrene is not 3 allowed to vary in this calculation; A" HCON produces 3A" FCON. The results (Table I-S) show that AB: is fairly constant for the three lowest electronic states. The AES indicate that fluorine substitution destabilizes the electronic states of FCON relative to the states of HCON, 33 FIGURE I-8 Relative Energy Levels of Carbonylnitrenes 4 3 20oonIo «m & 3 .<_ “A onnIu Av: as muH wusmflm 200... 28: GmoFeIz aerz l<fl a<fl IN” IN” 8v x. 3.9 as -\ I .<_ i d. 83 ||I| 83 .<— I Anvv <— s a. so as .unt of n electrons in the A' state. This may be due tc>.f1uorine stabilizing the pg configuration relative to 2 01‘]. 36 oa.l mm.a Hm.h MH.I Hm.a NN.N MH.I Hm.H NN.n m mH.I vm.H mm.o oa.l NH.H mm.o HH.I ma.a mm.o 0 v0.1 mv.a Ho.m vo.l mm. mo.m mo.l mm. mo.m Z om. mm. vh.v hm. mm. vh.v mm. mm. vb.v 0 Q : mg .mc w 3mg .mc mu =wc .wc .AQH :fin—H =¢m ECU.“ mcmupficamconnwoonosHm mo mammamca coflpmasmom oIH OHQMB 37 C. Methylcarbonylnitrene Methylcarbonylnitrene was selected as a molecule of interest because the methyl group is an electron donat- ing moiety and it would be interesting to examine the effect of the methyl group on the electronic structure of carbonylnitrenes. The methyl group was assumed to be of tetrahedral disposition with CH bond lengths of 2.06 au. The geometry of the CON group was chosen as the com— 3A" state. No puted equilibrium geometry of HCON in the geometry search was attempted for this molecule due to the amount of computer time and money it would have involved. The HFR equations were solved for the p; configuration and a selected number of single and double excitations from the reference configurations (Figure I—6) was used in the CI. The aim was not to be complete, but to be consistent with the other molecules in the series. The results of the calculation are presented in Tables I-2 and I-3 and in Figure I-8. In Table I-2, note that 3 1 1 only the energies of the A", A", and A' states of HCON and FCON correspond to optimized geometries for those states. All other energies in the table were computed at geometries which closely approximate the optimal geom- etries of the molecules. The energies for 1A'* and 3A' states of FCNO and HCNO were computed at the geometry of ‘the 1A' state. The energies of the methyl and methoxy charivatives were computed at the optimal CON disposition of HCON in the ground state. In methoxycarbonylnitrene, the orbital chosen as "x is the antisymmetric combination of hydrogen 5 orbitals with the carbon p. /® mH \o The "active" orbitals are "X’NCO’ pN, 00, ON, NEG, and are used to generate determinants for the CI function. The population analysis is in Table I-7. The computed energy level pattern is not much different from the other nitrenes that were studied, and this fact implies that the carbonyl group has an insulating effect on the nitrogen, shielding it from perturbations due to various substitu- ents. Since the electronic structure of the carbonyl- nitrenes depends almost entirely on the electron distribu- tion on nitrogen, it follows that all these species will be very similar in structure and reactivity. 1A' state and there The most interesting state is the is about 1.5 electrons in the pN orbital of this system. In other respects, methylcarbonylnitrene is very similar to HCON and FCON. The charge on the carbonyl carbon is .18 electron. 39 mo. III: mm. mo. IIII mm. mo. 1111 mm. m HN.I mo.H mH.m mm.l vo.H ma.m mm.l vo.H ma.m .0 hH.I mH.H No.5 HH.I mo.H mo.n ma.l 0H.H mo.n O oo.l hv.a mm.m no.1 mm. mo.w no.1 mm. mo.m 2 ma. mm. mm.v ma. mm. mm.w ma. mm. mm.¢ 0 @ 2mg .mc 0 :mc .wg W :mc .mc .«H : fiH 34m EO“4 msmnufisawconuwoamnumz mo mammamcd c0apmHsmom suH magma 40 D. Methoxycarbonylnitrene This molecule, which closely resembles ethoxycarbonyl- nitrene, can provide some insight into the chemistry of the experimentally observed species. The STO-3G basis was again used and the CON geometry was chosen to be that of HCON in the ground state. The methyl group was assumed to be tetrahedral while the CO'C bond angle was chosen to be 111°. The CSSCF solution was for the 1A' state with 30fl2 G2n2 2 o' o COpN' orbitals used to generate the CI function were no,, 0 the configuration (core) The six "active" 0’ NCO, pN, 00, "20. The same technique is used to form the CI function as was done in previous cases, i.e., use selected single and double excitations from the reference configurations (Figure I-6). The results for this molecule are contained in Tables I-2, I-3 and I-8, and in Figure I-3. The CI function con- sisted of 87 structures of A" symmetry and 142 structures of A' symmetry. This species is about the same as the 1 other nitrenes except that for this molecule the A' and 1A" states are almost degenerate. The carbonyl carbon carries a charge of .24 electron and the methoxy oxygen is more negatively charged than the carbonyl oxygen and the excess charge is gathered through the sigma bonds. There is very little donation of the methoxy oxygen's n electrons. 41 oa. IIII om. OH. 1111 om. 0H. IIII om. m vm.l vm.H o¢.o om.l mm.H H¢.o oN.I mm.H H¢.o .0 mo.| mo.H mo.m wo.l mo.H mo.m mo.l mo.H mo.m .U bH.I vN.H mm.o MH.I 0H.H hm.m vH.I hH.H hm.o 0 oo.1 mN.H mh.m «0.1 mm. mo.m v0.1 mm. mo.o z ow. om. mh.v vN. mm. mh.v on. mm. mb.v U 0 =mc - ”a 0 = mg . ”C Q =mg - ”a .4H 3 «H = + Cl|0N> + 1:2 CiloNoi> 2 10 wherele>==}4I(core) pNapNB, etc. The ci's were de- termined variationally and the natural orbitals were con- structed. The resulting NO's were used as input for another CI (the iterative natural orbital technique (15.22)) and the natural orbitals of this second iteration were taken as appropriate starting orbitals for a more extensive CI. The SCF and CI comparisons for the two basis sets are 1 shown in Figure I-lO. The SCF energy of the A" state is 3 taken to be the ROSSCF energy of the A" state plus twice the exchange integral between the nitrogen o and p orbitals 3 1 (computed with A" orbitals). For the A' state, the SCF energy is taken as the result of a two by two CI using the p; and a; configurations with the best set of natural orbitals obtained by the INO technique. The dramatic lowering of the 3 A' state demonstrates the changes in energy level patterns that can be obtained by using orbitals appropriate to the state of interest. To construct the final CI function for each of the four low-lying states, the number of orbitals was reduced 49 FIGURE I-lO Comprison of the Results of the STD-3G Basis with the [35,2p/Zs] Basis. [5(kaflfiwde) 9O 70 40 IO 50 , II! ' (90) j\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ 3A' (73.5) \ ‘s \ I ' * s‘ A (70.5) 3191' (655) ‘A’ (47.6) \ 'A' (35.5) 3A0 3A0 CI(STO-36) c1(9s,59/2s) Figure I-lO “a SCF(95,5 P/2S) 51 by eliminating highly antibonding functions localized on a single center and by maintaining the first three orbi— tals, which are essentially ls orbitals on C, N, and O, doubly occupied. This reduced the number of Mo's from 29 to 22. The approximate first order (15,23) wavefunction was then generated. The correct first order function contains three classes of excitations: 2 . . . wi + wixi Single exc1tations wi + wiwk doubles in valence shell 0% + wkxi doubles out of valence shell In these equations, w is a valence-like orbital and X is any other virtual orbital that is not a valence type or- bital. The first order wavefunction picks up that part of the correlation energy which is associated with low- lying valence orbitals which are not fully occupied in the SCF solution. In the case of formylnitrene the first set of excitations described above are omitted since the SCF functions are of good quality. Matrix elements between the SCF structure and structures formed by these single excitations will be very small or zero. In the third class of functions, excitations which involved promotion of electrons from two different valence orbitals were excluded (split-shell excitations). The only exceptions to this were excitations from the pNoN configuration. The second class of excitations is the most important in 52 obtaining good energies and all of these were generated. Of all the structures, the best one hundred were chosen for each electronic state. This was accomplished by ordering the structures on the basis of a two by two CI with the structure representing the first natural con- figuration of each electronic state, e.g., the pNON con- 3A" state. The CI calculation was then carried figure the out with these structures. The results are presented in Figure I-lO along with the CI results for the STO-3G basis. Although the actual energy in the larger basis is 1250 kcal/mole lower than in the STO-3G basis, the energy level pattern is only slightly affected. The actual energies in the larger basis are presented in Table I-9. The natural orbitals substantiates the minimal basis results with the main exception being the occupation of the o 1 N and pN or- bitals in the A' state. In the minimal basis these or- bitals host .74 and 1.26 electrons, respectively, while in the more flexible basis, the occupations are .61 and 1.38 electrons. Although the asymmetry of the nitrogen charge distribution is increased, the results of the [35,2p/Zs] basis concur with the minimal basis results in that they predict a significant 0 occupancy. The popula- N tion analysis results are presented in Table I—lO. The triplet-triplet transition energy is now 65 kcal/mole or about 75% of that for NH. All other conclusions drawn from the minimal basis results are substantiated by this more complete calculation. Energy of the Low-Lying States of HCON Computed with the [352p/25] Basis 53 Table I-9 3A" ~167.64519 1. A -167.58862 1.. A '167.58097 3. A -167.54076 Vertical Transition Energies of HCON in the [332p/Zs] Basis au eV kcal/mol 1 . 