Volume 23, Issue 8, 01 August 1955
Index of content:
23(1955); http://dx.doi.org/10.1063/1.1742312View Description Hide Description
Reaction of O‐atoms, produced by mercury photosensitized decomposition of nitrous oxide, with ethylene has been studied at room temperature and 123°C. The primary step of the process appears to be a direct addition of the atom to the double bond to form an energy‐rich intermediate which undergoes further reactions. The following products are formed: aldehydes (CH3CHO, C2H5CHO, and C3H7CHO), CO, CH4, C2H6, C3H8, C4H10, H2, very small amounts of some other substances and traces of HCHO. At 24°C O‐atoms react with ethylene 22±5 times as readily as with n‐butane. Assuming comparable pre‐exponential factors for the two processes ΔE= — 1.9 kcal/mole, so that the activation energy of addition of oxygen atoms to the double bond in ethylene is estimated at close to 3 kcal/mole. In the presence of molecular oxygen the character of the reaction is profoundly changed.
Use of Molecular Quantum‐Mechanical Approximations Exemplified in the Energy Calculation of the H3 — System23(1955); http://dx.doi.org/10.1063/1.1742313View Description Hide Description
Certain approximations of molecular quantum mechanics are examined. The size of the resultant error due to approximate methods for specific multicenter integral evaluations are cited. Examination of the Sklar approximate multicenter electron repulsion formula together with the generalizations and extensions of Vroeland and Löwdin allow application of the underlying approximation principle to three center nuclear attraction integrals. Specific approximate formulation of these integrals is given.
The accumulation of computational errors arising from the many individual integral approximations in the energy evaluation of the linear H3 — system is specifically examined. The ground state energy of this system found from the Heitler‐London and Wang‐like treatments were — 70 kcal/mole and — 84 kcal/mole for the Sklar type approximations, — 1 kcal/mole and — 19 kcal/mole for the Mulliken type approximations, and — 26 kcal/mole and — 41 kcal/mole for exact integral evaluations.
It is concluded that the Sklar type approximations, while preferable to the Mulliken type in a few specific evaluations lead like that formulation to considerable discrepancy between the over‐all energy obtained for a polyatomic system. The Sklar approximations have the further disadvantage of requiring two‐center integral evaluations not needed in the over‐all energy calculations. However, both the Sklar and the Mulliken approximations do give a rapid method for evaluating ``best'' distance and effective charge parameters for use with more extensive exact integral variational method calculations.
23(1955); http://dx.doi.org/10.1063/1.1742314View Description Hide Description
Azulene was found to fluoresce with a frequency corresponding to transition from the second excited singlet level to the ground state. No phosphorescence was observed. No emission corresponding to transition from the lowest excited singlet to the ground state was observed. These phenomena are discussed.
23(1955); http://dx.doi.org/10.1063/1.1742315View Description Hide Description
The experimental quenching cross sections of H2, HCl, CO2, and H2O for the sodiumD lines arising from the optical dissociation of the sodium iodide molecule have been determined by measurements with a photomultiplier tube. The velocities of the excited sodium atoms after dissociation of the sodium iodide molecules by ultraviolet light of different wavelengths have been determined by using an approximation for the repulsive potential curve of an excited sodium atom and normal iodine atom obtained from a measurement of the excitation of the sodiumD lines as a function of exciting ultraviolet wavelength. The variation of quenching cross section for HCl and CO2 with relative velocities of excited sodium atom and quenching molecules shows a decrease of cross section with increase in relative velocity. The cross section for H2 shows little change with relative velocity in the range of velocities used. H2O shows little quenching action. The source of ultraviolet light was a hydrogen discharge lamp and the wavelength of the exciting ultraviolet light in the 2100 A—2500 A range was controlled by a reflection gratingmonochromator.
Shape of the Consolute Curve and its Consequences for the Theory of Binary Liquid Mixtures of Nonelectrolytes23(1955); http://dx.doi.org/10.1063/1.1742316View Description Hide Description
The experimental results on the shape of the consolute curve of binary liquid mixtures with components A and B, especially the cubic dependence of the concentration on temperature, imply certain peculiarities in the behavior of the excess free energy of mixing ΔG exc. The term ΔG exc is expressed as W · x(1 — x) ·Ψ (x). Here 2W can be considered as the change in free energy, times Avogadro's number, when z contacts A—A and z contacts B—B are changed into 2z contacts A—B; x is the molefraction of one component, and z the coordination number. Moreover, Ψ(x) is much more sensitive than ΔG exc, and has a theoretical significance, being proportional to the ratio of the number of A—B contacts in the system investigated to their number in a perfect solution if the assumptions of the strictly regular model are regarded as valid.
