.5 : :. 2:...1. I. . ' TAI ES'coN CKEL (II) co; N NON LNG NONE; ¥ . .NJ . MACROCYCLIC gLI’GANDS' ‘MONOHYD MP BENZIL AND evcuc DERIVED FROM Thesis for the “Dame Of 'hPh . D. MICHIGAN ' COURTNEY le~ :ift.......::...,._..:_......_.T.,.b: 5:31! an”) .533 In 5:. Wars. 57.. .p an» :2, . r. C! p. 3.53 Jammy {1.1.}... r. 1 ”mania . fl mtg—‘1 s .: a... l 2. $135.»: .r.::.€_ .’,..r,.:.:w\.l. «.17; ,7, 1:4 £514.51: V (3.1 This is to certify that the thesis entitled Nickel (II) Complexes Containing Non—Cyclic and Macrocyclic Ligands Derived from Benzil Monohydrazone presented by Courtney Michael Kerwin has been accepted towards fulfillment of the requirements for Ph.D. Chemistry degree in :4 y / . Date—{ii 0-7839 LIBRARY Michigan Stabs University ABSTRACT NICKEL (II) COMPLEXES CONTAINING NON-CYCLIC AND MACROCYCLIC LIGANDS DERIVED FROM BENZIL MONOHYDRAZONE BY Courtney Michael Kerwin As a result of this project, a new type of non- cyclic Ni cis N202 complex, a new l3-membered macrocyclic system, and an unusual non-cyclic Ni N30 type system have been characterized. The non—cyclic Ni cis N202 complexes, nickel ketazines, result from the reaction of some ketones, RlRZCO, and benzil monohydrazone in the presence of nickel (II) ions. The resulting dinegatively charged tetradentate ligands are coordinated in a square planar arrangement about the nickel (II) ion. A characteristic of this type of condensation reaction is that the carbon atom from the carbonyl group of the ketone forms the sole bridge between terminal nitrogen atoms of two coordinated benzil monohydrazone residues. The Ni N4 macrocycles and the Ni N30 type compounds result from the direct reaction of the coordinated oxygen atoms of the nickel ketazines with di- or mono—amine Courtney Michael Kerwin compounds. The reactivity of coordinated carbonyl groups has been demonstrated previously, but an apparent need has been noted for B (meso) carbonyl substituents in coordinated Ni cis N202 type compounds in order for cyclization reactions involving aliphatic diamines to occur. In the nickel ketazine system, the coordinated -CO group reacts with aliphatic amines at room temperature or higher to yield cyclic or noncyclic products depending on the reacting amine. For 1,2 diaminoethane and 1,2 diamino— propane, both coordinated oxygens are replaced and Ni N4 tetradentate macrocyclic compounds are obtained. For ethylamine, at 100° in a pressure tube, and for 1,3 diaminopropane, at from room temperature to 100°, only one coordinated oxygen is replaced, resulting in the formation of unusual Ni N30 type compounds; in the case of 1,3 diaminopropane the ligand is potentially pentadentate. A mechanism is proposed for the reactivity of these coordinated -CO groups. The initial step in this mechanism is suggested to be the coordination of the incoming amine to the central metal ion. NICKEL (II) COMPLEXES CONTAINING NON-CYCLIC AND MACROCYCLIC LIGANDS DERIVED FROM BENZIL MONOHYDRAZONE BY Courtney Michael Kerwin A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1972 TO LILI ii ACKNOWLEDGEMENTS The author wishes to express sincere appreciation to Dr. Gordon A. Melson for suggesting this area of research, for his guidance, his personal interest, and his encouragement throughout the course of this project. The author wishes to express his gratitude to the Department of Chemistry, Michigan State University for the financial aid, in the form of a Graduate Teaching assistant- ship, which made this research project possible. The author also wishes to thank the Dow Chemical Company for a Summer Fellowship and for their generous assistance in obtaining high resolution mass spectra. TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . LIST OF FIGURES. . . . . . . . . . . . I. INTRODUCTION. . . . . . . . . II. NOMENCLATURE. . . . . . . . . . . III. EXPERIMENTAL. . . . . . . . . . . Synthesis of l, 2, 8, 9- -Tetraphenyl- 3, 4, 6, 7- Tetrazo— 5, 5— —Dimethy1nona- l, 3, 6, 8— —Tetraen -l, 9- dioxynickel (II) . . . . . Other Ketazine Complexes . . . . . Nickel Hydrazone Complex . . . . Reaction of Benzil Monohydrazone with Acetone in the Absence of Nickel (II) Ions . . . . . . . . . . . Other Variations Investigated . . . . Other Bridging Species . . . . . . Other Metal Ions . . . . . . . Reactions of NiMMK with Amines . . . [3, 3-Dimethy1—6, 7,12,13-Tetraphenyl- l, 2, 4, 5, 8, ll— Hexazacyclotrideca- l, 4, 6, 12- Tetraenato(2- )-N}N5N$N11] Nickel. . . . . . . . . . [d'[[l-[[2-[[3 Aminopropyl)amino]-l, 2- Diphenylvinyl]azo]— —l- -Methylethyl]azo] -a- Stilbenolato(2- )] Nickel. . . iv Page vi 11 ll l4 l4 l6 l6 16 20 21 21 23 Reaction of NiMMK with Ethylamine . Dehydrogenated Compound. . Reactions of NiMMK With Other Amines Other Reactions Investigated. . Reactions of Benzil Monohydrazone With Amines. . . . . . . . . IV. PHYSICAL MEASUREMENTS. . . . . . V. RESULTS AND DISCUSSION . . . . . Nickel Ketazines. . . . . . Reaction Products of NiMMK . . . . Macrocyclic NiN4 Type Compounds . Non—Cyclic NiN3O Type Compounds . Reactivity of Coordinated CO Groups Mechanism of the Reaction of NiMMK With Amines. . . . . . . . . Suggestions for Future Work . . . VI. CONCLUSIONS . . . . . . . . . . BIBLIOGRAPHY. . . . . . . . . . . Page 24 26 27 28 3O 33 34 34 51 52 62 69 72 79 82 83 LIST OF TABLES Table Page 1. Analytical Data for the Complexes . . . . . 22 2. Common Absorptions Bands in the Infrared Spectra of Nickel Ketazines . . . . . . 36 3. Proton Magnetic Resonance Spectra of Nickel Ketazines . . . . . . . . . . . . 37 4. Mass Spectra of Nickel Ketazines . . . . . 41 5. Results of High Resolution Mass Spectroscopy Investigation for NiMMK, (R1 = R2 = Me) . . 44 6. Ultraviolet and Visible Spectra of Nickel Ketazines . . . . . . . . . 46 7. Common Absorptions Bands in the Infrared Spectra of NiN4 and N1N3O Compounds. . . . 53 8. Results of High Resolution Mass Spectroscopy Investigation for Ni M cyclo l3 . . . . 55 9. Mass Spectra of NiN4 and NiN3O Compounds. . . 56 10. Ultraviolet and Visible Spectra of NiN4 and NiN3O Compounds . . . . . . 57 vi LIST OF FIGURES Figure l. 100 MHz lH nmr Spectrum of Nickel Methylethyl Ketazine, (NiMEK) . . . . . . . . . 2. Electronic Absorption Spectra, in Benzene, of NiMMK, Ni H cyclo l3, and Ni M cyclo 13. 3. Electronic Absorption Spectra, in Benzene, of NiApSo and NiESo . . . . . . . . 4. Circular Dichroism Spectra, in Benzene, of R-Ni M cyclo l3. . . . . . . . . vii Page 39 58 58 61 I . INTRODUCTION Chemists as early as Werner have recognized the altered properties of ligands coordinated to metal ions, however until just recently the inorganic chemists who have been studying coordination compounds were mainly interested in the effect of the ligands on the metal atoms. Recent advances in the field of biochemistry though have shown the importance of coordinated metal ions in reactions of biological significance. Also the discovery of metal ion catalyzed syntheses of polymers has emphasized the roles of coordination compounds as reaction intermediates. The recent literature dealing with reactions of coordinated ligands is quite vast.l The effects of the metal ion are readily studied in systems where the ligand remains coordinated to the metal ion during and after a chemical reaction. The metal ion may serve to catalyze unusual reactions, impart special steric effects, stabilize otherwise unstable molecules, or have several other effects on the reactivity of the coordinated ligand. One of the most interesting aspects of this area is the ability of the metal ion to function as a template in holding reactive groups in proper positions for sterically selective, multistep reactions. Many researchers have made use of this "template effect" to synthesize compounds which are difficult or impossible to make using other techniques. One area in the field of reactions of coordinated ligands, which has received a great deal of attention, is the condensation of metal amine complexes with carbonyl compounds. Reference here will be made to two of the numerous review articlesz’ 3 whose contents describe this area of chemistry. This type of reaction has proved useful in the synthesis of both cyclic and non—cyclic compounds. These condensation reactions fall into two general classes. The first results in the formation of a Schiff base from the coordinated amine, Structure I, whereas the second type is characterized by the linking of two coordinated amine groups by a three carbon bridge, Structure II. Depending on the metal amine complex employed, macrocyclic compounds with 13—16 member ring systems can be synthesized. The 13 member ring systems, Structure III, have previously been obtained only from condensation reactions of metal triethylenetetramine complexes. Condensation reactions have resulted from reactions of coordinated aldehydes or ketones with amines.4 Schiff first used a reaction of this type to prepare bis (salicylaldimino) copper complexes in 1869 and many such C :ZJH/ H\/ II C/ HZC—N HZN—CHZ H / N H;:' \h+2 \CHZ C//’ H\\\ +2/// \\ClH \ / \ / HZC— NH2 N——— CH2 l' \“M/ \NH /Cl2 CH3/ \C H3CIC/ |\C/I\HCH3 Structure I H2 CH3 Structure II /CH CH .12Cm//\/Ww\ Ham/Kw /\§ /3\.. 3.. Structure III 2 2/ \/\O reactions involving coordinated salicyclaldehyde have been performed since then.5 This Schiff base condensation is facilitated by the presence of a second donor atom which assists in the formation of a strong chelate ring. Then, polarization of the carbonyl caused by coordination makes it more susceptible to nucleophilic attack by the lone pair of the amine. For a diamine there is a further "chelate effect" which facilitates the reaction of the other amine group in a similar manner. Condensation reactions of coordinated carbonyl groups do not usually lead to the formation of macrocyclic compounds. However, Jager6 has synthesized some macrocyclic complexes by reactions between tetradentate B-ketoamine complexes and primary diamines. Other authors7’ 8’ 9 have also investigated the reactions of coordinated carbonyl compounds with amine compounds. These cyclization reactions using aliphatic diamines have failed in the absence of a substituent B or meso to the coordinated carbonyl group, see Structure IV. This requirement does not hold for aromatic diamines,6' 7 but other factors, such as the steric rigidity of aromatic diamines, the differences in basicity between aromatic and aliphatic amines, or the higher temperatures used, may be involved. This dissertation is a report on some nickel (II) complexes derived from benzil monohydrazone and the products of the reactions of one of these compounds with R?