3.. A + A .0566 1.539 35.50 1" 3" A + A .0642 1.747 40.30 3 . 3.. A + A .1044 2.842 65.53 1A" + 1A' .0076 .2082 4.80 3.. 1.. A + A .0402 1.094 25.23 54 mm. mm. mm. 11:: mm. mm. 1111 mm. m sm.l mm.H vo.> mm.| mm.H oo.w mm.l mm.H mo.n o oa.1 mN.H Hw.m HH.| om. mH.m Ha.» om. ma.m 2 OH. on. wH.m mo. be. va.m mo. en. va.m o I inc .0: m =mc .mc I :0: .mc .fiH =mH :mm eoud mammm flmm\mmmmH cw ozom mo mflmwamcd coHumHsmom oHIH OHQMB 55 G. Carbonyldinitrene Calculations were carried out on this molecule mainly because of the unusual properties it might possess. Each nitrogen atom in the molecule has the potential of retain— ing two unpaired electrons which implies that the mole— cule as a whole can have a spin state with four unpaired electrons. This would lead to a quintet ground state which would be very unusual. The basis set is the familiar STD—3G basis and the geometry has the angular disposition of formylnitrene along with its equilibrium bond lengths. In this particular geometry, the molecule has C2V symmetry and the orbitals are labelled as the one dimensional representations of the point group. A CSSCF calculation reveals the orbital ordering shown in Table I—ll. A CI calculation used the lbl' 8al, 4b2, 2bl, laz, 5b2, and 3bl molecular orbitals as the "active" orbitals from which the determinants were formed. A full CI within this orbital set would require 1225 determinants. In order to reduce the number of determinants, two restrictions were imposed on the orbital occupancies: l) The lbl (COW) orbital is either doubly or singly occupied at all times, and 2) If lbl is singly occupied, then 3bl (C0;) is also singly occupied. If lbl is doubly occupied then 3b1 is not occupied. 56 Table I-ll Molecular Orbitals of CNNO * 3b1 COTr .311 5b2 NO virtual .008 1a2 NTr .001 2b1 NTr -.235 4b2 00 occupied —.315 8al N0 -.348 7al sp bonding -.498 lb CO -.506 11’ 57 This restriction reduces the number of determinants to four hundred, which are distributed in the following manner: 106 of Al symmetry 98 of B1 symmetry' 98 of A2 symmetry 98 of B2 symmetry The energies of the states are presented in Table I-12 along with the vertical transition energies. Interestingly, the ground state is a 5 A1 and therefore there are no di— pole allowed transitions from the ground state to any excited state. Contour plots of the 8a and 5b2 orbitals 1 are shown in Figures I-11 and I—12. These orbitals are localized on each nitrogen which again demonstrates the insulating effect of the carbonyl group. Population analysis on the ground state shows the nitrogens hosting the largest negative charge (Table I—13). The most important aspect of this molecule to be pur- sued is the calculation of the correct equilibrium geometry. By assuming the geometry of the 5Al state of CNNO to be the same as the geometry of 3A" of HCON, we may have intro— duced a large error. The electrons on each nitrogen may interact thereby causing the molecule to reduce the NCN angle significantly. There may actually be formation of an N—N bond causing the molecule to be a closed ring. 58 Table I-12 Energy of States of CNNO Al -218.57536 3 B2 ~218.54993 3 A1 -218.50810 1 A1 -218.50500 Vertical Transition Energies of CNNO au eV kcal .0254 .692 15.96 .0673 1.83 42.20 .0736 2.00 46.18 .0419 1.138 26.24 59 FIGURE I-ll Orbital Contours of 5b2 Orbital of CNNO. HH-H mucosa 61 FIGURE I-12 Orbital Contours of 8a1 Orbital of CNNO. 63 Table I-13 Population Analysis of Carbonyldinitrene 5 Atom A1 110 nn 5 C 4.83 .94 +.23 N 6.14 .96 —.10 N 6.14 96 — 10 O 6.89 1.14 —.03 PART II KINETICS AND MECHANISM OF THE FORMATION OF A NON-CYCLIC NICKEL(II) COMPLEX INTRODUCTION During the past fifteen years, a considerable amount of work has been completed regarding the synthesis and characterization of transition metal complexes containing macrocyclic ligandsZ6-32. Some of this research has originated to find new new synthetic methods that can be used to form these metal complexes in higher yields, while other research has been concerned with the measurement of physical prOperties of transition metal macrocyclic com- plexes in order to elucidate the mode of bonding in these systems. Many important biological systems are known to contain a metal ion in macrocyclic environments, and there is considerable interest in the chemistry of these naturally occurring transition metal macrocyclic complexes. In general, the true biological process occurs in a compli- cated environment which cannot be observed effectively; if the active sites of these processes can be identified, model systems that mimic the behavior of the active site can often be devised33-36. Synthetic transition metal macrocyclic complexes can be used in the study of these biological sites, and by understanding the nature and function of the model, we can infer information about 37 the actual biological process . Transition metal complexes containing macrocyclic 64 65 ligands are also useful as model systems in the study of other chemically interesting problems. Some macrocyclic ligands can stabilize metal ions in unusual oxidation states such as Co(I)38, Ni(III)39, and Cu(III)40. The metal complexes are usually prepared with the metal ion in its "normal" oxidation and then the metal is oxidized or reduced by electrochemical methods. Macrocyclic complexes are also useful as models for the study of ligand interactions with the axial components 41’42. Because the metal is coordinated of metal d orbitals to the macrocycle, the planar environment around the metal is essentially static, and any observed spectral or mag~ netic changes can be assumed to be related to axial elec— tron redistribution. In this manner, one can selectively study these processes closely related to the naturally occurring ones, such as the interaction of small diatomic gases with metalloporphyrins. Macrocyclic complexes of transition metals can be formed in two convenient ways: the first is reaction of a metal salt with the previously prepared macrocyclic 43,44 I ligand and the second is the reaction of the organic components, in the presence of the metal ion, to form the 45’46. The desired transition metal macrocyclic complex latter method is usually governed by the "template" effect because the metal ion acts as a template by holding the organic reactants in the required orientation for reaction 66 to take place. This is a reaction of a coordinated ligand and the metal ion directs the course of reaction. In many cases the presence of the metal ion is necessary to produce the macrocyclic ligand; reaction of the organic species without the metal ion leads to nonmacrocyclic products. The template effect was used by Kerwin and Melson in the reaction between acetone and benzil monohydrazone, in the presence of nickel(II) ions, to form the complex, a,a"' -(isopropy1idenebis(azo))di-a-stilbenolatonickel(II), NiMMK, (Figure II-l)47'48. The condensation reaction, which forms the complex, takes place between the amine moiety of two monohydrazones and the carbonyl group of acetone. In the absence of nickel(II) ions, the product isolated from the reaction mixture is benzilacetone azine. C6H5\ c6115 //C_ \\ O NN=c(c (13);, Although considerable technique has been developed in the synthesis of transition metal macrocyclic complexes, there is a paucity of data concerning the kinetics and mechanisms of these processes. Since many transition metal macrocyclic complexes result from the reactions 67 FIGURE II-l Reaction of NiMMK With 1,3-propanediamine to Form NiApSo* *There is no evidence that the amine substitutes at the five membered ring. Substitution can occur at the six membered ring also. 68 HIHH whflmflm WIOU .” —N_ ome< _ z I 2 s: z Fw— _~_ J F / \ I \UHU/ /U ”U/\ NIZINI oxz z / 0\ z / \ / z zmI o z PNI o\ /z\<~_ ._. I _. I _ _.