A study was made to ascertain the mechanism by which ortho-positronium reacts with nitroaromatics in solution. These reactions are two to three orders of magnitude faster than expected by simple pickoff. Conjugation of the nitro group with the aromatic ring was found to be a necessary condition for this behavior. The rate constant for each of these reactions was measured in suitable solvents over a wide temperature range, 200-500°K. Distinct departure from the Arrhenius law was found. In general, the rate constant increased linearly with increasing temperature up to a point, and then decreased linearly as the temperature was raised further. The following mechanism is postulated: Ps + M [stacked right and left arrows with k₁ above and k₋₁ below] PsM [right arrow with k₂ above] M + 2γ. Ortho-Positronium reacts with the nitroaromatic, M, to form a complex in a reversible step. The complex may either decompose into a positronium atom and nitroaromatic molecule or it may go on to annihilate the positron. A steady-state concentration ot the complex is assumed and the observed rate constant for the process is K_obs = k₁k₂ /(k₋₁+ k₂) The two limiting cases are: (1) k₂ >> k₋₁ causing k_obs = k₁ . Typical Arrhenius behavior is expected and observed in this region, and (2) k₋₁ >> k₂ causing k_obs = K_EQk₂ . In this region H_EQ + E_a(2) <O and the observed rate constant decreases with increasing temperature. This corresponds to a stable but temperature sensitive PsM complex. An approximate molecular orbital study has been made to study the possible existence of positron and positronium complexes with a select number of organic molecules. The CND0/2 approximation is used for all electronic integrals and for electron positron Coulomb integrals. The core Hamiltonian matrix elements for the positron were estimated by combining the Wolfsberg-Helmholtz and Cusachs-Cusachs approximations. This approach shows that the positron should become bound to the molecules. Using the same criteria, the formation of stable positronium complexes is more improbable. The calculated binding energies are strongly dependent on the value of the Wolfsberg-Helmholtz proportionality constant for positrons. / Doctor of Philosophy

This thesis reports the preparation of new pentafluorosulfanylimino (SF₅N=) derivatives, which have been synthesized from pentafluorosulfanyl isocyanate, SF₅NCO, pentafluorosulfanylamine, SF₅NH₂, and pentafluorosulfanyliminosulfur difluoride, SF₅N=SF₂. Reactions of SF₅NCO, SF₅N=SF₂, and SF₅NH₂, with appropriate substrates have produced SFN=S(CH₃)₂, SF₅N=CHC₆H₅, SF₅N=SCl₂, (SF5N=)₂C, (SF5N=)₂S and SF₅N=PCl₃, some of which are new compounds and some of which represent improved routes to compounds previously reported. When N,N'-bis(pentafluorosulfanyl)urea, (SF₅NH)₂CO, reacted with carbonyl fluoride SF₅NCO was formed. The reactivity of SF₅NCO in several basic reaction types including nucleophilic substitution, addition, exchange and coupling reactions was examined. Pentafluorosulfanyl isocyanate and dimethyl sulfoxide reacted to produce a crystalline product identified by its H-1 and F-19 nmr and in spectra as SF₅N-S(CH₃)₂. Pentafluorosulfanyl isocyanate and benzaldehyde reacted to produce a yellow solution and carbon dioxide. The solution was determined by spectroscopic mean to contain SF₅N=CHC₆H₅, a novel, electron-deficient Schiff base. Pentafluorosulfanylisocyanate and PCl5 reacted readily at 60-80° producing SF₅N=CCl₂ and POCl₃. A slow reaction between SF₅NCO and excess AgF₂ took place at room temperature. Infrared analysis of the reaction mixture at increasing temperatures gave evidence for formation and subsequent decomposition ofSF₅N=NSF₅. Pentafluorosulfanyliminosulfur difluoride (obtained from irradiation of N≡SF₃) reacted at room temperature with PCl₅ to produce SF₅N=SCl₂, a pale yellow liquid which rapidly attacked mercury, and reacted with AgF₂ (producing SF₅N=SF₂) and with SF₅NH₂(producing SF₅N=S=NSF₅). Pentafluorosulfanylamine, SF₅NH₂, and PCl₅ reacted at room temperature to give SF₅N=PCl₃ a pale yellow liquid which reacted rapidly with mercury. Thiazyl trifluoride reacted with metal carbonyls (Ni(Co)₄ , Fe(CO)₅, Mo(CO)₆, Mn(CO)₆) and ferrocene to produce thiazyl trifluoride - transition metal complexes. Chemical similarities between SF₅NCO and SF₅N=SF₂ with respect to nucleophilic substitution, exchange and coupling reaction reflect the similarity in bonding expected in these two systems. However, the failure of SF₅N=SF₂ and compounds containing the -N=SF₂ group to undergo additional reactions with polar reagents indicates some gross discrepancy from the usual behavior of these multiply bonded systems. Thus, a theoretical study of -N=SF₂ system was undertaken to clarify the role and magnitude of d-orbitals in the bonding of -N=SF₂ compounds. Theoretical calculations (CNDO) of the total energy of SF₅N=SF₂ , CF3N=SF₂ , C₂ F₅N=SF₂ , and FCON=SF₂ as a function of rotation about the N-S(IV) multiple bond showed that each total energy curve possessed a broad, flat minimum. Rotamers derived with only slight excitation would possess equal energy. Therefore, this multiple bond is nonrigid in contrast to the usual concept of the pπ-pπ double bond. The d-orbital contribution accounts for approximately 50% of the total π-bonding and is practically independent of the nature of the substituents on the nitrogen atom. Low temperature F-19 nmr studies showed that splittings occurred which can be explained on the basis of the total energy curves derived from the calculations. / Doctor of Philosophy