\\= ‘///RY 0V0 \ 8 YR N///’ M\\\\N —_—_—_A7 R8 NBN\/ Structure IV some amine compounds. This study has resulted in the characterization of a new type of condensation reaction of a metal amine type complex with a carbonyl compound, in which the carbon atom from the carbonyl group of a ketone forms the sole bridge between terminal nitrogen atoms of two coordinated benzil monohydrazone residues. The dinegatively charged tetradentate ligands resulting from these condensation reactions are shown in Structure V (p. 9). A compound of this type was first proposed by Taylor, Callow, and Francis10 in 1939 when they isolated a small quantity of red needles from a synthesis of the nickel complex of benzil monohydrazone in boiling acetone. These authors suggested that a condensation reaction had occurred, but they were unable to ascertain the structure of the new complex. This earlier study was the starting point for this project which has now resulted in the characterization of a series of these nickel ketazine compounds. The second part of this project was undertaken to investigate the reactivity of the coordinated carbonyl groups of one of these nickel ketazines, NiMMK, towards mono- and di-amine compounds. In view of other workers' investigations of the reactivity of coordinated carbonyl groups (vide supra), it seemed appropriate to investigate the reactivity of the nickel ketazines. It was hoped that if these coordinated carbonyl groups were to react with amine compounds, that data might be obtained which would contribute to a better understanding of the conditions necessary for these reactions to occur and of the mecha- nism involved in this type of reaction. This has resulted in the characterization of two macrocyclic Ni N4 type complexes, a noncyclic Ni N30 type complex, and a potenti- ally pentadentate, non-cyclic NiN4O type complex. These compounds are shown in Structures VI and VII (pp. 9—10). II. NOMENCLATURE The first non-cyclic nickel complex synthesized in the course of this project is shown in Structure V, along with its IUPAC name. The second name listed under the structure is the name under which this compound was previously published. To remain consistent with previous publications these complexes will be referred to as nickel ketazines in the abbreviations used here and are further related to the ketone used in the condensation reactions by which these compounds are synthesized. For example, Ni MMK corresponds to the nickel ketazine formed by use of methyl methyl ketone (acetone). The macrocyclic compounds derived from Ni MMK are illustrated in Structure VI. The fundamental ring system involved in these compounds is named, numbered, and oriented as shown below: H 2 N = N _ C _ N = N 1, 2, 4, 5, ll— —Hexazacyclotrideca- l 2 3 4 5 :Q\( 1,4, 6,12——tetraene 7 CH 13 12 11 10 9 8/ HC = C - N — C - C N H H2 H2 H In abbreviating these names the key words "cyclo 13" will refer to the generalized compound in Structure VI and an abbreviation for the substituent on carbon #9 will serve as a prefix, i.e., Ni M cyclo 13 will indicate the presence of a methyl substituent at position #9 in the macrocycle. Structure VII shows another type of compound which has been synthesized whose IUPAC names are related to that of the parent compound, Ni MMK, by the "stilbenolato (abbreviated So)" stem. These compounds differ only in the amino group bonded to the #2 carbon, so an abbreviation for these substituents, plus "So," for the stilbenolato stem, can represent these compounds; i.e., Ni-Ap So corresponds to Structure VII with R = 3-aminopropyl. The compound shown in Structure VIII is a complex of benzil monohydrazone with nickel (II) and will be referred to as the nickel hydrazone complex. *The names for these compounds were supplied by Kurt L. Loening, Director of Nomenclature for the Chemical Abstract Service. Structures v. Ph Ph \ / c = C / \ o N = N\ CH \ 3 N/ c: 0/ \N _ N/ CH3 \C ___ C/ / \ Ph Ph VI Ph\ Ph C — C/ / \ N N — N CH3 /\ /1 2\/ I C 7 \N\ N - N CH3 R C = C/ / \ Ph Ph IUPAC Names--Abbreviation [[a.,a '-[Isopropylidenebis (aZO)]-di-a-stilbenolato] (2-)1 nickel (Previously published as: l,2,8,9 tetraphenyl — 3,4,6,7-tetraza-5,5-dimethyl— nona—l,3,6,8-tetraen-l,9— dioxynickel (11)) Ni MMK a. (R=H):[3,3-Dimethyl- 6,7,12,13-tetraphenyl-l,2,4, 5,8,11-hexazacyclotrideca- 1,4,6,12—tetraenato(2—)- N1,N5,N8,Nll] nickel Ni-H cyclo 13 b. (R=CH3):[3,3,9-trimethyl- 6,7,12,l3—tetraphenyl-l,2,4, 5,8,11-hexazacyclotrideca~ l,4,6,l2-tetraenato(2—)— N1,N5,N8,Nll] nickel Ni-Mcyclo l3 lO Structures IUPAC Names-—Abbreviation a. (R=(CH NHZ):[a'[[l-[[2— 2’3 [(3Aminopropyl)amino]-l,2- / \ O\\\ ’//N = N\\ CH diphenylvinyl] azol-l- 3 //Ni\\ /C<: methylethyl]azo]-a-stilbenolato R - N N = N CH3 (2~)] nickel Ni-Ap So \2 l/ , i2==0.5 B.M. for a diamagnetic nickel(II) ion and ~ 3 B.M. for a paramagnetic nickel (II) ion. Reaction of Benzil Monohydrazone with Acetone in the Absence of Nickel (II) Ions Procedure II.-—To a hot solution of 2.25g (0.01 moles) of benzil monohydrazone in 100 ml of hot 95% ethanol was added 3.7 ml (0.05 moles) of acetone. The solution was refluxed and stirred for 7 days; on cooling and re- ducing the volume, yellow crystals were obtained. After recrystallization from methanol, thick yellow, hexagonal crystals were obtained, removed by suction filtration, and dried under vacuum at room temperature; mp 76-77°C. This compound was characterized as benzilacetone azine, 0=C(C 6 3)2. Anal. Calcd for C17H16N20: C,77.23; H,6.ll; N,10.60. Found: C,77.05; H,6.06; H5) — C(C6H5) = N—N = C(CH N,10.50. A molecular ion, at a m/e of 264, was observed in the mass spectrum. Other Variations Investigated Other Bridging Species a.2,4 Pentanedione.-—In the above described Pro- cedure I, 109 (10.3 ml; 0.10 mole) of 2,4 pentanedione was substituted for the acetone. Upon mixing the reagents the 17 solution turned a dark yellow-green color. After 20 hours the reaction mixture was filtered by suction and the filtrate allowed to cool. On cooling some pale green crystals precipitated out and were removed by suction filtration. When dry the precipitate appeared to be a mixture of green and yellow crystals. These were separated by dissolving the green substance in water in which the yellow was insoluble. Both compounds were then recrystallized from methanol. The green substance was identified as hydrated nickel acetylacetonate on the basis of its infrared spectrum. The yellow compound was obtained as thick yellow hexagonal crystals. Analysis of this compound showed it to be benzilacetylace- tone azine O = C(C6H5) - C(C6H5) = N — N = C(CH3) - CH — 2 19H18N202; C,74.48; H,5.93; N,9.15. Found: C,74.32; H,5.85; N,9.19. C = O(CH3). Anal. Calcd for C In a second attempt to form a ketazine complex with 2,4 pentanedione a molar ratio of 0.013 : 0.01, moles of 2,4 pentanedione to moles of nickel (II) was used. Upon initial mixing of the reactants the usual dark red color resulted After five days of refluxing and stirring, the solution was filtered by suction and the filtrate allowed to cool. A small amount of irri— descent-red residue remained in the fritted funnel. The infrared spectrum of this material was similar to that 18 of the ketazines. No further batches of this material could be isolated and the small amount obtained was in- sufficient for characterization. This was probably the desired product and probably could be synthesized in larger quantities by carefully controlling the molar ratios of the reactants. The other products of this reac- tion were a red oil, a tan powder, and some yellow crystals. The yellow compound was probably benzilacetylacetone azine, based on the earlier experiment, and the tan powder probably contained Ni(OH)2.xH20, based on its infrared spectrum. Aldehydes.--Attempts were made to substitute an aldehyde for the acetone in Procedure I. A molar ratio of 0.20 : 0.01 moles of butyraldehyde to nickel(II) ions was used. Immediately after mixing the reactants the resulting mixture turned dark red and, after stirring at reflux temperature for seven days the solution turned violet. No residue remained after filtering this hot solution, but the filtrate turned red when stored over- night in a refrigerator and a small amount of green mater- ial (probably nickel hydroxide) precipitated out of the solution and was removed by filtration. Step-wise reduc- tion of the volume of this filtrate resulted only in a viscous red oil after the removal of all solvent. This oil was very soluble in 95% ethanol, but insoluble in petroleum ether, so a mixture of these solvents was tried 19 to induce crystallization. The result was a two-layer system, the lower being a purple alcohol solution which was removed with a separatory funnel. These solvents are normally miscible, which suggests that the separation may have been induced by saturating the alcohol with the pur- ple material. A purple solid was obtained from the alcohol by solvent evaporation. Attempts to recrystallize this material from acetone, or a solution of three parts ethanol and one part benzene, resulted in purple tars. No meaningful mass spectra could be obtained as the compound decomposed too rapidly in the mass spectro— meter. The infrared spectrum of this compound is listed below: 1 (very broad), 1580 cm-1 vs (broad), 3100 — 3600 cm- 1055m, 1030m, 725m, 6833, 622m, 395m, and 310m. Both butyraldehyde and acetaldehyde showed promise of yielding pure new compounds if sufficient work were in- vested in the project, but they were not pursued beyond the preliminary stage described above. Acetaldehyde per- formed similarly to butyraldehyde, so only the work with the latter is described here, as a guide to future workers. Di—chlorides.