~_ _ 69 of coordinated ligands, the kinetics and mechanisms of these reactions are included in this discussion. The first kinetics study was carried out by Blinn and Busch49 on the reaction of benzyl bromide with various nitrogen substituted bis(B-mercaptoethylamine) nickel(II) complexes. Since the benzyl moiety adds to the sulfur atom of the complex, the reaction is one of a coordinated ligand. Blinn and Busch propose a two step process for this reaction: the first step is a pre-equilibrium between the nickel complex and benzyl bromide, while the second step is the actual attachment of the benzyl group to the coor- dinated sulfur atom. A five coordinate intermediate in which the benzyl bromide occupies the axial position of a square pyramid is prOposed. AH* is between 8.8 kcal/mole and 10.1 kcal/mole, while AS* ranges between -25 en and -35 en for the various nickel complexes. A second study, also by Blinn and Buschso, concerned the ring closure of 2,3-pentanedionebis(mercaptoethylimino)- nickel(II), Ni(PE), with a-a'-dibromo-o-xy1ene. In this reaction the product is a transition metal macrocyclic complex in which each of the sulfur atoms of the nickel complex react with the dibromo—o-xylene to form a new chelate ring. The rate of this reaction was shown by Blinn and Busch to be influenced by the nickel ion which holds the mercapto groups in the cis positions, and this directive influence was called the "kinetic coordination template effect" by the authorsSI. They were able to 70 ascertain that the first step of the reaction was slow and the second step was rapid; once one sulfur had reacted with the reagent, the second sulfur and halide were in position for the ring closure to take place. Since the rate of the second step (ring closure) is rapid, there is no measurable concentration of intermediate formed and it was not detected. As in their previous study49, the activa- tion parameters support the two step mechanism. AH* was 9.6 kcal/mole while AS* was -27 eu. This latter number signifies an activated complex which may be formed by an associative process such as the authors propose in their mechanism. Burkesz, et. al. studied reactions of bis(B-mercapto- quinoline) nickel(II) complexes with various organic halides. This work was an extension of the work done by Blinn and Busch, discussed earlier. Burke, gt_31. computed activation entropies of approximately -35 to -40 eu. These values indicated an associative mechanism and they proposed mechanisms similar to those of Blinn and Busch in which the alkyl halide coordinated in the axial position of the nickel complex. To test the effect of solvent on the 153 used a solvent mixture of rate, Burke and Campbel benzene and methylene chloride in which the mole fractions of each component were varied to give a series of solu- tions of differing dielectric constant. The reaction rate of the mercaptoquinolinenickel(II) complex was measured as a function of dielectric constant and the data were 71 interpreted to indicate an activated complex that was less polar than the reactants. This was also evidence for an associative mechanism in these reactions. Another kinetics study was done by Leussing and 54 on the formation of zinc(II)-Sa1icy1a1dehyde McQuate Schiff bases with ethylenediamine and 1,3—propanediamine. Their results covered a wide range of reaction conditions and implied a reaction path which was second order in amine concentration. The intermediate, which was proposed for this reaction, was a "carbinolamine", i.e., 55 obtained some preliminary data for Nafisi—Movaghar the reaction of NiMMK with ethylenediamine (en) and 1,3 propanediamine(1,3-pn), and his results showed that the reaction of metal complex with 1,3—pm did not produce a macrocyclic ligand but the reaction of the metal complex with en did produce a transition metal macrocyclic complex. The reaction to form the transition metal macrocycle was observed to occur in two measurable steps; the build- up of an intermediate, shown to be similar to the product of the reaction of NiMMK with 1,3—pn was detected. There~ fore, the reaction between NiMMK and 1,3-pm may be related to the first step in the formation of these transition metal macrocyclic complexes. If this first step can be 72 understood, attention can be focused on the second step in which the transition metal macrocyclic complex is formed, and a complete mechanism can be formulated which will describe the formation of the macrocyclic product. SCOPE OF THE THESIS WORK The reaction of the metal complex, NiMMK, with 1,3- propanediamine was studied in detail, and factors which influence the rate of reaction were examined so that a complete mechanism which is consistent with the experi- mental facts may be pr0posed. The reaction scheme is shown in Figure II-l. The product, [a'-[[l-[[2-[(3-Aminopropyl)amino]-l,2-dipheny1- vinyl]azo]—l-methylethyl]azo]-a-stilbenolato]nickel(II), NiApSo, is known to be formed quantitatively from the start- ing material48. The absorption spectra of both NiMMK and NiApSo is presented in Figure II-2. The rate of reaction was monitored at 340 nm and 390 nm, which correspond to absorption maxima in the spectrum of NiMMK. 73 74 FIGURE II-2 Absorption Spectra of NiMMK and NiApSo at Equal Concentrations. a = NiMMK b NiApSo 75 000 own oon omv NIHH musmflm 8.5 I oov m Com m mmm oom mum OAV NAV .Eo oAV on 0; mg 0; m._ AHN aouquosqe EXPERIMENTAL A. Preparation of Materials NiMMK. The metal complex, d,a"'-(isopropy1idenebis(azo))- di-a-stilbenolatonickel(II), (NiMMK), was prepared by methods previously reported by Kerwin and Melson47. The red-orange needles were recrystallized from l-butanol and the identity of the complex was confirmed by mass spectrometry and by its UV-visible spectrum. Solvents. Tetrahydrofuran (THF) was refluxed over calcium hydride to remove water. It was then distilled from cal- cium hydride and stored over molecular sieves (Linde Type 5A) in a brown, stoppered bottle. Absolute ethanol was used without further purification. Amines. 1,3 Diaminopropane(l,3-pn) (Aldrich) was twice distilled from sodium hydroxide under an atmosphere of dry nitrogen, and then stored in a dry nitrogen atmosphere. All other amines (Aldrich) were stored under dry nitrogen and were used as supplied. Sodium Hydroxide Solution. An ethanolic solution of sodium hydroxide was prepared by dissolving 7.2 grams of reagent grade sodium hydroxide pellets (Mallinckrodt) in cold, de- gassed, absolute ethanol. 300 milliliters of degassed ethanol was measured into a Schlenk flask which was fitted with a gas adapter. The flask was placed in an ice bath 76 77 and dry nitrogen was allowed to flow over the solution, while the sodium hydroxide pellets were slowly dissolved in the ethanol, which was stirred by a magnetic stir plate. Time for complete dissolution of 7.2 grams of pellets was about six hours. After the pellets had dis— solved, the stopcocks on the flask and gas adapter were closed and the flask was removed from the ice bath and transferred to a dry box which had a dry nitrogen atmos- phere. In the dry box the solution was filtered by suction through a medium fritted funnel. The filtered solution was transferred to a plastic bottle which had been capped with black tape to prevent light from affecting the solution. The bottle was stoppered, taken out of the dry box, and stored in a refrigerator. Exactly 25 m1. of this solution was diluted to 250 ml. and 25 ml. aliquots of the dilute solution were standardized by titrating against dry po- tassium hydrogen phthalate to the phenophthalein end point. In the other methods we used to prepare the sodium hydroxide solution, extensive decomposition occurred within a few days. This decomposition was due in part to the hydroxide catalyzed oxidation of ethanol, which was evi- denced by a change in color of the solution from colorless to brown. The procedure presented above was the method of choice in the preparation of the ethanolic hydroxide solution, because the solution was more stable and decom- position did not occur for at least three weeks. 