The stereochemistry of the energetic ³⁸Cl for Cl substitution was studied in diastereomeric 2,4-dichloropentane and 2,3-dichlorobutane as well as in enantiomeric 2-chloro-1-propanol in solution. It was found that the stereochemical course of the substitution reaction exhibits a strong dependence on the concentration and nature of the solvent added when the diastereomeric compounds were used as substrate. Similar results with a lesser degree of solvent dependence were observed in the 2-chloro-1-propanol system. A conformational analysis of 2,4-dichloropentane was made. It showed that the relative conformer population is not affected to any significant degree by the nature of the solvent. Furthermore, an examination was made of the physical and chemical properties of the solvent which could be responsible for the drastic changes inflicted upon the stereochemistry of the ³⁸Cl for Cl substitution reaction. The experimental results indicate that the stereochemical course of the substitution process is predominantly and directly controlled by the properties of the solvent molecules, most likely by the dielectric constant and molecular size which in turn are responsible for the magnitude of intermolecular interaction between reactants and solvents. Although the experimental results seem to indicate that the radical-radical recombination mechanism is the most likely mechanism as compared to the hot one-step substitution process, they offer no conclusive distinction between these two mechanisms. It appears that strong intermolecular interaction favors substitution via retention of configuration. whereas in solvents having a low dielectric constant or large molecular diameter the retention/inversion ratio decreases. / Ph. D.

Sodium borohydride in aqueous dimethylsulfoxide has been shown to be a good reducing agent for converting certain activated aromatic halo-compounds to their corresponding dehalogenated products. The order for ease of removal of the halogen is I>Br>Cl. The activating groups are those that are strongly electron-withdrawing. The reactivity for activating aryl halides substituted with groups such as -NO₂, -CF₃, -F, -Cl, -Br, and -I is in the order ortho > meta > para, thus indicating that the activating effect has its origins in inductive rather than resonance effects. The removal of halogen appears to be occurring via a displacement on-halogen. This has been demonstrated in certain cases by using deuterium oxide instead of water in the reaction. When this is done, the halogen is replaced by deuterium instead of hydrogen. This indicates that the halogen leaves without its bonding electrons, thus leaving a carbanionic site on the aromatic ring. The carbanion is subsequently quenched by a proton from the water. This reaction has also been shown to be applicable to aromatic systems other than benzene. On-halogen type displacement by the hydride ion also occurs on certain polyhalogenated alkanes with the formation of a guasi-carbanionic intermediate which can be quenched by a proton from water, undergo alpha elimination, or undergo beta elimination. The alpha elimination occurs when a good leaving group is not present on the beta carbon. Attempts at trapping the carbene type intermediate from alpha eliminations were unsuccessful, evidently because of the presence of the water in the reaction medium. In two cases eliminations have occurred through a benzene system to generate para-xylylenes. Diazonium salts were also shown to undergo direct reduction with sodium borohydride, thus providing a new route for the deamination of aromatic amines. / Ph. D.

Ph. D.