——Methylene chloride and 1,1—dichloro— ethane were substituted for acetone in Procedure I. The principle product obtained from these reactions was a brown- reddish brown solid with characteristics similar to the nickel hydrazone complex (see page 14). These reactions were investigated to determine whether 1,1-dichloro compounds 20 could serve as bridging agents in the reaction in which the nickel ketazines were formed. A possible "driving force" for these reactions might have been the formation, and subsequent removal from the reaction mixture, of hydrOgen chloride, similar to the formation of water when a ketone was used. 1,1,1 trifluoroacetone.--1,1,l trifluoro- acetone was substituted for acetone in Procedure I. The uniform result of several such experiments was a brown solid with characteristics similar to the nickel hydra- zone complex (see page 14). Other Metal Ions Copper (II), mercury (II), zinc (II), and manganese (II) acetates were substituted for nickel acetate in Procedure I, but no products similar to the nickel ketazines were obtained. The reaction mixture before the addition of the metal had considerable re— ducing ability as evidenced by the formation of elemental c0pper and mercury after their acetate compounds were added. In the zinc case, zinc hydroxide was produced. The manganese acetate solution in ethanol was unstable, and once carefully placed in solution in ethanol and added to the reaction mixture appeared again to undergo decomposition to the hydroxide. The products of these 21 reactions were not completely characterized, but they did not resemble those found in reactions using nickel acetate. Reactions of Ni MMK With Amines [3,3—Dimethyl-6,7,12,l3- tetraphenyl—1,2,4,5,8,Il- hexazacyclotrideca:1,4,6,12— tetraenato(2:7-NI,N5,N8,NIL] nickel Procedure IIIa.--To 10ml of ethylenediamine was added 0.549 (0.001 moles) Ni MMK under a dry nitrogen atmOSphere. The reaction mixture was stirred and heated to a temperature just below the boiling point for 30 minutes. After cooling the solution to ambient tempera- ture a bright red powder was removed by filtration. A second batch of product was obtained by the complete rotary evaporation of the ethylenediamine from the fil- trate. The two fractions of product were combined for recrystallization from acetone. The product was obtained as a bright red powder, or as dark red crystals. For analytical data see Table 1. Procedure IIIb.--TO 10 ml Of ethylenediamine was added 0.54g (0.001 moles) Ni MMK under a dry nitrogen atmosphere. The reaction mixture was stirred at ambient temperature for 48 hours, during which time the solid component changes color to a brighter shade of red. At the end of this time a bright red powder was removed by 22 Table 1.--Analytical Data for the Complexes. Compound % Yielda % Calculated % Found C H N C H N Ni MMK 78.5 68.27 4.82 10.28 68.16 4.89 10.30 Ni MEK 60.9 68.71 5.06 10.02 68.91 5.25 10.22 Ni EEK 39.4 69.12 5.28 9.77 69.19 5.33 9.87 Ni MPrK 58.2 69.12 5.28 9.77 68.76 5.35 9.70 Ni MBuK 60.3 69.52 5.50 9.54 69.32 5.32 9.44 Ni MPhK 14.5 71.18 4.66 9.23 71.05 4.76 9.14 Ni Hcyclo 13 b 69.60 5.32 14.76 69.22 5.28 14.71 Ni Mcyclo 13 b 69.99 5.54 14.41 69.28 5.54 14.36 Ni ApSo b 67.89 5.71 13.98 67.54 5.93 14.75 aBased on nickel (II). b . . . . . . Due to difficulty 1n recrystallizatlon accurate yields could not be attained. 23 filtration. A second batch of the product was obtained by the complete rotary evaporation of the ethylenediamine. The two fractions of product were combined for recrystal- lization from acetone. The product was obtained as a bright red powder or as a bright red crystals. This material was shown to be identical to the product of Procedure IIIa. Racemic 1,2 diaminOpropane, when substituted for ethylene diamine in Procedures IIIa and b yielded similar products. A reaction involving Z-l,2 diamino— propane was carried out using Procedure IIIa and re- sulted in the formation of an optically active product. This product was obtained as a waxy red material. Comments.--The products described in this section could not be crystallized well from a wide range of available solvents. The products generally dissolved with difficulty and then precipitated only after the loss of most of the solvent. The reaction times indicated in Procedures IIIa and b have been shown to yield the desired products, but shorter reaction times may also have been effective if they had been investigated. [a'—[[1-[[2—[(3 Aminopropyl) amino -1,2-diphenyviny1]azo]— l-methylethyl]azol-d-stilbeno— lato(2-)] nickel When 1,3 diaminopropane was substituted for ethyl- enediamine in Procedure IIIb a red powder was obtained. 24 This material could be recrystallized from methanol or ethanol, yielding a bright red crystalline material. For analytical data see Table 1. When 1,3 diaminopropane was substituted for ethylenediamine in Procedure IIIa, some of the starting material, Ni MMK, was isolated, as well as a red oil which could not be induced to crystallize, even by a chromatographic purification procedure. Reaction of Ni MMK with Ethylamine * Procedure IV .--Use of Pressure Tubes. To 0.54g (0.001 moles) Ni MMK, which had been placed inside of a pressure tube, was added 15 m1 of ethylamine. A stopcock was attached to the end of the tube by means of a short section of vacuum tubing. The reaction mixture in the tube was gradually cooled to -196° by immersing the tube in a liquid nitrogen bath. After being frozen, the tube was evacuated by pumping for 15 minutes. The stopcock was then closed and the tube was allowed to warm to near room temperature, at which point the tube was shaken and then reimmersed in the liquid nitrogen bath. The cycle of gradually freezing, pumping, warming, and shaking was repeated three times before the contents of the tube were finally frozen and sealed with a torch. The tube was slowly warmed to room temperature and then clamped in an *This procedure was used to control the atmos- phere over a reaction and also the temperature at which the reaction was allowed to occur. 25 oil bath behind a shield. The oil bath was heated to about 100°, but not over 115°, and held there for three hours, then it was allowed to cool to room temperature. At room temperature, the tube was removed from the oil bath, wiped clean, and gradually cooled to -196°. While the tube was still cold, the glass was broken to Open it. The tube was warmed slowly in an upside down position, while wrapped to catch condensed water. The solution dripped into a receiving vessel when it reached its melt- ing point, and on further warming the amine evaporated, leaving the red product behind. A pure sample of this product could not be isolated, in part this seemed to be caused by decomposition during purification steps. Caution! During the heating cycle of one of these reac- tions the end of the pressure tube shattered and the reaction mixture shot out. Comments.--Both the time and temperature indicated in this procedure appeared to be necessary for the reaction to occur. 1,3 Diaminopropane was allowed to react with NiMMK in a manner similar to Procedure IV. The changes re- flected the higher boiling point of the diamine (160° vs 14°). The product was initially obtained as a red oil, from which a white substance could be removed by dissolv— ing the red material in methylene chloride. A thick red oil was obtained by rotary evaporation of the filtrate. 26 An attempt to purify this material by chromatographic means yielded Ni Ap So. p-Toluenesulfonyl derivative of Ni Ap So.—-To a solution of sodium hydroxide in 10 ml of water was added 0.1 g (excess)p-toluenesulfonyl chloride. A second solu- tion, composed of 0.10 g of Ni Ap So dissolved in 5 ml of benzene, was shaken together periodically with the first solution over a period of several hours, after which the benzene layer was removed and the solvent was removed by rotary evaporation. This residue contained unreacted p-toluenesulfonyl chloride which was removed by sublimation, leaving behind an orange-red material. This product was identified as a mono-p-toluenesulfonyl deriva- tive of Ni Ap So by its parent ion at m/e 784 in its mass spectrum. Dehydrogenated Compound A solution of 0.24g Ni Ap So in 350 m1 of n-butanol was refluxed for six hours. After reducing the volume of the solvent and cooling, a red solid was ob- tained. The mass spectrum had a parent ion (m/e 596) which was four mass units lower than Ni Ap So. A summary of the infrared spectrum of this product is presented in Table 6. 27 Reactions of Ni MMK With Other Amines To 50 m1 of liquid ammonia was added 0.5 g Ni MMK. The reaction mixture was stirred under a dry nitrogen atmosphere for 1 hour while refluxing with a dry ice con— denser. The ammonia was then allowed to evaporate and be carried off by the nitrogen flow. The red product which remained in the reaction vessel, was recrystallized from n-butanol to give red needles. The mass spectrum of this product identified it as Ni MMK. In separate reactions, aniline and 3,3'imino- bispropylamine [H2N(CH2)3NH(CH2)3NH2] were substituted for ethylenediamine in Procedure IIIb. In both cases, the red product obtained was recrystallized from n-butanol and identified as Ni MMK by its mass spectrum. 