78 Dielectric Constant. The dielectric constant of each solu- tion used in the kinetics experiments was obtained at 25° with a Wissenschaftlich-Technische Werkstatten Dipolmeter Type DM-Ol. A calibration curve of instrument reading as a function of dielectric constant was obtained from the ex- perimental readings and the literature values of the di- electric constants of ethanol, THF, methylene chloride, and l—propanol. B. Procedure for Obtainingythe Kinetics Data For each kinetics experiment, about 0.0845 grams of NiMMK was weighed and dispensed into the reaction flask, which was a 100 m1. round bottom flask fitted with a standard 24/40 ground glass joint, 10/30 ground glass thermometer port, and side arm fitted with a stopcock. THF was added and the nickel complex was dissolved. Speci- fied volumes of ethanol and the ethanolic sodium hydroxide solution were then added to the solution, and a condenser fitted with a drying tube containing Mallcosorb to remove CO2 and Drierite to remove H20 was attached to the flask. The solution was heated to reflux temperature (68°) with a heating mantle, and a specified quantity of 1,3-pn was injected into the flask through the side arm, and the timer was started. A summary of the experimental conditions is in Table II-l. Atspecified time intervals, 0.23 ml. aliquots were withdrawn from the reaction mixture with a hypodermic 79 as 0.4 He es 0.4m as ea mmmcmno meme. moo. mmosxm muspmumIEme as I.N Ioum>uom m.mm mmaum> N.Hm.we momm. mac. mMOme Iona QQIMIH o . ol o II o II \ mmaum> >ue me 8 mm as ea m +m we mmaum> Sac zmoaxm calm H as m.m es e.~m as on N.Hm.me momm. moanm> mmoaxm -Io calm H> mmB> quOB> moum> Aoove :QIm.H hnmo H mwzg omaum> muflucmso musmaflnwmxm mowpmcfix How wcofiuflocoo HIHH OHQMB 80 syringe fitted with a Chaney adapter. The aliquots were injected into 10 ml. volumetric flasks containing ethanol at room temperature, and diluted to the final volume with ethanol. The concentration of nickel complex after dilu— tion was approximately 8 x 10"5 M. The absorbance of each sample was obtained at wave- lengths of 390 nm and 340 nm with a Beckman DB-GT grating spectrophotometer which was calibrated for wavelength with a Holmium filter. In some cases spectra were also re- corded as a function of wavelength on a Unicam SP800 spectrophotometer, which was also calibrated with a Holmium filter. The absorbance data which were used to extract the rate constants were taken by use of the Beckman spectro- photometer since it is more precise. Agreement between the two instruments was usually 0.05 absorbance units. For the experiments that were not carried out at reflux temperature, a thermostated oil bath which was kept constant to i.1°C was used. The bath was equipped with a variable speed stirrer and a two component thermostatically con- trolled heating element. An atmosphere of nitrogen gas, saturated with THF and ethanol continuously flowed through the condenser. This was necessary to prevent air, which decomposed the complex under these conditions, from enter— ing the system (Figure II-3). The reaction mixture, without 1,3-pm, was placed in the bath and the temperature of the solution was allowed to become constant before the amine was added and the reaction initiated. A thermometer, 81 FIGURE II-3 Experimental Apparatus 83 which was immersed in the reaction solution, measured the temperature which was approximately 1° lower than the bath temperature. The reaction mixture was stirred by a Teflon coated stirring bar which was activated by an airIdriven magnetic stirrer which was immersed in the oil bath. Four milliliters of 1,3-pn, preheated to 60°C, were injected into the reaction vessel through the sample port by means of a hypodermic syringe. Aliquots were then withdrawn at selected time intervals and analyzed as previously described. C. Treatment of Data In all kinetics experiments the concentrations of 1,3- pn and sodium hydroxide are greater than the NiMMK con- centration and therefore the reaction is characterized by pseudo first order kinetics. The absorbance of NiMMK is directly related to its concentration and the change of absorbance in time is a measure of the rate of reaction. The absorbance-time data, which were obtained from a particular kinetics experiment, were treated by using KINFIT, a nonlinear curvefitting computer program. The data were fit to the expression for a first order decay process (A — Am) = (A0 ~ Aw)e"kobst 84 by using the KINFIT program and the Michigan State Univer- sity CDC 6500 computer. In the above expression, t is the time andkobS is the observed first order rate constant. A0, Am, and A represent the absorbances eat zero time, in— finite time and any other time, respectively. Usually, fifteen pairs of absorbance—time data were input for each kinetics experiment. The variances (square of the standard deviation) in the absorbance and the time were also input to the program. These variances were estimated at (.005)2 for the absorbance and (.01)2 minute for the time. The computer program found the solution to the decay process by an iterative method and it gave as output the best estimates of k, A0, and Am. The calculated A0 and Am usually agreed with the observed values to within 1-3%. The standard deviation in k was always less than 10% for any acceptable data set. The observed rate constants were calculated at both 390 nm and 340 nm to check that the data were consistent. The rate constant at 390 nm agreed with the rate constant at 340 nm to within 5%. RESULTS Depending on the initial conditions, the reaction between NiMMK and 1,3—pn to form NiApSo is effectively complete in two to fifteen hours. Spectra recorded during one of the reactions is shown in Figure II-4. In this figure, spectrum 1 is the spectrum of NiMMK and spectrum 6 is that of the product, NiApSo. The observed decrease in the absorbance of the starting material as a function of time is depicted in Figures II-5 and II-6. The calculated decrease in the absorbance, fit to the expression for first order decay, is also shown in these figures. The observed data usually fit this kinetics scheme to better than 7%. The half life for the reaction ranges from 10 to 180 minutes under these initial conditions. A. Hydroxide Dependence of the Rate The first reactant which is varied to obtain its re- action order is the hydroxide. The amount of ethanol in solution is kept constant at ten milliliters and cor- responds to a concentration of 3.26 M_in the final solu- tion, which has a volume of 52.6 milliliters. The 1,3-pn concentration is adjusted to 0.590 M throughout all these experiments and the concentration of hydroxide is varied from 0.027 M’to 0.092 M. These extremes represent hydroxide to NiMMK ratios of nine to thirty, so the hydroxide is in excess at all times. Duplicate experiments are made at 85 86 FIGURE II-4 Spectra Recorded During the Reaction of NiMMK with 1,3-pn. 87 1’1 J 1 1 1 l l 1 o m 9 .. cw o e ‘I <1 Q aoueqlosqe 600 550 500 450 400 350 325 300 .K(nnn Figure II-4 88 FIGURE II-5 Computer Fit of First Order Decay of NiMMK at 390 nm. 89 mnHH magmas con vwbccuc at u h muzwczwamc.wc:hdzuarwh x::»/ cattacstonoomoocstrccasc czu Fv-000m0000V0000m0000m0900V-0-0m0000m0000r000'm'---U0000m0003m0000m'--. m.... U.... U I '9'3'--- mp- III-luau--- 8% In.-. - m. n ”C. t-O-O-ht-U’ v—b o—oo—U‘t-v-t- u—V hunt—Hr v—t—I-I-vv I-vOI-‘O-‘h—U‘Mv—mo—o—t-u-Uo-uwu—V uni-op...- o-uu—t-vha—u puppy-pu—u—v—ou-o—HPW'WP-b—V'U'P Orv-thou o—v c- o- “ >-u-v-.-'u 'v-www- u-c—u—u—p.u IC“ m-a--mu--uv--u4won-nw-u-uuuuuuwuonumnuu-unuucvunonmuuoowuuu:wu-u-v-uunm-:-uu-u-uw---:m----v--acv--q .. > «you: >¢ x «pow: wzcm II» I“ was »2_co :wpcsrtsqt :vc La»:w:~cuoyu (I uzcw: n mcwbhcAa m» c—vua 2w!) cww: Vn n >JZC~ bznoo cub¢42L4.~oow~o_. n-0c» a1» pa wzoq: .zawta er .I.n.. v. .a epau> ccouawo. u hzuzwcoz~. comcwo. u bro—a mzp kc w:.<>.ocowo~m. u sumo or» ~< uzoas ..h.o¢:x.y ac so...» u. cocoa 'IGURZ 11-6 Co-putor H: of Hut Order Decay of mm at 140 nu. 91 esHH mucosa _. . FTQZ— k». C?“ . . oem vwwacwc me n b wc7wc2mowc hazhccuazur xsrb. otttottttttotttvto-tooo¢ czu wwu--v---wv000-r--0-v-.l00w---'W-‘00m0001w0000U0000V0-00m0000m0000wu'00'$-999V0008m'8'0U98I-0Mv0000m9000W ' n v wu—u—uu w-«o—o-oo—u war-«bunt moat-ow ‘P-O-HO-‘u mo—u ~ u—o—o—n—u 'p-I—o-w—u ~»op-p.g'>-p..