The preparation of 2-(3-chloro-l-naphthylmethyl)bromobenzene was achieved by the cross-condensation reaction of 3-chloro-l-naphthylmagnesium bromide and 2-bromobenzyl bromide, as well as by the reaction of this Grignard reagent with 2-bromobenzaldehyde, followed by reduction of the resulting carbinol with lithium aluminum hydride and aluminum chloride. It was found that 2-bromophenyl-1-(3-chloronaphthyl)carbinol thermally decomposed into the corresponding methylene compound and ketone. A study of the thermally induced reaction of the carbinol was made, and the products were quantitatively analyzed by means of gas chromatography. It was concluded that the anomalous products of the reaction of an aryl Grignard reagent with a benzaldehyde were actually p~duced by the thennal disproportionation of the resulting carbinols during the distillation step. The keto-acid, 2-(3-chloro-l-naphthylmethyl)- 2’-carboxybenzophenone was prepared by the inverseaddition of the Grignard reagent of 2-(3-chloro-l-naphthylmethyl)bromobenzene to phthalic anhydride. Cyclization of this keto-acid with an acetic and hydrobromic acid mixture gave 6-chloro-7-(2-carboxyphenyl)benz[a]anthracene. Methyl ester derivatives were prepared from both this acid and the precursor keto-acid. The cyclodehydration of either 2-(3-chloro-l-naphthylmethyl)-2'-carboxybenzophenone or 6-chloro-7-(2-carboxyphenyl)benz[a]anthracene with polyphosphoric acid gave 14-chlorodibenzo[hi,l]chrysen-9-one.· Treatment of this ketone with lithium aluminum hydride and aluminum chloride gave the reduction derivative, 14-chloro-9H-dibenzo[hi,l]chrysene. The unequivocal synthesis of dibenzo[hi,l]- chrysen-9-one was achieved by the dehalogenation of 14-chlorodibenzo[hi,l]chrysen-9-one with 10% palladiumcharcoal catalyst and hydrazine. The dehalogenated product was shown to be identical to the compound produced from the cyclodehydration of 7-(2-carboxyphenyl)benz[a]anthracene. The ketone, 2-(3-chloro-l-naphthylmethyl)benzophenone was prepared by the inverse-addition of the Grignard reagent of 2-(3-chloro-l-naphthylmethyl)- bromobenzene to benzoyl chloride. It was found that a small amount (16%) of 6-chloro-7-phenylbenz[a]-anthracene was formed during the distillation of the precursor ketone. The cyclodehydration of this ketone failed when various standard cyclizing media were employed, and the reason for this is discussed. Cyclization attempts with polyphosphoric acid or alumina gave dibenzo[a,l]pyrene as the only identifiable product. This unusual reaction obviously involves a rearrangement. A study was made and a mechanism for this reaction was postulated which is consistent with the experimental observations. The ketone, 2-(3-cyano-l-naphthylmethyl)benzophenone was prepared by the reaction of the corresponding chloro ketone with cuprous cyanide in N-methylpyrrolidone. 6-Cyano-7-phenylbenz[a]- anthracene was also produced in small quantity in this reaction as a consequence of the presence of the corresponding chloro compound in the ketone prior to reaction. Naphtho[3, 2, l-fg]naphthacen-9-one was prepared by the treatment of 6-cyano-7-phenylbenz[a]anthracene with a hydrobromic and acetic acid mixture at 180°, and also by the treatment of the precursor cyano ketone with polyphosphoric acid. The novel use of polyphosphoric acid in cyano group hydrolysis is discussed. Phenalo[2, 3, 4, 5-defg]naphthacene-4, 8-quinone was prepared by the treatment of 6-cyano-7-(2-carboxyphenyl)benz[a]anthracene with a hydrobromic and acetic acid mixture at 180°. An attempted procedure for the. preparation of this quinone involved the oxidation of 7-(2,6-dimethylphenyl)benz[a]anthracene to the corresponding diacid with aqueous sodium dichromate. Unfortunately this new method of oxidation failed in this case. The partial resolution of 7-(2-carboxyphenyl)-benz[a]anthracene was achieved with the use of brucine. Only one optically active isomer was obtained, and this was racemized by treatment with boiling ethanol. An empirical rule used to quantitatively determine the resistance of optically active biphenyls to racemization was applied to this acid, and the experimental observations were supported. During the course of this investigation, sixteen new compounds were prepared and were all properly characterized, except 6-cyano-7-(2-carboxyphenyl)- benz[a]anthracene, which did not give acceptable analytical data. The reason for this is discussed. Infrared and ultraviolet spectra of all new compounds were recorded. Infrared spectral observations were made which gave further support to the assigned structures of the isomeric compounds naphtho[3,2,l-fg]- naphthacen-9-one and dibenzo[hi,l]chrysen-9-one. / Doctor of Philosophy