3,3' Iminobispropylamine was also substituted for the ethyl— amine in a procedure very similar to Procedure IV. A red solution was removed from the pressure tube and subjected to rotary evaporation to remove the amine. The product obtained was a red oil which could not be induced to crystallize. The mass spectrum of this oil did not indi— cate the presence of Ni MMK nor any of the products which were expected to be formed in this reaction. When o—phenylenediamine was substituted for either the ethylenediamine in Procedure IIIa, or the ethylamine in Procedure IV, the red products obtained from both of these reactions were identical. These 28 products were both identified as Ni MMK by their mass spectra. Ni MMK was allowed to react with o-phenylene- diamine by the method of Jager6 (similar to Procedure IIIa). The products of this reaction appeared to be decomposition products of the starting materials, o-phenylenediamine and Ni MMK. When 1.1 g o-phenylenediamine was added to a hot solution of 0.54 g Ni MMK in n-butanol and stirred at reflux temperature for three hours, a red product was obtained in the form of red needles. The mass spectrum of these needles identified them as NiMMK. Other Reactions Investigated About 30 ml of ethyl iodide was distilled (66-68°) into a round bottom flask containing 0.27 g of Ni MMK. The reaction mixture was stirred at reflux temperature for 6 hours, after which the ethyl iodide was removed by rotary evaporation, leaving behind a red residue. The red material was recrystallized from acetone and identi— fied as Ni MMK on the basis of its infrared and mass spectra. To 0.27 9 Ni MMK was added 50 ml of pyridine. The Ni MMK dissolved completely and the solution was stirred for 3 hours. The pyridine was then removed by rotary evaporation and the red residue was recrystallized from n-butanol. The product was obtained as red needles and 29 was identified as Ni MMK on the basis of its mass Spec— trum. Also a 4.97 x 10-5 M solution of Ni MMK in pyridine was prepared. The UV-visible spectrum was taken both im- mediately after the solution was prepared, and several days later. In both cases the spectrum observed was identical to that of Ni MMK in benzene. Two 3.68 x 10-3 M solutions of Ni MMK in chloro- form were prepared under a dry nitrogen atmosphere and amounts of 1,2 diaminopropane were added to give respec- tive molar ratios of amine to Ni MMK of 25-1 and 120-1. The volumetric flasks were sealed with electrical tape and stored in the dark for 7 days. Aliquots were then taken and diluted to give a theoretical concentration of Ni MMK of 7.4 x 10-5 M. At this point purple crystals were observed in the original flasks. The UV—visible spectrum of the diluted solution was taken and compared to SMNiMMK the spectrum of a known solution of 7.4 x 10- in chloroform. The solution with the amine indicated the presence of Ni MMK, but in a lower concentration than the known solution, indicating that decomposition had taken place. When 1,3 diaminOprOpane was substituted for 1,2 diaminopropane in the above procedure, similar results were obtained. 30 Reactions of Benzil Monohydrazone With Amines To a hot solution of 2.25 g (0.01 moles) benzil monohydrazone in absolute ethanol was added 0.34 ml (0.005 moles) ethylenediamine. The clear yellow solution was refluxed for 24 hours and then cooled overnight in a refrigerator. The next day a white crystalline precipitate was removed by filtration. This material was identified as benzil monohydrazone on the basis of its infrared and mass spectra. This was the only product of this reaction. This reaction was repeated but with the addition of 1 drop of concentrated sulfuric acid to serve as a catalyst. The white crystalline product of this reaction was identified as benzil monohydrazone on the basis of its infrared and mass Spectra. 0.34 ml (0.005 moles) ethylenediamine was added to 175 m1 absolute ethanol in a round bottom flask. 0.25 g (0.01 moles) benzil monohydrazone was placed in an asbestos thimble of a soxhlet extraction apparatus. The ethanol solution was refluxed and the condensate was used to leach the benzil monohydrazone from the thimble into the round bottom flask containing the diamine. The white product isolated after the reaction was identified as benzil mono- hydrazone by its infrared and mass spectra. To a 95% ethanol solution containing 0.01 moles of benzil monohydrazone and 0.012 moles of ethylenediamine 31 was added 0.005 moles of hydrated nickel chloride. The reaction mixture was stirred at reflux temperature for 6 hours. The white product obtained from this reaction was identified as benzil monohydrazone on the basis of its infrared and mass spectra. To 50 ml of fresh 91—93% ethylenediamine was added 2.25 g benzil monohydrazone. The solution was stirred at reflux temperature for two hours. The excess amine was then removed by rotary evaporation and the yellow residue was recrystallized from 95% ethanol. The product was obtained as yellow crystals and was identified as 2,3 diphenyl-l,4 diazine* on the basis of its mass spectrum. To a hot ethanolic solution of 0.01 moles of benzil monohydrazone were added 1 drOp of concentrated sulfuric acid and 0.005 moles of o-phenylenediamine. The solution was stirred at reflux temperature for 3 hours. After reduction of the solvent volume, a small amount of white needles were obtained. This product was identified * N /\.\ h ' P HC \c/ 2 HZC 32 as 2,3 diphenyl quinoxaline* on the basis of its infrared and mass spectra. To a hot ethanolic solution of 0.01 moles benzil monohydrazone were added .005 moles of hydrated nickel chloride and 0.005 moles of o-phenylendiamine. The reaction mixture was stirred at reflux temperature for 2 hours. A pale green precipitate was removed by filtration. After cooling overnight in a refrigerator, a lavender crystalline material precipitated from the filtrate and was removed by filtration. The green product was identical to a green material produced in 87% yield when the reaction was repeated in the absence of benzil monohydrazone. This substance was Ni (o-phenylenediamine) C12. The lavender crystals could not be characterized. No mass peaks higher than m/e of 160 were seen in the mass spectrum. The infrared spectrum gave strong broad absorptions between 3100 and 3500 cm-1, probably due to N-H absorptions also several unusual bands of medium intensity were observed between 2100 and 2300 cm-1. The following is a list of the major infrared absorptions of the material in cm-1 3100-3500 s(br), 2260m, 2180m, 2120m, 16555, 1608vs, 1590vs, 1328vs, 1273vs, 1155v5, 10965, 10755, lOlOvs (br), 981vs, 7205, 660vs(br), 525vs, 501vs, 325m. 'k /N\\C/¢ ' l [::)\\N¢C\\¢ IV. PHYSICAL MEASUREMENTS Infrared spectra were obtained using Nujol or Fluorolube mulls on a Perkin-Elmer Model 457 spectrometer. Visible and ultraviolet spectra were obtained by use of a Unicam SP 800 B spectrophotometer. The reported proton magnetic resonance spectra were determined using a Varian HA-100 nmr spectrometer by a repetitive scanning technique. H2804 was used as an external reference for the instrument to "lock on" during the scanning. Chemical shifts were determined by the subsequent introduction of TMS as an internal reference. This procedure was necessitated by the high molecular weights and low solubility of the compounds in suitable solvents. Mass spectra were determined with an Hitachi-Perkin-Elmer RMU-6 O, and CEC 21-110 B double-focusing mass spectrometer. Accurate masses were measured with a Nier type peak matching unit attached to the latter instrument (accuracy ~ 4ppm). Elemental analyses were performed by Spang Microanalytical Laboratories, Ann Arbor, Michigan, by Galbraith Labora— tory, Inc., Knoxville, Tennessee, by Chemaltyics, Tempe, Arizona, and by the Water Resources Department of Michigan State University. 33 V. RESULTS AND DISCUSSION Reaction between benzil monohydrazone and a series 1RZCO in the presence of nickel (II) ions has of ketones R been found to yield a series of neutral Ni cis N202 tetradentate complexes containing a ligand formed by condensation between the ketone and two benzil monohy— drazone residues. The reactivity of one of these co- ordinated ligands toward amine compounds has resulted in the formation of two types of complexes, the macrocyclic Ni N4 and the non-cyclic Ni N30 complexes. These new compounds are formed by the condensation of the nickel ketazine with one or two amine groups and the elimination of one or two molecules of water. The analytical data which were obtained for these complexes are presented in Table 1. The nickel ketazines and the compounds formed from them will be discussed in separate sections below. Nickel Ketazines The structures of the ketazine ligands were . . l deduced by their infrared, H nmr and mass spectra. The infrared spectra of the nickel ketazines exhibited numerous 34 35 absorptions but many similarities existed in the series. Table 2 contains a list of the common features of the infrared spectra of these complexes. These spectra differed significantly from the spectrum observed for benzil monohydrazone, which exhibited strong absorption bands at 3390, 3275, and 3180 cm“1 assigned to v(NH2) and a strong broad band at 1620 cm-1 assigned to 6(NH2) and v(C=O). The absence of these absorptions in the nickel ketazines was accompanied by a new strong band centered near 1280 cm_1 attributed to v(C-N) and v(C-O). Also the strong complex band at 1530 cm_1 in benzil monohydrazone, assigned to v(C=C) from the phenyl rings and v(C=N), is generally absent in the nickel ketazine spectra. These changes are consistent with the NH2 groups of benzil monohydrazone having undergone condensation with the reacting ketone. A summary of the 1H nmr spectra obtained from the nickel ketazines is presented in Table 3. The solubility of the nickel ketazines was very low in solvents suitable for nmr studies, including CCl4 in which the spectra were obtained. As a result of this problem, repetitive scan, time averaging techniques were employed to obtain reasonable resolution of the spectra. However, even using these techniques, some of the more complicated resonances could not be resolved. The spectra consist of two regions, an alkyl region, 0 0.9-2.2 and a phenyl region about 36 m m E m mmm swam mam ascend mo > “mlmmovo m> m> m m> «as -mmn Hos NHmQOHumunfl> mcflnumoun mcflu mcmnemm m m E E mooanmmm oooa m> m> a m mmoanmmoe mmoa ouo In omusuflumnsm ocoe m> m m m whoauesofl mace n omusuepmnsm oaos uouo lmxmmovo m m> m m> ooeaummae mmaa o-o KSmmoco m m> a m seeflnqoafl «see malzuov> “ouo e m> m m oama-aame meme ouo m> m> m m maveumaea cave in omusuhumnsm oaoe m m a s omeeuemva wave ZIO luznzl “mGOAumHOQO oaumEoum 3 E 3 E owmalmema Emma -zuz- “one omummsnhoo a a a m Hooeummma some mmaeumpms how muamecmhmma .m.H mgmzhz Msmzaz sumzhz xmmhz smzez xzsz mmsmm hmucmo pcmm pcmm wcflumumm/muflmemch .mocflumumm me0fiz mo muuommm pmumumcH on» Ga mpcmm GOAUQHOde 50EE0011.N mqmda 37 TABLE 3.