—g r— u-u—b—u-U‘ u—h-n—vl—U‘ Hu—t—o—U' bur-Hr b-b-h- MU Mind“ WhU‘Hb-O—O—U‘t-o—Ho-U hush—1r 0— x C > «some >1 I (some wxaw m1» :_ mac >2_ca subtotuott eta qurm: shows If uzem: u mcwbbcaa w» c—vua tux) cwm: m~ u >42C. hzuca owbagthac a v7oooowmooo u-n:p mi» he mtg”: . ~Juo 2. ...a.» up .c..haw> ooouuwo. u hzuzwauz—o coucwoo u bro—a um» ha w:4<>oooowo~m. u hung II» ha u:4(> ..I.Inrx.y I: .p...y v» ucaua 92 selected concentrations and the agreement of the rate constants is less than six percent. The observed first order rate constants and their estimated standard deviations (0) in percent are presented in Table II—2. A graph of the average observed rate constant, (k390 + k340)/2, as a func— tion of hydroxide concentration (Figure II-7) implies a two step mechanism; the first reaction is a pre-equilibrium forming an intermediate which is used in the second, slower reaction. The reaction order changes from first order at concentrations below approximately 0.045 M to zero order at concentrations greater than 0.07 M. When other param- eters are varied, the kinetics experiments are performed at hydroxide concentrations of approximately 0.09 M; at this concentration, the observed rate is independent of the concentration of hydroxide. B. Influence of the Solvent on the Reaction Rate The effect of the composition of the solvent on the rate of reaction was examined by varying the ratio of ethanol to THF in the solvent system. The mole fraction of ethanol, X, in solution was varied from 0.154 to 0.382. These extremes were dictated by the reaction conditions: at less than six milliliters of ethanol in solution (X = 0.154), there was precipitation of sodium hydroxide; at more than sixteen milliliters of ethanol (X = 0.382) the rate of decomposition of NiMMK became too large to 93 Table II~2 Observed Rate Constants as a Function of Hydroxide Ion Concentration [03’] k390x102 k340x102 kobsxio2 “ (min-l) (min-1) (min-1) O(%) .092 1.87 1.72 1.80 i 5 .091 1.75 1.80 1.78 i 8 .067 1.67 1.49 1.58 1 9 .060 1.64 1.83 1.74 i 8 .045 1.22 1.28 1.25 :10 .045 1.25 1.21 1.23 1 5 .027 0.675 0.651 0.663 1 8 94 FIGURE II-7 kobs as a Function of Hydroxide Concentration. -103 95 euHH mucosa Cot—\mfloev NO. x PIC”— t. m m e m m _ o _ _ J I _ 1 _ AVVAV annz N_nV QInv IONO. p-561 mno 96 neglect. The dielectric constant, D, was measured for each solution to assess the influence of the solvent system on the reaction rate. The equation that relates the dielectric constant of the solution to the value obtained with the dipolemeter is D = 0.5636 M + 3.757 where D is the dielectric constant and M is the number read from the dipolemeter. This equation is obtained by a least-squares analysis of four standard solutions: ethanol (D = 24.3), THF (D = 7.23), methylene chloride (D = 9.08), and 1-propanol (D = 20.1). The correlation coefficient for the straight line fit is 0.9990. The dielectric constant for each solution used in the kinetics experi- ments is then obtained by using the above equation. All measurements are for solutions at 25°. The equation used to interpret the kinetics results was derived by Laidler and Eyring56 for the general case of ion—dipole, ion-ion or dipole-dipole reactions. An approximate form of this equation is zze2 1 1 l ink=1nk+—(_—1)[ -_-] 0 2kBT D rA r?! + ___3_ (2 - 1) [u3 + if; — 1%.] (II-1) 8kBT D :3" r3 r3 B (3 ¢ IIIIIIIIIIIIIIIIIIIIIIIII::::—___________________________________::7_7——-"W 97 where k is the observed rate constant, k0 is the rate con- stant in a medium of infinite dielectric constant, Z is the charge on the ion, e is the electronic charge, kB is Boltzmann's constant, T is the absolute temperature, and rA, rB, rC are the sizes of the ion or molecules. r? is the distance of closest approach of the reactants in the activated complex and the u are the dipole moments. The values of in k as a function of l/D are given in Table II-3 and these values are plotted in Figure II—8. The kobs’ which is plotted, is the average value of k390 and k340. The rate of reaction increases with decreasing di— electric constant. The slope of the line is 69.1 I 3.2 and the intercept is -1l.l as given by a least-square analysis. C. 1,3—pm Dependence of the Rate The reaction order with respect to 1,3—pn was determined at a hydroxide concentration of 0.092 M. The solvent system was composed of varying amounts of tetrahydrofuran and 10 ml of ethanol; the 1,3—pm concentration was varied from 0.290 M to 1.13 M. The observed rate constants were found to be a linear function of the square of the 1,3-pm concen— tration. The results are presented in Table II—4 and in Figure II—9. These data imply that the reaction is second order with respect to 1,3—pn. A least—square analysis of data 98 Table II-3 Solvent Dependence of the Rate 2 2 . —1 _ 1/D x 10 kobs x 10 (min ) 2n kobs 0(%) 8.87 .654 5.03 i9 9.50 1 19 4.43 :8 10.19 1.77 4 03 17 10.73 2.54 3 67 i7 99 FIGURE II-8 kObs as a Function ofthe Inverse of Dielectric Constant. 100 N. muHH magmas No. x a: o. q 101 Table II-4 Observed Rate Constants as a Function of [1,3-pn]2 2 1 _22 .- [1,3 pn] M kObs x 10 (min ) 0(%) 0.087 0.571 i9 0.348 1.78 i9 0.743 4.66 :5 1.29 8.00 :6 102 FIGURE II-9 kobs as a Function of the Square of the Concentration of 1,3-propanediamine. 103 .v; “N; muHH magmas «28:86.65 «cam: my. man mvnv ago a _ — — Na! ¢A¥ mAV m5! not ATEEomeox 104 produces a slope of 0.0615 I .002 M51 min_1 and an inter- cept of zero. The fact that the intercept is zero implies that there is a one term rate law in this case. For the least-square analysis, the rate constants at both 390 nm and 340 nm are input to the computer program, but only the average value of these two constants (Ekobs) is reported in the table. D. NiMMK Dependence of the Rate The rate of reaction with respect to NiMMK is first order as shown by Nafisi-MovagharSS. No extensive data concerning NiMMK dependence were gathered during this thesis work, but the observed data always fit the first order decay of nickel complex to better than 10%. In many experiments, the fit of the data was as good as 2—3%. A plot of n (AO — Am)/(A - Am) as a function of time is given in Figure II—10. In this equation A00 and A0 are experimentally determined. The A0 and A00 calculated with the computer agree with the observed values within 3%. E. Effect of Temperature on the Reaction Rate The procedure outlined in the experimental section was used to obtain the rate of reaction as a function of tem- perature. Unfortunately, good data could not be obtained at temperatures below 60°C because of extensive decomposi- tion of the NiMMK. This decomposition was evident because 105 FIGURE II-10 A0 - A... 2n -— as a Function of Time for NiMMK. A-A an 106 OHIHH musmflm 26:. mm so NF om ms on VN N. o 107 the initial absorbance of NiMMK as well below its theo- retical value and in some cases up to fifty percent decom— position had taken place before the reaction could be initiated. All the reactants were kept constant for the tempera- ture experiments. The concentration of hydroxide was 0.092 M and the concentration of amine was 0.89 M_in all cases. The results of these experiments, presented in Table II-5, can be interpreted in terms of the Arrhenius equation where A is the pre-exponential or "frequency" factor, Ea is the activation energy, R is the gas constant and T is the temperature in degrees Kelvin. A plot of in k versus l/T is shown in Figure II-ll, and the slope of the line, ~4285°K, is defined as ~Ea/R; the intercept, 9.73 is in A. In this specific case, k = Robs/[1,3-pn]2 (see Mechanism). The activation energy of 8.5 I .3 kcal/mole, and the inter- * cept, in A, can be related to AS , the entropy of activa- tion by the Eyring equation: * kBT -AH*/RT AS /R Te 9 k(sec-1) = if the units of k are min-1, 60 k T * * 1) B e AH /RT eAs /R k(m1n — h 108 Table II-5 Temperature Dependence of the Rate 1) 1 T(°K) 103/T(°K’1) kobs(min- k(min- ) -£nk <fl%) 341.5 2.928 .0466 .0595 2.820 3 5 339.0 2.950 .0425 .0543 2.914 t 8 336.5 2.972 .0391 .0499 2.997 1 4 334.1 2.993 .0353 .0451 3.10 1 5 109 FIGURE II-ll in k as a Function of the Inverse of the Temperature. HHuHH mucosa 110 me. . cit. 8N main. had wad mad .va mad NQN _m.N — _ fl _ . A _ A 19.01 J mo.m.u x5 ._ mom- .. was: 1 Raw- 111 where: R = 1.987 cal/mol-deg k8 = 1.38 x 10"16 erg/deg 27 h = 6.62 x 10- erg-sec * AH enthalpy of activation AS entropy of activation * For reactions in solution, AH = Ea — RT and it can be shown that AS* = R(2n A - 27.85 — in T) * By using the appropriate values of the parameters, AH338 * is found to be 7.8 i .3 kcal/mol and AS ==-48i2 eu. 338 The activation parameters are to be regarded with caution because of the narrow temperature range in which they apply. DISCUSSION The rate of formation of the non-cyclic complex, NiApSo, is directly proportional to the concentration of NiMMK, and the rate is also proportional to the square of the