The chemical adsorption and reaction of methyl fluoride, methyl chloride, methyl bromide, methyl iodide, methyl sulfide, dimethyl sulfide, dimethyl disulfide, and bis(trifluoromethyl) disulfide on clean, polycrystalline nickel surface (at 25°C) were investigated using an X-ray photoelectron spectrometer (XPS) and a thermal desorption system with mass spectrometric analysis (TD). The ion implantation of the methyl halide series on nickel was investigated by XPS. The nickel surface was cleaned by high temperature heating in vacuum or argon ion sputtering. An industrial lubricant, FC-43 (tri-hepta-fluorobutyl amine), was adsorbed on a gold-plated nickel surface at 25°C. Tne FC-43 adsorption system was investigated by TD. The adsorption of the methyl halide series on nickel was predominantly dissociative with several co-existing surface species being observed. The XPS results show that two carbon and two halogen species are formed. The possible structures for carbon are a CH₃-type and a CH₂-type or CH_nX (n = 0-3) adsorbed species. The halogen species appear to be neutral and ionic species. The TO results support the dissociative adsorption process for methyl halides on nickel. The adsorption of methyl fluoride and methyl chloride on nickel studied by TD resulted in the corresponding nickel halide being detected. The adsorption of methyl bromide and methyl iodide on nickel studied by TO resulted in the corresponding hydrogen halide, HX, being detected. The XPS results of the ion implantation of the methyl halides resulted in only one type of carbon that is similar to graphite. A halide type of halogen was also formed upon ion implantation. The XPS results for the adsorption of methyl sulfide, dimethyl sulfide, and dimethyl disulfide on nickel showed that only one carbon and one sulfur were observed. The possible structure for carbon is a CH₃-type of adsorbed specie. lhe sulfur species appears to be a sulfide S⁼-type adsorbed specie. The XPS results for the adsorption of bis(trifluoromethyl) disulfide on nickel showed that only sulfur was observed on the nickel surface. The TD results- for the methyl sulfide adsorption process was shown to be predominantly an associative type. The TO results for the dimethyl sulfide adsorption process were a mixture of dissociative and associative adsorption processes. The dimethyl disulfide adsorption process was determined by TD to be predominantly a dissociative type. The bis(trifluoromethyl} disulfide was determined by TD to be completely dissociative upon adsorption and reacted with nickel to form nickel sulfide. The TD results of FC-43 adsorption on gold-plated nickel showed that the predominant adsorption process was the associative type. / Ph. D.

Previous investigations have revealed that some polymer solutions exhibit dielectric dispersion (variation of the dielectric constant with frequency) which is apparently related in some way to the molecular weight of the polymer (1) (2) (3). One of these investigations (1) was carried out on solutions to cellulose acetate in dioxane. It was found that the change in dielectric constant was so small that rather concentrated solutions had to be used in order to obtain reproducible results. The possibility of intermolecular interactions in concentrated solutions made dilute solution measurements desirable. Consequently, the object of the present investigation was to carry out such measurements with sufficient precision to define the shape of the dielectric dispersion curve over the necessary frequency range. The fundamental difficulty was that the change in dielectric constant of a solution containing less than 1% cellulose acetate was only about one part in five hundred or less in a frequency range of about three decades. The course ot such a small change is difficult to detect over such a wide band of frequencies. Capacitance bridges; which are very good for making measurements at a single frequency, particularly in the range of frequencies under investigation; were commercially available. However, bridges have certain inherent disadvantages which make difficult the coverage of such a wide band of frequencies with the necessary precision. The use of ordinary resonance methods was desirable, but they are limited by the practical sizes of the components to frequencies well above the lowest frequencies which were expected to be encountered. As a result of these limitations, a modified resonance procedure was adopted. A resistance-capacitance oscillator was substituted for the more conventional inductance-capacitance circuits. Null indications were obtained by substituting a frequency counting system of great precision for the "zero beat" method. It had been found earlier that the critical frequency was related to the weight average molecular weight of the polymer sample, the critical frequency being defined as the frequency at which the dispersion was 0.5. Based on this it seemed reasonable to assume that the dielectric dispersion curve might be related in some way to the molecular weight distribution curve in the case of a polydisperse sample. An attempt was thus made to obtain the dielectric dispersion curve in as much detail as possible, since this served the two-fold purpose of permitting conclusions to be drawn concerning the validity of the molecular weight relationship in dilute solutions and giving an indication of any influence of the molecular weight distribution on the dielectric dispersion. Apparatus was developed which gave a precision of 0.01 uuf in a total of about 250 uuf. This provided the means of measurement. Measurements were made on several fractions and three artificial blends. The results of measurements on fractions indicated that the form of the critical frequency-molecular weight relationship determined earlier (1) was correct. The dispersion curves obtained for the three blends of varying width indicated that the slope of the dispersion curve at the critical frequency decreased as the width of the blend was increased. This was the only indication of an influence of the width of the distribution on the dispersion curve. It was noted, however, that the slopes of the dispersion curves for the fractions were less than those for the narrowest blend in several cases. The most reasonable explanation of this observation seems to be that the fractions and blends have much wider distributions than was believed. It is also possible that chain flexibility might be a complicating factor. (1) Scherer, P. P., Levi, D. W., and Hawkins, M. C.: J. Polymer Sci., 24, 19 (1957) (2) Scherer, P. C., Hawkins, M.C., and Levi, D. W.: J. Polymer Sci., 31, 105 (1958) (3) Scherer, P. C., Hawkins, M. C., and Levi, D. W.: J. Polymer Sci., 37, 369 (1959) / Ph. D.