—-Proton Magnetic Resonance Spectra of Nickel Ketazines. Compound Chemical Shifta Type Assignment NiMNK 1.68 singlet —CH3 7.06-7.47 multiplet -C6H5 NiMEK 0.96b triplet -CH2.CH3 1.65 singlet -CH3 ___ 2.17C quartet —CH2.CH3 7.06—7.44 multiplet CgHS NiEEK 0.92b triplet -CH2.CH3 2.20C quartet —CH2.CH; 7.05-7.40 multiplet -C;H5 NiMPrK 0.91b triplet —CH2.CH2.CH3 ~1.40 multiplet -CH2.CH2.CH; 1.65 singlet —CH3'_—_ 2.12b triplet -CH2.CH2.CH3 7.06-7.39 multiplet -C;H5 NiMBuK 0.92b triplet ~CH2.CH2.CH2.CH3 1.22-1.43 multiplet -CH2.CH2.CH2.CH; 1.64 singlet —CH3 ___-___- 2.15b triplet -CH2.CH2.CH2.CH3 6.94-7.49 multiplet —C;H5 NiMPhK 2.06 singlet -CH3 7.11-7.30 multiplet -C6H5 ap.p.m. downfield from tetramethylsilane. bCenter of triplet. cCenter of quartet. 38 0 7-7.5, both downfield from TMS. As an example of the spectra obtained, the spectrum of NiMEK is presented in Figure 1. No resonances which might be attributable to N-H groups, were observed. The relative areas under the peaks appeared to be in ratios consistent with the assignments made (coupling constants, J -H’ were calculated to be H 7.0-7.7 cps from spectra exhibiting triplets and quartets). For all of the ketazine complexes containing R1 = Me, except the case where R2 = Ph, a sharp singlet, attributable to this methyl resonance, was observed close to 6 1.65 However, when R2 = Ph, this methyl singlet was shifted downfield to 6 2.06, indicating that the methyl group had been deshielded by the phenyl group. A molecular framework model of NiMPhK based on the proposed structure showed the methyl group to be positioned, relative to the phenyl group, in such a way as to make possible the observed deshielding. The observed lH nmr spectra are consistent with the formulation of the tetradentate ligands of the nickel ketazines as proposed in Structure V. Further confirmation of the structure of the ketazines was provided by a study of their fragmention patterns in their mass spectra. For each pure compound in the series, the set of highest mass peaks corresponded to the singly charged molecular ion for the compound. The 39 mod .Asmzezv .mceumuwx assum assume meoec mo ashuommm has me he: OOH .H .oem See a 03 o3 ooN on.“ no.0 ooN onN e e a e _ l e l ll Z H II Z CN/ CPh\ Structure V peak of highest intensity in this set corresponded to an ion containing 58Ni followed by peaks at higher masses with lower intensities representing the presence of other isotopes of nickel, carbon, and nitrogen. The fragmentation pattern for this series of compounds was quite uniform and is summarized in Table 4. The tentative assignments made in this table can easily be accounted for on the basis of the proposed structure. A notable feature of these residues is the stability of the five-membered chelate ring, possibly stabilized by electron delocalization. A similar stability of a six-membered chelate ring has been observed by Cummings and Sievers in the mass spectra of two nickel (II) macrocylic complexes.l3 One of the most notable characteristics in these mass spectra was the feature of the three highest mass peaks. The highest peak corresponded to the molecular ion 41 TABLE 4.--Mass Spectra of Nickel Ketazines. M/ea Assignment P [Parent Ion]+ P'28 [P - N2]+ P-56 [(P - N2) - CO]+ P Ni 1- / ~\ R I \ / P-222 O\\ . N—N C \ ’ / /c— c\ \{2 Ph Ph : + Ni /, \\ 280 O \ /‘ N—‘—N \" / /C-C\ Ph Ph L + 1’ Ni ‘ // ~\ 266 q\\ giq C-C / \ Ph Ph L— .4 240 [Ni.O.C. (Ph)2]+ 238 [Ni.N.C.(Ph)2]+ 224 [Ni.C.(Ph)2]+ /Ni\ 203 o; ‘N-‘-N /C-—C (Eh l 192 [Ph.C.C.(N).Ph]+ 178 [Ph.C.C.Ph]+ 42 TABLE 4.--(Con't.). a M/e Assignment 166 [C.(Ph)2]+ 135 ? 117 [Ph.C.CO]+ or [Ph.C.N2]+ 105 [Ph.CO]+ 99 ? 91 [Ph.N.]+ 77 [Ph]+ a . . . . . . . For nickel containing spec1es, the value for 58N1 15 glven. 43 and the next two peaks occurred at P-28 amu and at P—56 amu (again a separation of 28 amu from the next higher peak). These losses could correspond to the loss of CO or N2, or some combination of the two, since they both have the same nominal mass of 28 amu. High resolution mass spectrometry was used to resolve the uncertainty. Table 5 presents the theoretical possibilities and the results of an exact mass measurement for NiMMK. Thus it can be seen that the [P-28 amu]+ peak for NiMMK represents the loss of an N2 fragment and that the [P-56 amu]+ peak represents the loss of a CO fragment from the [P-28 amu]+ residue. It was further shown that these losses were consecutive rather than concurrent by a measurement of metastable ions by the defocusing technique. A similar consecutive fragmentation pattern has been observed in the mass spectra of some triazinones* (compounds which contain both > C=O and -N=N- groups).14 This similarity strongly suggests the presence of an -N=N- linkage in the nickel ketazines and is consistent with the proposed structure (see Structure V). On the basis of these measurements and the similarity of the mass spectra of the other nickel ketazines, it was concluded that their fragmentation also proceeds by an * /. /N\N l R .\C/N\R ll O 44 TABLE 5.--Results of High Resolution Mass Spectrosc0py Investigation for NiMMK, (R1 = R2 = Me). Exact Mass. m/e Possible Assignment Theoret. Measured 544 m? 544.1409 544.1387 516 mf-NZ 516.1348 516.1343 mf-co 516.1460 488 NT-Nz-co 488.1399 488.1375 mf—zNZ 488.1286 mf-zco 488.1511 45 initial loss of N followed by a loss of CO from the 2, initial residue. However, for benzilacetone azine, the initial loss of 28 amu probably represents the elimination of a CO fragment because the two nitrogen atoms are linked by only a single bond and their loss would break the molecule into several fragments. The ultraviolet-visible spectra of the nickel ketazines, from 300-850 nm, were obtained in approximately 5x10"5 M benzene solutions (in which they were quite‘soluble). Care was taken to run the spectra soon after preparation of the solutions because benzene appeared to accelerate photodecomposition. All of the nickel ketazines gave very similar spectra, which suggests that the nickel (II) ion is in a similar environment in all of the complexes. These spectra are presented in Table 6. The absorption band near 495 nm has been assigned to the 1A + 1B transition for 1 1 nickel (II) in a Ni cis N202 planar environment,15’ 16 which satisfies the requirements imposed by the nature of the tetradentate ligand. If this assignment is made, then the other bands may be assigned as charge-transfer and ligand bands. The position and molar extinction coef- ficients of the lower lying bands are probably influenced by the strong ligand bands at 390 and ~ 340 nm. All of the nickel ketazines were obtained as dark red to orange-red crystalline solids which were found to be diamagnetic (ueff : 0.5 B M) when pure. As initially 46 .mcomcme .Ee 5H Q CH omcflmuno muuowmm Hadm m.¢a mmm e.wa omm va.m em mmw me.e mmv memzflz m.vH mwm m.oa omm me.m Sm owe mv.n wow msmzflz m.va mvm n.oa omm vo.m gm owe Hm.> mmv mumzflz N.mH mvm m.ea omm mm.m em owe vm.n mmv Mmmflz m.mH mvm m.eH omm oo.m gm owe mm.e mme mmzflz e.mH vvm m.na omm om.m em owe no.5 mme Mzsz Ioa x nv> mIOH x 2o nm> mnoa x 20 Qm> mnoa x 2w QH> pssomEoo m .mmsflumuwm meoez mo muuommm manflmfl> can umaofl>nmuuaaun.m mamme 47 isolated, the complexes usually contained small amounts of ferromagnetic impurities. These impurities could be removed by repeated recrystallizations (if stored in the dark) or by passing a benzene solution of the complex through an alumina column (see experimental section for details). All of the data obtained for the nickel ketazines support the structure as shown in Structure V. In the alternate structure the adjacent nitrogen atoms would be linked by a single bond and the coordinated nitrogens would form a four-membered ring with the nickel ion and the bridging carbon atom. However, the presence of a nitrogen- nitrogen double bond has been demonstrated and molecular framework models have indicated that the alternate structure would involve greater strain than Structure V. This is the first example in which a ketone reacts with coordinated nitrogen atoms to form a single carbon atom bridge. Thus, it represents a new type of condensation reaction, previous types of condensation reactions resulting in the formation of new chelate rings have necessitated the presence of, or the formation of a three carbon atom bridge between coordinated nitrogen atoms. This type of condensation reaction offers a direct route for introducing a large degree of unsaturation into metal chelate ring systems. Unsaturated chelate ring systems have been a goal of previous synthetic investigations,3 48 one reason for this being attempts to approximate the conjugated chelate ring systems of naturally occurring compounds. These previous attempts to introduce un- saturation have often required additional dehydrogenation steps after the initial formation of the metal complex. The chelate ring formed by the single bridging carbon atom in this new type of condensation contains six members, this is the same size ring as has been observed in other condensation reactions2 and the stability of five— and six— membered chelate rings has been observed here and in . . . . 