concentration of 1,3-propanediamine throughout the range of concentrations that was studied. Hydroxide ion catalyzes the reaction; the rate of reaction is proportional to the concentration of hydroxide ion when the concentration is less than .07 M, and it is independent of the concentration of hydroxide ion at concentrations greater than .07 M. In the absence of hydroxide ion, decomposition of NiMMK occurs rather than the formation of NiApSo. The reaction rate also depends on both the dielectric constant of the solvent and the temperature. The rate of reaction increases with decreasing dielectric constant, and this behavior is indicative of either the reaction of a neutral molecule with an ion or the reaction between dipolar molecules which form an activated complex that is less polar than the reactants (Equation II—l). The values of enthalpy and entropy of activation suggest an associative mechanism for this reaction. The small value of AH*, 7.8 kcal/mol, indicates that there is no heterolytic cleavage of the nickel-oxygen bond in the rate determinig step, ‘while the large, negative value of AS*, -48 eu, indicates that an associative process may be occurring. The analysis of the experimental facts not only implies a mechanism 112 113 for this reaction, but it also indicates experimental conditions that may improve the synthesis of this type of nickel complex. The experimental rate law is ~dLNiMMK]_ _k[NiMMK][1, 3-pg] 2[0111 dt 1 + k' [OH'J (II-2) and a mechanism which reproduces the experimental rate law is NiMMK + 2 am : NiMMK-2am k1,k_1 NiMMK-2am + OH' + NiApSo k2 where am is a shorthand notation for 1,3—propanediamine, and NiMMK'Zam is assumed to be a short-lived species. From the stoichiometry, d[NiApSo] = -d[NiMMK] dt dt and d[NiApSo] dt = k2[OH-][NiMMK°2am] By using the steady state approximation for NiMMK'Zam, one obtains 114 d[NiMMK°2am] . 2 dt = 0 = k1[N1MMK] [am] - k__1 [NiMMK°2am] - k2 [NiMMK-2am] [OH-] or kl[NiMMK] [am]2 [NiMMK-2am]= _ (II-3) k_1 + k2 [OH ] and so k2k1[NiMMK] [am]2 [on‘] -d[NiMMK] = _ k - " obs k_1 + k2 [OH ] dt [NiMMK] (II-4) The experimental rate constants can be related to the individual constants by choosing appropriate experimental conditions. At high hydroxide ion concentration, Equation II-4 reduces to (assuming k_l < k2 [OH-J) -d[NiMMK] = dt k1 [NiMMK] [am]2 = k [NiMMK] obs , is and the observed rate constant, kobs 2 kobs = kl [am] A graph of kobs as a function of [am]2 yields k1 as the slope (Figure 11-9); kl is found to be 0.061 t .002 £72 min-1. At low concentration of hydroxide ion, the rate law (Equa- tion II—4) becomes 115 . k k _ “‘3”;ng = _2__1_ [NiMMK] [am]2 [OH ] = k [NiMMK] k obs -1 thus k k k = _2 1 [am]2 [on ] (II-5) obs k 1 The slope of the line of kobs as a function of [OH’] yields kzkl [am]2/k_1. Since k and [am]2 are known to be 0.061 1 and 0.348, respectively, the ratio kz/k_1 can be determined to be 13.6 M. By inverting Equation II—4, one obtains k 1: -1 1+ 1 obs kzkl [am]2 [69’] k1 [am]2 k A graph of the increase of kobs against the inverse of [OH-] will have a slope equal to k_1/k2k1 [am]2. The slope of this function is 3.5 i 0.2; by using the apprOpriate values of k1 and [am]2, the ratio k2/k__1 is found to be 13.3 M. To obtain k2 or k_1, the equilibrium constant for adduct formation would have to be determined. The postulated adduct, NiMMK'Zam, has not been observed, and there is no spectral evidence for this species in any reaction that has been studied. By hypothesis, it is a reactive intermediate which reaches a steady-state concen- tration that may be too low to detect by visible spectros— copy. The adduct may be detected by using electron spin resonance spectroscopy. If the amines coordinate in the 116 axial positions of the nickel complex and the interactions between the nickel ion and the amines are strong enough to decouple the paired electrons in the dxy orbital, the adduct will be paramagnetic and its esr spectrum might be observed at liquid helium temperature. The six coordinate adduct forms the products by reaction of the carbonyl carbon of NiMMK with the uncoordinated end of the diamine. The intermediate in this reaction is probably a carbinolamine, a well known intermediate in Schiff Base condensations, which is formed by nucleophilic attack by the uncoordinated nitrogen atom of the diamine on the carbonyl carbon atom (Figure II-12). The hydroxide ion may enhance the nucleophilicity of the amine by ab- stracting a proton from the uncoordinated nitrogen atom. Another possibility for the structure of the adduct is one where the amines approach the NiMMK in the equatorial plane. In this way there is the possibility for attack at both carbonyl groups, forming products in which two amine molecules have reacted with both carbonyl groups. Products expected from this reaction are not detected in any reactions between NiMMK and various diamines and so there is no evidence for this mode of attack. This type of reaction may hinder macrocycle formation because there is a free amine blocking the reaction site from the coordinated amine which reacted once and must now react again. The reaction NiMMK with ethylenediamine or 1,2-propane— diamine to form macrocyclic products proceeds in two steps55 117 .uouppd may mo soflumucmmmumwm oaumEmsow .NHIHH madmam IIouI zmI V2\// 2 I o\\Vz z /z o\ z I /2 ////;H\\\. /*D\ A ,///;r\\\\ l4RW. O\. /Z\z .81 O\. /Z\Z {IL /I III\ NIz 118 The first step leads to an intermediate similar to NiApSo, the product of the reaction between NiMMK and 1,3-pn; the second forms the macrocyclic product from this intermediate. The intermediate is formed in appreciable concentrations and can be isolated from the reaction mixture and this fact implies that the first and second steps in these reactions are of comparable ratesss. The data obtained by studying the reaction of NiMMK with 1,3-pn may give a good repre- sentation of the first sequence in the formation of macro- cyclic complexes containing nickel(II). FUTURE WORK A number of interesting points need to be clarified so we may better understand this type of reaction, and the most important point is to determine the location of nucleophilic attack by the amine. Two carbonyl carbons are present in NiMMK, and yet the amine attacks only one of them because only one isomer is formed. A number of techniques (high resolution mass spectrometry and Extended Hfickel calculations) have failed to distinguish between these two carbon atoms. X—ray diffraction studies would give the structure of NiApSo, but suitable crystals have not been grown and this technique has not been implemented. The function of hydroxide ion poses another interesting problem. There is no definite evidence that the role of the hydroxide ion is to abstract a proton from the amine. If a number of different bases can be used to catalyze the reaction, then a correlation of rates as a function of basicity may lead to some insight regarding the role of the base in these reactions. Variation of the amine and the metal ion can also provide useful information about the chemistry of these reactions. The ring closure of NiMMK seems to be steri— cally controlled since macrocyclic complexes have only been formed when the amine possesses a chain containing two carbon atoms, 219;! ethylenediamine. All other diamines that were used in these condensation reactions lead to 119 120 noncyclic products similar to NiApSo (See Appendix). It may be possible to form macrocyclic complexes with the longer chain amines if the metal ion in NiMMK could be varied. The metal ion plays a crucial role in the formation of many macrocyclic complexes and by varying the metal ion, the rate of reaction could be observed as a function of the electronic configuration of the metal ion. This data may lead to a better understanding of the nature of the template effects which are so important in the formation of macrocyclic complexes of transition metal ions. APPENDIX A. STO-3G Basis Set The basis set employed in the calculations on carbonyl- nitrenes is the STO-3G basis of Pople16 , et al; These basis functions are linear combinations of nuclear centered Gaussian functions which are least-squares fit to unscaled Slater-type functions (STO's). The scale factors used in this work are those designated by Pople as the "standard molecular" scale factors. The form of the function, Xi’ which mimics the STO is 2 2 = 4 m n ‘C aikr xi(r) Ni X Y Z dik e "hdw k 1 where Ni is a normalization constant, aik are the least- squares determined exponents, and dik are the coefficients of each