Physisorption of argon, krypton, and nitrogen on Pyrex and iron was studied by a static technique in a constant volume system, in the temperature range 77.4 to 90.2°K and in the pressure range 10⁻⁹ to 10⁻⁴ torr. Analyses via Electron Spectroscopy for Chemical Analysis (ESCA), Scanning Electron Microscopy (SEM), and Neutron Activation Analysis (NAA) were made to characterize the solid sample surfaces. Both Pyrex and iron surfaces were degreased with ethylene dichloride. The iron surface was further cleaned using the Diversey process. Both surfaces were given a mild bake in vacuum. A modified Dubinin-Radushkevich-Polanyi (D-R-P) theory was used for interpreting the experimental results. This theory assumes that the adsorbed phase is not necessarily liquid-like and is a function of adsorption temperature. Comparisons were made between the modified and the original theories and it was found that the modified theory had a wider applicability and gave better data correlation for all five systems in the present study. It was found that the Dubinin-Radushkevich (D-R) parameter B, and the limiting (maximum) adsorption energy, (π/4)^½ (B), ^-½ of all systems were primarily dependent on the adsorbate. A correlation between the limiting mean adsorption energy and the polarizability of adsorbate was made. From the D-R parameter, N_m, it was found that the effect of the Diversity process on iron was to increase the surface area. Isotherms of the argon/Pyrex system showed that Henry's Law behavior was reached at low coverages and high temperatures. Isosteric heats of adsorption were calculated for all systems. Isosteric heat versus coverage curves indicated that the Pyrex surface was heterogeneous and the iron surface was less heterogeneous. The magnitude of isosteric heats also indicated that all three gases were physisorbed on the Pyrex and iron. Work function measurements were taken simultaneously with the adsorption isotherm measurements. The results were marginal but confirmed that nitrogen was physisorbed on iron. Water contact angles on Pyrex and iron surfaces were also measured. It was found that the contact angle was very sensitive to the surface contamination and that ordinary laboratory air may contaminate the sample surface very rapidly. A finite water contact angle on Pyrex was found instead of a zero contact angle as is generally used in many works. / Ph. D.

A balanced density slurry-packing apparatus was developed and used to pack sixteen HPLC columns having different lengths and internal diameters with 10 µm Lichrosorb Sl-100. The evaluation of these columns led to several conclusions: the column efficiency, expressed in plates per meter, is a function of the ratio of column internal diameter to column length and increases as this ratio increases; a new definition for preparative efficiency, Time Yield Factor (TYF), is proposed. This factor differentiates among HPLC columns with different lengths and internal diameters using samples of the same volume and concentration; the column that showed the highest TYF value was used for the preparative separation and identification of three positional isomers of high molecular weight (634): 3,4-di(p-bromophenyl)-1,2,5-triphenylcyclo-2,4-pentadien-1-methyl ether; 2,3-di(p-bromophenyl )-1,4,5-triphenylcyclo-2,4-pentadien-1-methyl ether; and 1,2-di(p-bromophenyl )-3,4,5-triphenylcyclo-2,4-pentadien-1-methyl ether, obtained in the synthesis work. / Ph. D.