13 prev1ous investigations. The reaction resulting in the formation of these ligands is best described as a template reaction3 since it is dependent on the presence of the nickel (II) ion. In the absence of nickel (II) ions, benzil monohydrazone and acetone undergo a Schiff's base condensation reaction to give benzilacetone azine. Benzilacetone azine may function as an intermediate in the formation of NiMMK, in that Taylor and associates10 demonstrated that NiMMK could be formed, in low yield, by refluxing nickel acetate and benzil monohydrazone with this compound. However, excess benzil monohydrazone is necessary for this reaction to take place. When nickel acetate is added to an ethanolic solution of benzil monohydrazone a dark red-brown product is formed immediately. This same compound appears to form 49 as an initial precipitate in the condensation reactions of the ketones. Both of these dark reddish-brown substances have infrared spectra similar to the nickel hydrazone complex (gage infra). In the case where the ketone is benzophenone, the only product obtained is a dark reddish- brown crystalline compound which is probably a pure form of this compound. Earlier workers in this area also reported the formation of a similar dark red solid, but they were unable to obtain consistent analyses. The crystalline form of this dark solid (serendepitiously isolated!) has now been characterized as N12 (L-H)2 (L-2H), where (L—H) and (L-2H) are benzil monohydrazone minus one or two amine hydrozen atoms respectively (see Structure VIII). It is likely that the amorphous solids obtained by mixing nickel acetate with alcoholic benzil monohydrazone or as the initial precipitate in the ketazine formation reactions are similar to Ni2 (L-H)2 (L—2H), but polymeric or more randomly complexed species may also be formulated using the same chemical units. The extremely viscous nature of reaction mixtures containing this substance immediately after its formation may lend credence to these possible species. If this did occur, it might help to explain the inconsistent analyses. Two other possible explanations for these earlier inconsistencies are that the material is insoluble in many organic solvents and does not recrystallize well from those in which it does dissolve and also that it undergoes rapid decomposition in solvents such as benzene which are used 50 to remove the nickel hydroxide impurities of these reactions. Ph Ph \ _ / /C — C\ /H /Ph O N = N //O——— C\ \Ni/ \Ni \C—Ph / \ / \ / o. N = N N = N \ / \ C = C H / \ Ph Ph Structure VIII As the ketazine formation reactions progress, this initial complex slowly disappears (probably through a thermodynamic equilibrium favoring other product formation) and the stable orange-red nickel ketazines are formed. The ketazine formation reaction has not been studied in great detail, but it appears to involve an equilibrium concentration of benzil monohydrazone, with the nickel (II) ions coordinating the intermediates or products which are formed. To summarize briefly, a new series of non-cyclic Ni cis N202 tetradentate complexes have been synthesized. These complexes have been characterized by infrared, lH nmr, ultraviolet-visible, and mass spectroscopy as well as elemental analyses and magnetic susceptibility measure- ments. In the next section, the reactivity of one of these complexes (NiMMK) toward amines will be examined. 51 Reaction Products of NiMMK At about the same time that the nickel ketazine compounds were being characterized, other authors6' 7’ 8 were reporting condensation reactions of coordinated carbonyl groups with amines resulting in the formation of macrocyclic complexes, in compounds similar to the nickel ketazines. Previously, the coordinated carbonyl groups of salicyclaldehydes had been shown to react with amines to form Shiff base type compounds, but the formation of macrocyclic complexes was a new and interesting develop- ment. Since the conditions necessary for these cyclization reactions to occur and the mechanisms of these reactions were not entirely understood, an investigation of the reactivity of a nickel ketazine towards amine compounds seemed like a potentially interesting and useful area for research. Reaction between NiMMK and some neat mono- and di- amine compounds yields two types of neutral complexes, a macrocyclic NiN4 type (Structure VI) and a non—cyclic NiN3O type (Structure VII). These compounds are conden- sation products of the ketazine resulting in the replace- ment of one or both oxygen atoms by nitrogen atoms and the elimination of one or two molecules of water. The analytical data for these compounds have been presented in Table 1. The structures of the ligands have been deduced from the infrared and mass spectra of the complexes. 52 Ph Ph Ph \C = C/ \C= CPh/ / \ c)/ C\ //N N = N O\\\N ///N = N H2]: \ \ /CH3 \/CH3 . C C HC / \cH3 Rl— N/N \N = N/ \CH3 R/\N‘ N = N N\C = C/N / Cph\ \C = C/ P. PH/ \Ph Structure VII Structure VI H nmr spectra have been unobtainable and purification of some of the compounds has proven to be difficult due to their low solubility in most solvents. The character- ization of these two types of compounds will be discussed in separate sections below, beginning with the macrocyclic NiN4 complexes. Macrocylic NiN4 Type Compounds These macrocyclic complexes are formed by reactions of NiMMK with neat 1,2 diaminoethane or 1,2 diaminopropane at room temperature or higher in a dry nitrogen atmosphere. Their infrared spectra were complicated in that numberous absorptions were observed, but they bore some resemblance to that of the parent compound. The common features of the infrared spectra of these compounds are reported in Table 7. Also presented in this table are 53 commumcfi Hgmmcflsme m m>fig on Ehom musm maucmflofiwmsm M GA pmumaomfl mQ nos pagoo ommazm .mwmomusm comflummEoo mom pmumfla mum MEZHZ mo mcoflumuode Q .Eduuommm 3 E 3 mam loom mam Emmm E 3 3 3 omm Imam omm m3 m> m> m> Hoe 1000 ooe Nfimmovo m 65 E 8 mm» -mme mme E E E 3 E mew navw vvm m E E mmm namm mmm mamcoaumuQfl> msanmmuQ mcau mcmuch E E m E E Nooanemm oooa mamlu oaumEonm is oopsuflqudm ocoE m meoa m m> m m mnoasoooa onoa Nemuo oepmaOHM mlmmovo m m Ashcm>eoefl m» m> eeeeloeea mafia E E m E E Namanmoma oama m>ommH m m> m> m> memaloEma mama mafizuuv? m> m> m> m> womanmmma mama NaAZIUV9 m mama m> m> . . mama m> m> m> m> m> mvmalmvma ovma “e pmpsuflquSm ocoE Emmva m> m movanmmva mmva znonlznzn xmcoHQQEOQO oaumEoum E E 3 E nemanmema enma Neono luznz- uouo emnmmshcoo m m5 m E s mooa-emme mama MEZHZ ommmloaz ommmflz ma oHo%osz ma oohomaz mmcmm umpcou mucmEcmHmma team .m .H Q ocmm ocmm pcsomEom/AuamsmucH m.m©csomEou Omzflz pcm vsz mo ouuommm pmnmumcH mQQ Ea mpcmm coaumuode EOEEOOII.e mqmde 54 some infrared absorption bands of NiMMK which is presented for comparison purposes. In these spectra 1 the strong band at 1420 cm- v(C-O) of the nickel ketazines is absent and a strong new band at 1295 cm"1 1 it seems that the v(C-N) appears. Below 1300 cm- bands in common with NiMMK are shifted slightly in the direction of lower wave numbers. Other notable features of these spectra include the consistent intense absorptions at 700 cm-1, in contrast to a frequent, but not 1 of the ketazines, and the characteristic band at 704 cm- band at 512 cm-1 of weak to moderate intensity which is unique to these daughter compounds. In the mass spectra of these compounds, singly charged molecular ions corresponding to a 58 Ni isotope, have been observed for each of the compounds. The envelope of mass peaks surrounding the molecular ion is very similar to that observed for the nickel ketazines. The compounds also exhibit the loss of an N2 fragment, as shown by high resolution mass spectroscopy (see Table 8). A summary of the fragmentation patterns of these compounds in their mass spectra is presented in Table 9, along with tentative assignments based on the proposed structures for the reported mass peaks. The table is divided into two sections in order to illustrate the similarities observed for these compounds. In the first section, the leaving groups are shown because for each compound the residue 55 TABLE 8.—-Resu1ts of High Resolution Mass Spectroscopy Investigation for Ni Mcyclo 13. Exact Mass m/e Assignment Theoret. Measured 582 NT 582.2042 582.1995 544 NT—N2 554.1980 554.1930 would be different, while in the second section the remaining molecular fragments have lost the disimilar parts of the parent ions, so the identical residues are shown. The mass spectra of these compounds display a similarity to those of the nickel ketazines in some of the residual fragments and also in the stability of the five-membered chelate ring system. The loss of an N2 fragment suggests the presence of a nitrogen-nitrogen double band and the resemblance to the nickel ketazine mass spectra suggests structural similarities. The data obtained from the mass spectra of these compounds supports the proposed Structure VI for these compounds. The ultraviolet and visible spectra over the range of 300—850 nm are reported in Table 10, along with the circular dichroism spectrum of Q-NiM cyclo 13. Both the location and the measured absorbances of these bands are probably inaccurate due to the strong charge transfer bands at the high energy end of the spectra. This is shown in Figure 2, where the spectra of the macrocycles 556 TABLE 9.—-Mass Spectra of NiN4 and NiN3O Compounds. m/ea Assignment Leaving Group Assignment Residue P Parent Ion b p 15 CH3 P-28 N2 P-56 NC(CH ) c 3 2 C P 84 N-N f‘CHa’z N b c P 131 Ph c N (CH2)2 p-159 Ph-C-N=N-C(CH3)2c P-187 Ph-C—N=N-C(CH3)2-N=N P-276 Ph-C=C(Ph)N=N-C(Cfl3)2c p-290 Ph-C(N)=C(Ph)N=N-C(CH3)2c p-332 -(CH2)-N-(Ph)C=C(Ph)N=N-C(CH3)2NC 306b F N N + \\ // Ni N /\’:I\ n- \c=-c/ /’ Ph Ph.J 292b FN\ " 1 Ni N\/: ) /N-N c=c fh/ Ph_ — ‘1 278 Ni\ + N<::') N-N ‘ / /c=c\ Ph Phi 265 - Ni _ . N/' 9N-.. '\ ‘ / c=c / \ Ph PhJ 250b [Ni-N-C(Ph)-C(Ph)l+ 236 [Ni-C(Ph)-C(Ph)]+ 202 Ni\| * H-N/C) N- \ ./ c-c / Ph 193 [PhC(NH)CPh]+ 178 [PhCCPh]+ 77 [Ph]+ aFor nickel-containing species the value for 58Ni is given. bFor macrocyclic compounds only. cResidue contains a five-membered chelate ring. 57 TABLE 10.