primitive Gaussian in the series. Ci is the scale factor appropriate for each function and atom, and r is the radial distance of the electron from the nucleus. Angular anisotropy is induced into the function by the pro- gymz“. If £=m=n=o the duct of cartesian coordinates X function is an s-type. If £=m=o and n=l, then the function is of pz symmetry. The functions used in these calculations are all either 3 or p types. There are larger STO expan- sions available some using as many as six Gaussian primi- tives to mimic each Slater function. These expansions were not used due to the increased amount of time required 121 122 to compute the integrals in the primitive basis. A minimal basis was used in all of these calculations which is to say that the number of basis functions was equal to the number of atomic orbitals with distinct n and l quantum numbers. For example, the minimum basis set for a carbon atom in- cludes ls, 25, 2px, Zpy, and 2pz functions. The STO-3G basis functions used in this series of calculations are listed in Table A1. B. 3sZp/23 Basis Set A larger basis calculation was used to check on any inadequacies of the minimal basis study and to improve the computed results. A number of larger bases are available but the Dunning18 [3sZp/25] contraction of Huzinaga's (9sSp/4s) basis seemed most suitable to our needs. Basically, a contracted basis is one where the ratios of coefficients of the primitive functions are fixed which reduces the number of integrals to be evaluated24. For example, if the full (9sSp/4s) basis of Huzinaga was used in the cal— culation on HCNO there would be (9+15)x3+4 = 76 functions or 8778 one electron integrals and 3.853 x 107 two electron integrals to evaluate. By using the contracted [332p/Zs] basis, these numbers are reduced to 1305 and 94,830 re- spectively. The efficacy of such a procedure is clearly apparent when one considers the time for an SCF iteration 4 is proportional to n where n is the number of basis 123 Table A1 Carbonylnitrene Basis Set: STO-3G Coefficients and Exponents 6.464805 -.0999672 .155916 Exponent(§2a) d(s) d(p) Basis Function 3.53052790 6.00313672 Carbon ls 13.0450913 7.22760730 71.61681857 2.08363733 .22229003 1.5—T929061 .88416047 JCarbon 2—T—s, p .68348211 .90120499 1.37078856 2.94125016 -.22§§0190 .35170890 74T88568202 7.65935849 Nitrogen ls 18.05230544 9.22165160 99.10614297 2.65849772 .28571453 1.90643191 1.06730942 Nitrogen 2s,2p .87849538 1.08788461 1.65473982 3.78045691 -.27221337 .42456345 6.44363704 9.42642031' Oxygen ls 23.80885689 11.3491442 130.70928710 3.27182975 ,11_ 4f_ .38038916 2.36288812 1.32285487 Oxygen 23,2p 1.16959444 1.34835673 2.0509335 5.03315269 -.33738930 .52621650 .16885616 .61395448 Hydrogen ls .62391349 .73918388 3.42525002 .21309834 78.216857 .444635 Fluorine ls 30.36080 .535328 170.34090 .154329 ‘_‘ .48858877 .700115 .391957 Fluofine2s,2p 1.502279 .399513 .607684 124 functions. The notation (9sSp/4s) implies 9 s-type functions and 15 p-type functions for each first row atom and 4 3- type functions on hydrogen. The [] notation means that the basis is a contracted set. The explicit functions for HCNO are listed in Table A2. This basis provides much more flexibility in the valence regions than does the STO—3G basis. A scale factor of (1.2)2 was used to multiply the exponent of the hydrogen functions to get the exponents listed in the table. C. The CI The Hartree-Fock function is only the first approxima- tion to the solution of the electronic Schroedinger Equa- tion. In many cases it can be improved by configuration interaction (CI). In the CI formalism the true eigen- function is approximated by a linear combination of Slater Determinants which are built from the SCF molecular or- bitals. If the combination of determinants was infinite, then the true wavefunction would be obtained. In practice, however, the series is truncated after a finite number of terms and only an approximate eigenfunction is constructed. In HCNO, a CI expansion of 100 determinants is used. These determinants were constructed from all excitations of * co’ pN’ ”0’ ON’ "c0' There are six electrons (3a, 38) in these orbitals in the 5 2 3) the five valence-like orbitals: n SCF solution, so there are ( = 100 determinants to be 125 Table A2 3sZp/Zs Basis Set For Formylnitrene Exponent Coefficient Carbon 3 4232.6100 0.002029 634.8820 0.015535 146.0970 0.075111 42.4974 0.257121 14.1892 0.596555 1.9666 0.242517 5.1477 1.000000 0.4962 0.542048 0.1533 0.517121 Carbon p 18.1557 0.018531 3.9864 0.115442 1.1429 0.386206 0.3594 0.610080 0.1146 1.000000 Nitrogen s 5909.4400 0.002004 887.4510 0.015310 204.7490 0.074293 59.8376 0.253364 19.9981 0.600576 2.6860 0.245111 7.1927 12000000 0.7000 0755237371" 0.2133 0.508031 Nitrogen p 26.7860 0.018257 5.9564 0.116407 1.7074 0.390111 0.5314 0.637221 0.1654 1.000000 Oxygen s 7816.5400 0.002031 1175.8200 0.015436 273.1880 0.073771 81.1696 0.247606 27.1836 0.611832 3.4136 0.211205 9.5322 1.000000 0.9398 0.563459 0.2846 0.497338 126 Table A2 - Continued Exponent Coefficient Oxygen p 35.1832 0.019580 7.9040 0.124189 2.3051 0.394727 0.7171 0.627375 0.2137 1.000000 Hydrogen 5 19.2406 0.032828 2.8992 0.231208 0.6534 0.817238 0.1776 1.000000 127 constructed. These determinants are the MS=0 components and they are presented in Table A3. In the table, a determinant is represented by five columns; each column denotes a spatial molecular orbital. The occupancy of these orbitals is represented by a series of ones and asterisks. The numeral one signifies an electron, and an asterisk denotes no electrons. The core occupancy, which is always 16 electrons, is not displayed in this table. As an example, consider the 3A" state of formyl— nitrene. The core of eight doubly occupied orbitals can be represented as 11/11/11/11/11/11/11/11/, where the solidus (/) separates each orbital. The first 1 in an orbital signifies an electron with spin a; the second 1 in an orbital denotes an electron with spin 8. The elec- trons outside the core for the 3A" state are represented as ll/ll/l*/*l/ and 11/1l/*l/1*/. These representations correspond to the two determinants which are the major 3 components of the A" state (the core electrons are omitted in this example): 1 1])(3An) z/IHOOO‘OOBWCOO‘NCOBPNO‘UNBI + IOOaOOBnCOaflCOBpNBONaI} In the calculations on larger carbonylnitrenes, a 128 Table A3 Determinental Basis for Formylnitrene . * Coefficient 00 "C0 pN ON "C0 Index I...” G 111111111“ § I.” 11111:... I119 111]” § § 1* l IG 5 G I.” ‘6 11% 1111...!“ .IO 11* 11.11111 #61», 9 § 15116 1.1119 19 11G llfi § § § 110 .G llllfl Isl-11¢ «1511611481 4 1111* Illtl§ 115 l Ii 9 1].“ 11¢ 11} 1119 11111.19 G U fill} #1111110 1% ll. lfi G .119 11¢ III-#1110 111.16 I... G ”19 #1119 111*19111 lllfiGilllfilll§§lll§1§l§lll§fill.” *fifi‘fiflfifl] ll1§ G fl .1111“ 111% llfi III-1'16 19116 G G fi 1% 0 § 1% G 1% III-.1331!” § § § 111111..” 1&0131110 § .l‘lfi IIG 1G .l§ IIG fl lull-1119 D * § 111111115 lfi 11.11% 111.11% 19 1! 11¢ 6 Q 0000000000000000000000 00000080000000000 oooooooooooooooooooooooooooooooooooooooo lllllllllllllllllllllllllllllll‘lllllllll I? 31.256780. 0173456789011? 3N5“ ? 129 Table A3 - Continued * Coefficient Trco 0N 1Tco PN 00 Index 15.... ##1§1§1§11§1111§§§ .91... .955. 1:.lfillllflllfilfifi § 1.111% 111d 111* fl 9 6.9 § 1% #111115 .111” §1§ 19 *111 16 111111§111§1 1G 11% 111.11% 11§ fl #1“ 11% 11 1§1§ §1§§ §1§1§11111§ 1G§1§ §§1§1§1§1§§§1111 filfifill§11§1§1§11111§§§§ §§§ .l §1§1§1§ *1111 #1111G111fi 1111§1111111§1§111§1§ 111.11% #1 § 1§ 1.11111 5.111111% 1111111§ 1% 111§ 1% 111.1% #§ 1 11% § 111% §1§ #1111G1fi 111111§ 1§111§1§ 111§ .1 1111§ Q #11§ § § 11111§ 1§ 11111116 15 111% 1% 111% 0000000000000000000000000000000000000000 on.000.000000000000000.0000...0000000000 111111111111111111111111111111111111.1111 789017345673901734§67 0.0 555666666666677777777778 ‘56 130 Table A3 - Continued 1Tco 0N 0o 1Tco pN Coefficient Index 1Q 111111§ 111.1% .1” 1111 § 11111111§ 11* 1” 11111 9 1111” 1111” 11 .111“ § § 1 1§ G § 111.1110 11% 11.1 11“ fl .1” 11.11116 § § 1111§ 1.1” 1% 1? 111.1% .15 §x11111§ G .1 fi 1% 11# 111§ 1 1% 1.111% 1.11% § 111§ 1§ 1* .I. 1% 111% 19 1” §1§§§§§§§§ lfifififififlfififi 00000000000000000000 oooooooooooooooooooo 11111111111111111111 131 structure basis was used. Structures are Spin and space symmetry adapted linear combinations of determinants. It is easier to diagonalize the matrix representation of the Hamiltonian in a structure basis because only states of specified spin and space symmetry are included and the matrix is smaller. D. Syntheses A number of condensation reactions were attempted between various amines and NiRR'K where the R groups are attached to the carbon that is derived from the ketone (see Figure II-l for NiMMK; R=R'=CH3). The list combina— tions of R and R' can be seen in Table A5 which lists some results. These condensation reactions were undertaken to assess the ability of the R groups to enhance the reac- tivity of the carbonyl moieties. (See Table A4 for abbre— viations.) All nickel complexes were prepared as previously des- cribed. The amines (Aldrich) were used as supplied except for ethylenedeamine (en), and 1,3—propanediamine (1,3—pn) which were doubly distilled as described in the experimental section of this Thesis. Ethanol and THF were purified as described in the experimental section. The general technique used was to dissolve approxi- mately 5 x 10.4 moles of NiRR'K in solvent (either ethanol or THF), add 10-3 moles of a sodium hydroxide in solution 132 Table A4 Abbreviations l) Nickel complexes - NiRR'K M E Pr Bu Ph 2) Amines en 1,3-pn 1,4 DAB 1,5 DAP 1,6 am 1,7 am 1,8 am 3,3' am unsym CH3 CH3CH2 CHBCHZCHZ CHBCHZCHZCHZ C6H5 ethylenediamine 1,3 propanediamine 1,4 butanediamine 1,5 propanediamine 1,6 hexanediamine 1,7 heptanediamine 1,8 octanediamine 3,3' Iminobispropylenediamine unsym—dimethylenediamine 133 Table A5 Results of Synthetic Work Compounda Est. Time MW Yield (%) NiMEK en* 1 hr 583.38 82.6 NiEEK en* 50 min 597.41 81.4 NiMPr en* 45 min 597.41 85.0 NiMBuK en* 1 hr 611.43 86.0 NiMPhK en* 30 min 631.42 64.4 NiMEK l,3-pn 2-20 hr 615.42 84.3 NiEEKl,3-pn 2 hr 629.45 57.5 NiMPrK 1,3DAP 1-2 hr 629.45 81.1 NiMBuK 1,3DAP 2-3% hr 643.48 78.9 NiMPhK 1,3DAP 0-1% hr 663.47 69.2 NiMEK unsym 4-22 hr 627.43 81.2 NiEEK unsym 4-24 hr 641.46 51.9 NiMPrK unsym 4-22 hr 641.46 70.6 NiMBuK unsym 4k-23 hr 655.49 86.8 NiMPhK unsym 40 min 675.48 80.9 NiMEK 1,4DAB 35-2 hr 628.44 83.1 NiMEK 1,5DAP 3-6 hr 642.47 85.0 * Macrocycle. aThe first entry is the nickel complex and the second is the amine. 134 and heat the system to the boiling point. 5 x 10"2 moles of amine were added and the course of reaction was monitored spectrophotometrically. When the reaction was completed the product was precipitated by dispensing the reaction mixture into a quantity of cold water. The precipitate was filtered and dried in a drying pistol at 78°C. Some results are presented in Table A4. Since the macrocycle is formed by the condensation of both ends of the amine with NiRR'K, the molecular weight (MW) of the macrocycle equals the MW of starting material (NiRR'K) plus the MW of amine minus twice the MW of water. The MW of the dangling species is the MW of NiRR'K plus the MW of the amine minus the MW of water. All the above species are dangling species (only one carbonyl group of NiRR'K has reacted). Some synthetic work with NiMMK and various amines was also completed. The technique was the same as previously described, but in this case two different solvents were used. The results (Table A6) showed that it may be pos- sible to form macrocycles in a solvent with a higher boil- ing point than THF. Some macrocycles were sent for elemental analysis to Chemalytics, Tempe, Arizona. The results are: C N H theory found theory found theory found NiEEK—en 70.37 70.28 14.41 14.47 5.74 5.83 NiMPhK—en 72.28 70.23 13.31 12.95 5.11 5.20 135 Table A6 Solvent Effects on Macrocycle Formation NiMMK + THF Ethanol 3,3' am dangling mcycle m/e 657 m/e 639 1,7 am dangling mcycle m/e 656 m/e 638 1,6 am dangling dangling m/e 642 m/e 642 1,8 am no rx no rx 136 The trend implies that macrocycles derived from NiRR'K are most readily formed with amines that have a two carbon bite (343;, en); under suitable conditions, such as sol- vents with higher boiling points, the macrocycles may also be formed in good yield. A solvent with good potential is dioxane. Since the reaction that occurs is a nucleophilic attack on the carbonyl carbon, the rate of reaction should in- crease with the basicity of the amine. This trend may be true; compare the estimated completion times of NiMEK with 1,3 pn and 1,4 DAB (Table A5). l,4 DAB is a slightly stronger base than 1,3 pn and the estimated completion time for 1,4 DAB is faster than for 1,3 pn. The compari- son of‘basicity, an equilibrium property, with nucleo- philicity, a kinetics property, must be made with care and more complete data needs to be completed before any generalizations can be made. REFERENCES 3. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. W. Lwowski, Ed., "Nitrenes", Interscience, New York, New York, 1970. E. Wasserman, G. Smolinsky, W. Yager, 8. Amer. Chem. Soc., 88, 3166 (1964). E. Wasserman, R. Murray, W. Yager, A. Trozzolo, G. Smolinsky, 8. Amer. Chem. Soc., 88, 5076 (1967). G. Smolinsky, E. Wasserman, W. Yager, 8. Amer. Chem. Soc., 88, 3220 (1962). W. Lwowski and J. McConaghy, 8. Amer. Chem. Soc., 88, 2357 (1967). D. Cornell, R. S. Berry, W. Lwowski, 8. Amer. Chem. Soc., 88, 3626 (1965). R. S. Berry, D. Cornell, W. Lwowski, 8. Amer. Chem. Soc., 88, 1199 (1963). W. Lwowski and T. Mattingly, Jr., Tetrahedron Letters, 277 (1962). W. Lwowski and J. S. McConaghy, 8. Amer. Chem. Soc., 88, 4450 (1967). W. Lwowski and J. S. McConaghy, 8. Amer. Chem. Soc., 88, 5490 (1965). P. S. Skell and R. C. Woodworth, 8. Amer. Chem. Soc., .18, 4496 (1956). W. Lwowski, Ed., "Nitrenes", Interscience, New York, New York (1970), pp. 43. F. L. Pilar, "Elementary Quantum Chemistry", McGraw- Hill, New York, New York, 1968. C. C. J. Roothaan, Rev. Mod. Phys., 88, 69 (1951). H. F. Schaefer III, "The Electronic Structure of Atoms and Molecules", Addison-wesley, Reading, PA, 1972. W. J. Hehre, R. F. Stewart, J. A. Pople, 8. Chem. Phys., 88, 2657 (1969). R. Hoffmann, G. Zeiss and G. VanDine, J. Amer. Chem. Soc., 88, 1483 (1968). — T. H. Dunning, Jr., 8. Chem. Phy . 88, 2823 (1970). S. Huzinaga, 8. Chem. Phy . 88, 1293 (1965). 137 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 138 C. C. J. Roothaan, Rev. Mod. Phys., 88, 179 (1960). D. B. Neumann, H. Basch, R. L. Kornegoy, L. C. Snyder, J. W. Moskowitz, C. Hornback and S. P. Liebmann, Poly- atom (Version 2), Program 199, QCPE, Indiana University. C. F. Bender and E. R. Davidson, 8. Phys. Chem., 18, 2675 (1968). H. F. Schaefer, R. A. Klemm and F. E. Harris, Phys. Rev., 181, 137 (1969). An excellent discussion of contracted basis sets is given in reference 15 on page 70. R. Liedtke, Ph.D. Thesis, Michigan State University, 1975. D. H. Busch, Helv. Chim. Acta. Fasc. Extraord. Alfred Weiner, 1974 (I967). N. F. Curtis, Coord. Chem. Revs., 8, 3 (1968). D. H. Busch, Adv. Chem. Ser. 188, 44 (1971). D. St. C. Black and A. J. Hartshorn, Coord. Chem. Revs., 2, 219 (1972). L. F. Lindoy and D. H. Busch, "Preparative Inorganic Reactions", W. Jolly, Ed., Vol. 6, Interscience, New York (1971). PP 1-61. J. Christensen, D. Eatough, R. Izatt, Chem. Rev., Z4, 351 (1974). L. F. Lindoy, Quart. Rev., 3;, 379 (1971). C. L. Honeybourne, Inorg. Nucl. Chem. Lett., 11, 191 (1975). C. L. Honeybourne and P. Burchill, Inorg. Nucl. Chem. Lett., 18, 715 (1974). J. P. Collmann, R. Gagne, C. Reed, W. Robinson and G. Rodney, Proc. Nat. Acad. Sci. USA, 21, 1326 (1974). J. P. Collmann, R. Gagne, J. Keuba, H. Ljunberg-Wehren, 8. Amer. Chem. Soc., 28, 6801 (1974). R. G. Wilkins, "Uptake of Oxygen by Cobalt(II) Complexes in Solution", Adv. Chem. Ser., 198, 111 (1971). 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 139 J. Vasilevskis and D. Olson, Inorg. Chem., 88, 1228 (1971). F. Lovecchio, E. Gore, D. Busch, 8. Amer. Chem. Soc., 28, 3109 (1974). J. Vasileoskis and D. Olson, Inorg. Chem., 10, 463 (1971). ”“ B. Wayland and L. Olson, 8. Amer. Chem. Soc., 38, 6037 (1974). P. Lau and W. Lin, 8. Inorg. Nucl. Chem., 81, 2389 (1975). T. Truex and R. H. Holm, 8. Amer. Chem. Soc., 28, 285 (1971). R. Stotz and C. Stoufer, Chem. Comm., 1682 (1970). G. A. Melson and D. H. Busch, 8. Amer. Chem. Soc., 87, 1706 (1965). G. A. Melson and D. H. Busch, 8. Amer. Chem. Soc., 88, 4834 (1964). C. M. Kerwin and G. A. Melson, Inorg. Chem., 11, 726 (1972). C. M. Kerwin and G. A. Melson, Inorg. Chem., 83, 994 (1974). E. L. Blinn and D. H. Busch, 8. Amer. Chem. Soc., 28, 4280 (1968). E. L. Blinn and D. H. Busch, Inorg. Chem., 1, 820 (1968). Ibid., pp. 821. J. A. Burke and E. C. Brink, Inorg. Chem., 8, 386 (1969). J. A. Burke and S. Campbell, 8. Inorg. Nucl. Chem., 83, 1163 (1971). “ R. McQuate and D. Leussing, 8. Amer. Chem. Soc., 21, 5117 (1975). K. Nafisi-Movaghar and G. A. Melson, Inorg. Chem., 84, 2013 (1975). K. Laidler and H. Eyring, Ann. N. Y. Acad. Sci., 88, 303 (1940).