--U1traviolet and Visible Spectra of NiN4 and NiN3O Compounds. Compound A, nm (e)a Ni H cyclo 13 Ni M cyclo 13 (racemic) Ni M cyclo 13 (1) Ni ApSo NiESob CD Spectrum of Ni M cyclo 13 (2) 558(63.l4), 350(817.5) 546(82.02), 334(615.8) 546(82.15), 345(815.9), 558(€l.20), 420(812.2), 510(E7.5), 348(814.5) A, nm (As) 495(+0.11); (+0.44); 350(+O.63); 464(68.9), 534(64.7l), 518(63.64), 518(83.43), 320(619.3), 532(82.79), 340(810.8) 426(-0.07)7 459(812.2), 440(69.90), 450(89.0l), 300(824.0) 440(812.2), 405(817), 364 sh 307(-0.09) aAll a values are x 10—3 and all A are in nm. e are based on an estimated concentration. 58 1.0 ] Absorbance vs. Wavelength (nm) 1.0 0.8« 0.6J 0.4- 05% 0'0 1* l l I l 300 320 350 400 450 500 550 600 Fig. 2. Electronic absorption spectra, in benzene, of NiMMK (---), Ni H cyclo 13 ( ), and Ni M cyclo 13 (___-"'-)o 2.04 Absorbance vs. Wavelength (nm) L8< 16d L4- 1.2-( LO- 08F (165 (14- (12% (lOfif r 1 300 320 350 400 450 500 550 600 Fig. 3. Electronic absorption spectra, in benzene, of Ni ApSo (—-—-), and Ni ESo ( ). 59 Ph Ph \C = C/ N/ \N Hzc /NN\ /N =N\ /CH3 I / h / \ _ \.3 R \C= C/N M \. Structure VI are superimposed on that of NiMMK. Normally the substi— tution of a methyl group for a hydrogen atom on a chelate ring system produces little change in the electronic spectrum of the compound12 but some distinct changes are noted in this case. In the electronic spectra of square— planar nickel (II) complexes, theory predicts that three or four transitions, depending on the symmetry of the complex, should occur in the d-orbital manifold of the metal ion. For these compounds, the d—d transitions are probably represented by the low energy shoulders in the spectra; they appear to be partially obscured by the charge—transfer bands which occur at lower energies than in the nickel ketazines. These shoulders may represent the 1Al + lBl transition observed in the spectra of the nickel ketazines. Bands near 650 nm would be expected if these complexes were tetrahedral, so their absence suggests a square-planar 60 environment for the nickel (II) ion. The circular dichroism spectrum of k-Ni M cyclo 13 exhibits a positive cotton effect and is shown in Figure 4. This CD spectrum exhibits better resolved bands than the ultraviolet-visible spectrum, but little interpretation is possible from a single example. These compounds were generally obtained as red powders, or rarely, as red crystal platelets; they had a measured magnetic moment of approximately 1 B.M. This relatively high magnetic moment was probably due to the presence of impurities as difficulties were encountered in purifying these compounds, a similar problem was encountered with ferromagnetic impurities in the nickel ketazines which could only be removed by repeated recrystallizations or a chromatographic purification step. However, anomalous magnetic moments18 have been reported for nickel (II) complexes and this possibility could not be discounted. The coordination of the ligands to the nickel (II) ion is proposed to be as shown in Structure VI. The macrocyclic ligand contains thirteen members and is coordinated so as to form three five-membered and one six— membered chelate rings with the nickel (II) ion. The ligand donor atoms are situated in an approximately square planar configuration about the nickel (II) ion. Such coordination by l3-membered macrocycles is rare,2 but a recent X—ray structure determination on a nickel (II) 61 .MH ofiomo z Hz-a mo .mcmwamn an .mupommm Emflouzoflm umfisouflo .a .mflm ooo 0mm con omv 00V 0mm com own P p b p p p p! _d it D 3 9 {v .P I I! . S - «.0 T3 .3 . no 1 so A85 numcmamtwwz .m> mama/N. coapmuom r50 62 19 has shown that such complex of a lB—membered macrocycle coordination does exist in other compounds. The authors found slight tetrahedral distortions, but noted that similar distortions occur in macrocyclic complexes of larger ring size and in non-cyclic complexes. These authors have further suggested that square—planar coordination may be possible for lZ-membered macrocyclic complexes of nickel (II), based on the available data on Ni-N bond distances and N—N "bites." All of the previously reported synthetic l3—membered macrocyclic complexes have been derived from triethylenetetramine, so the synthetic route published in this dissertation represents a new route for the synthesis of these unusually small ring-sized macrocyclic ligands. Electron delocalization throughout the ring system is possible via the central metal ion and may stabilize the compounds. Models based on this structure show very little strain. By reacting NiMMK with optically active diamines it has been shown that optically active macrocycles can be synthesized. Non-Cyclic Ni-N30 Type Compounds A different type of complex is formed when NiMMK reacts with 1,3 diaminopropane at from room temperature to 100° or with ethylamine at 100° for several hours. These reactions result primarily in a Ni N30 type of 63 coordination, by displacement of only one of the oxygen atoms of the ketazine, as shown in Structure VII. The Structure VII formation of these compounds appears to indicate that the second oxygen atom of NiMMK is more difficult to dislodge than the first. Only a mono-ethylamine product has been detected, although the reaction took place under what would appear to be forcing conditions. This reaction occurred at loo-110°, about 90° above the boiling point of ethylamine, under several atmospheres of pressure in a sealed pressure tube and with precautions taken to exclude water from the reaction mixture. NiApSo is obtained in near quantitative yields from the reaction of 1,3 diaminopropane with NiMMK and can be recrystallized from common organic solvents. NiESo is obtained in low yield and has not as yet been isolated in a form pure enough for infrared or elemental analysis, but it appears that a purification can be 64 accomplished in the near future. Some of the infrared absorption bands of NiApSo have been presented in Table 6. This spectrum contains the intense v(C—N) band at 1295 cm-1 and the characteristic band of theSe condensation products at 700 cm-1; in many aspects this spectrum is similar to those of the nickel macrocyclic complexes. The primary method of characterization for these Ni N30 type compounds has been mass spectroscopy, and these compounds exhibited fragmentation patterns which were similar to those of the macrocyclic complexes. These compounds also displayed the stability of the five-membered chelate ring which was noted previously. Although NiApSo was the only compound of this group which did not exhibit an initial loss of 28 mass units, the majority of the data obtained from the mass spectra supports the proposed structure for these compounds, as shown in Structure VII. When 1,3 diaminopropane reacted with NiMMK at 100° in a pressure tube, a peak was seen in the mass spectrum of the product which corresponded to the macrocycle which could be formed in the reaction. However, the major component of the mixture obtained was NiApSo and an attempt to isolate this macrocycle by chromatographic techniques yielded only a fraction containing the non— cyclic NiApSo compound. Apparently the 1,3 diaminopropane macrocycle hydrolyzed on the water containing alumina column. Also a sample of NiESo was apparently hydrolyzed 65 to NiMMK in a similar way. Chromatographic purifications of NiESo have been accomplished, though, when the alumina and solvents used were treated to remove any moisture; a similar procedure might work for the 1,3 diaminopropane macrocycle. The reactions of ethylamine and 1,3 diaminopropane were temperature dependent. For 1,3 diaminopropane, a red oil was obtained from its reaction with NiMMK at the boiling point of the amine (~ 150°). Although this red oil was not characterized, it was probably a polymer. Jager6 has reported polymer formation in similar reactions with diamines, and a molecular framework model showed that the propyl chain was sufficiently long to enable the terminal amine group to react intermolecularly with a second nearby molecule of NiMMK. At loo-110°, about the boiling point of 1,2 diaminoethane and 1,2 diaminopropane, 1,3 diaminopropane reacted to form NiApSo and, what appeared to be, a small amount of a l4—membered macrocyclic complex. At ambient temperature, NiApSo appeared to be the only product of the reaction. For the case of ethyl- amine, NiESo was only obtained after heating the reactants to lOO-llO° for several hours. The product isolated when this reaction occurred at the boiling point of the amine, about 14°, or in a pressure tube at 60°, was the NiMMK starting material. In these condensation-reactions the temperature appeared to be a crucial factor in determining the course of these reactions. 66 The NiApSo compound contained a potentially pentadentate ligand. A five coordinate intermediate similar to this has been proposed by Green et al,8 but has not been isolated. The presence of the amino group was confirmed by the formation of a mono-p-toluenesulfonyl derivative. A similar reactivity was observed for nickel 20 where a terminal amino group, complexes of ornithine which was coordinated to the nickel ion, formed only a mono benzoyl derivative. A project is now underway to determine the structure of this compound in the solid state by X-ray crystallography. It is hoped that this structural analysis will confirm the five coordinate nature of this compound. After NiApSo was characterized and the temperature dependent nature of the reactivity of 1,3 diaminopropane with NiMMK was discovered, attempts were made to bring about a ring closure reaction to form a macrocyclic product. Heating NiApSo, at 100° for several hours under vacuum in a drying pistol containing phosphorus pentoxide, produced no reaction. The next idea tried was to dissolve NiApSo in a solvent and then heat the solution to reflux. It was hoped that under conditions of high dilution and elevated temperature the ring closure reaction might be induced to occur. It was thought that the applied heat might supply sufficient energy to overcome the apparent barrier to reaction at the second oxygen site, and under 67 conditions of high dilution the "template effect" might work to hold the terminal amine group in close enough proximity to allow a reaction to occur. The compound was soluble in alcohols, so it was decided to attempt the reaction in a series of alcohols of increasing boiling point to try to determine the height of the barrier to reaction. The first alcohol used was absolute ethanol (b.p. 78°) selected partially on the basis of its water absorbing abilities, but after refluxing a solution of NiApSo in ethanol for several hours, NiApSo was the only product isolated. The next alcohol tried was n—butanol (b.p. 114°). This appeared a likely candidate since its boiling point was close to that of 1,2 diaminoethane which was known to form macrocycles at reflux temperature. The red product obtained after refluxing NiApSo in n-butanol for several hours had a molecular weight of four less than NiApSo. The mass spectrum of this compound was quite different from that of NiApSo. This compound was apparently formed by a dehydrogenation of NiApSo with the resulting loss of two molecules of hydrogen. The nickel ketazines were routinely recrystallized from boiling n-butanol without undergoing any such change, which rules out dehydrogenation at the bridging carbon atom, so the loss probably occurred on the aminopropyl chain. Elemental analysis would probably not be sensitive enough to detect a loss of four hydrogens from a compound with a molecular 68 weight of 596 and which still retained 30 hydrogen atoms, so it was not attempted. The infrared spectrum showed changes in the N-H absorptions and showed bands which could be attributed to hydrogens attached to doubly-bonded carbon atoms. The available information, on the small sample obtained, suggested the following change in the amino propyl chain, from -II‘:IIi-CH2-CH2-CH2-NH2 of NiApSo, to -N -CH=CH-CH=NH of a dehydrogenated compound. Possible precedents for this type of a reaction might be found in the work of Bailar and co-workers* on platinum complexes which catalyzed selective partial hydrogenation of unsaturated fatty acids. A complete understanding of this unusual reaction awaits further investigation. As a result of this reaction it appears that NiApSo would rather undergo partial decomposition than to cyclize. A possible explanation of this behavior might be that the terminal nitrogen is coordinated relatively strongly to the nickel atom. This attachment could prevent the amine from reacting at the oxygen site. In the dehydrogenated compound, the terminal nitrogen could still be coordinated to the nickel atom, and the ring thus formed could be further stabilized by election delocalization over a pseudo conjugated system. *Private communication. 69 Reactivity_of Coordinated CO Groups The reactivity of the coordinated CO groups of NiMMK toward aliphatic amines is somewhat unusual for a carbon-oxygen-metal bond. Other authors6’ 7' 9’ 21' 22' 23 who have attempted to cyclize tetradentate cis-N202 type ligands with aliphatic amines have been unsuccessful in cases where a B (meso) carbonyl substituent was absent. Coordinated carbonyl groups have been shown to react readily with mono- or di-amine compounds, i.e., reactions of salicylaldehyde,l but these reactions normally stop short of macrocycle formation and that adds significance to the results of this project. For this reason the factors which appear to be involved in these reactions will be recounted here. Thus far these reactions have not been observed to take place in the presence of a solvent; they appear to occur best in dry amines, at room temperature for the diamines, but only at an elevated temperature for ethyl— amine. Another point is that the amines, especially the diamines, form very stable hydrated species which may serve to remove the water molecules formed in the conden— sation reactions from the reaction and perhaps to act as a Le Chatelier stress on an equilibrium which may occur. Also one oxygen atom of the ketazine appears to be more available for reaction than the other which is identical before the first amine condensation reaction; this 70 suggests that changing a Ni-O—C bonding system to a Ni-N-C system makes the second Ni-O-C bonding system stronger. Framework models show that the nickel ion in the nickel ketazines would have vacant axial sites available for possible further coordination and the preliminary studies of NiApSo suggest that an amine group is coordinated in one of these axial sites. However, attempts to form adducts with donor ligands such as pyridine have so far been unsuccessful. It seems that for 1,2 diaminoethane and 1,2 diaminopropane that the macrocyclic compounds are both kinetically and thermodynamically favored, as the same main product results at ambient or reflux temperatures. For 1,3 diaminopropane it seems that the products obtained are the kinetically favored products. NiApSo is formed at ambient temperature or 100°, but not at reflux temperature and the macrocycle which appeared to be formed at 100° hydrolyzed on an alumina column due only to adsorbed water on the alumina. Similarly for NiESo, it was only obtained from reactions at 100° and hydrolyzed readily to NiMMK due to adsorbed water on an alumina column. For NiESo this seems to indicate a large barrier to its formation, but a small barrier to decomposition. The NiESo compound potentially is a key to further under- standing of the mechanism(s) involved in these reactions as its ultraviolet-visible spectrum is intermediate to 71 those of NiMMK, Ni H cyclo l3, and NiApSo, a fact which may be useful to spectrosc0pic studies of the kinetics of these'reactions. Several attempts were made to incorporate o- phenylenediamine into a nickel cyclo 13 type of compound, however, all of these attempts were unsuccessful. This 6, 21 diamine had been used by Jager to synthesize macro— cyclic compounds from cis N202 compounds similar to the nickel ketazines. The reactivity of NiMMK towards ammonia, aniline, and 3,3' iminobispropylamine was also investigated. In all three cases for reactions at or below room temperature the starting material NiMMK was recovered from the reaction mixture. Only 3,3' iminobispropylamine was allowed to react with NiMMK at about 100° and the product of this reaction was a red oil which could not be crystallized nor characterized. This oil might have been a polymeric material as the carbon chain between terminal amine groups was relatively long and polymeric materials have been previously reported in similar cyclization reactions.21 Also in the course of this project, evidence was obtained which indicated that another product was formed in the room temperature reaction of NiMMK with 1,2 diamino- ethane. In the mass spectrum of an unrecrystallized product of one of these reactions, a peak was observed at 18 mass units above the parent ion. This peak probably *HN(CH2CH2CH2NH2)2 72 represented a compound intermediate between the nickel ketazines and the nickel macrocycles. Several possible structural explanations for this peak were examined of which the following three seemed most likely: one possibility was a compound similar to that formed in the reaction of NiMMK with 1,3 diaminopropane (zidg supra); a second was a compound in which one amine group had displaced an oxygen, but the other had only proceeded halfway in the displacement reaction, leaving O-H and N-H groups, possibly hydrogen bonded together, in this compound; a third possibility was that the peak represented a Ni H cyclo l3 molecule with a water molecule attached to it in some way. This last possibility was discounted because it seemed unlikely for a water molecule to remain attached for a sufficient length of time to be observed by the mass spectrometer at 250° and under the vacuum present in the instrument. The only other data obtained on this compound was that it was more soluble than Ni H cyclo 13, but this does not confirm any of the possibilities. Mechanism of the Reaction of NiMMK With Amines The ideas presented in this section are somewhat speculative, but an attempt has been made to link together the information which is known into a coherent and con- sistent pattern. There are other possible explanations, but only the ones which appear to be most likely will be discussed here. 73 The mechanism which appears to be most likely is as follows: (1) an amine group coordinates to the nickel ion in an axial site; (2) as a result of this the N-H bonds become weaker; and (3) a Ni-O bond may be weakened, but whether or not this occurs, the oxygen is close to the amine hydrogens and hydrogen transfer can occur giving an -N:H —O-H type of intermediate; (4) the -O—H group is now less firmly coordinated to the nickel and the -N;H may be more strongly coordinated; (5) now the second amine hydrogen is transferred to the O-H group, perhaps through a hydrogen bonded intermediate or through nickel ion catalysis, as a result a water molecule is eliminated and the nitrogen, now firmly bonded to the nickel ion moves in to replace the departing oxygen; (6) a similar process could then take place for the replacement of the second oxygen atom, but now assisted by the "template effect" which holds the reactive species in close proximity. For the 1,2 diamines it appears there would be some strain introduced into the molecule on coordination of the second amine group to the nickel ion and perhaps this compensates for the apparent greater strength of the second Ni-O bond. In the 1,3 diamine case, the chain length is such that little or no strain would be introduced and the coordination of the amine together with a stronger Ni—O bond could be strong enough to stop the reaction at this five coordinate Species. For ethylamine the same mechanism could apply, 74 .P §m_Z