With the assumption that the system is symmetrical, the consolute curve consists of the points at which the first derivative of ΔGm with respect to x vanishes. If the function Ψ is expressed as a power series in s=2x—1, one also obtains Tc—T, Tc being the critical temperature, as a power series in s. The experimental results on the consolute curves serve now to determine the coefficients of the power series, by which Ψ is given. Two additional assumptions are needed: (1) The coefficients of the high powers vanish in order to make use of the relation Ψ(1) = 1 and (2) the value of ΔG exc for T = Tc and x=0.5 is known approximately. With these assumptions the experimentally established flatness of the consolute curves leads to a curve for Ψ vs x having a double‐maximum Ψ being greater than unity for high values of s, i.e., for great dilution of one component. This confirms earlier conclusions drawn from vapor pressuremeasurements and investigations of the volume change on mixing.
From a molecular point of view one can express this result by saying that the solvation of the diluted component is increased as compared to a perfect solution, even if the mixture is so strongly endothermic as to separate into two phases. The theoretical interpretation is not yet clear.
23(1955); http://dx.doi.org/10.1063/1.1742317View Description Hide Description
The wave equation is derived from two completely asymmetric molecules undergoing mutual rotation about a common axis and subject to an arbitrary hindering potential. Matrix elements of the Hamiltonian are evaluated using the wave functions which describe the mutual rotation of two symmetric groups.
23(1955); http://dx.doi.org/10.1063/1.1742318View Description Hide Description
The study of dielectric relaxation in compressed gases requires an extension of existing theory. As the gas pressure is lowered below several hundred atmospheres, the inertial response of the dipoles gives rise to large deviations from the Debye equations. The absorption and dispersion depend on collision time and frequency of applied field in a manner which is sensitive to the molecular constants and to the state of dynamical order of the compressed gas.
These phenomena are investigated by studying the response to an alternating electric field of a dilute solution of dipolar molecules in a nonpolar compressed gas. The dipoles are described by a classical distribution function which is a function of angular velocity, orientation, and time. We assume that the duration of collision may be neglected compared to the time between collisions, the period of the applied field, and the mean thermal period. The distribution function satisfies a kinetic equation; the effects of collisions are described by a collision kernel. One kernel yields a soft impact theory which is a generalization of Debye's Brownian motion treatment to include inertial effects. The complex polarization is given in the form of a convergent infinite continued fraction. Other models studied involve strong collisions and partial specular reflection.
A common feature of all models is that the Debye relaxation shape is found at high pressures. Inertial corrections are important in the region between one and a few hundred atmospheres; the corrections depend on the model. Another common feature is the way in which the discrete rotational lines are linked with the Debye spectrum. The collision frequency 1/τ increases with increasing gas pressure; at low pressures it is proportional to the width of a rotation line. At high pressures when the Debye shape is found, the relaxation timet * is given by t * = 1/τ(I/kT). Here I is the moment of inertia of the dipole, k is Boltzmann's constant, and T is the temperature of the reservoir. Thus the relaxation time varies inversely with time between collisions. The physical reason is that collisions hinder the drift motion of the dipoles which is the cause of the relaxation of the polarization.
At lower pressures the main contribution to the polarization arises from dipoles rotating with an angular frequency close to that of the applied field. This is related to the known result that the main contribution to the static polarization is from librating dipoles.
23(1955); http://dx.doi.org/10.1063/1.1742319View Description Hide Description
The strong absorption of iodine dissolved in paraffin hydrocarbons in the region of 2200 A—2600 A is not shown either by gaseous iodine or by iodine dissolved in perfluoroheptane (PFH) or di‐perfluorohexyl ether, and is present only to a much smaller extent for iodine dissolved in chloroform.
The interaction of iodine with n‐heptane has been studied quantitatively in PFH solution. It is concluded that the above absorption is a charge transfer transition due to a very weak, or possibly collisional type, complex, and is not a solvent shift of the N – V transition of the iodine molecule.
23(1955); http://dx.doi.org/10.1063/1.1742320View Description Hide Description
The effect of foreign gases on the continuous absorption of bromine in the region 2200 A—5100 A has been studied. Contrary to the results of Bayliss and Rees, no increases in intensity were observed. The absorption of bromine in various inert solvents has also been measured. The results are similar to those obtained with iodine and have been interpreted in an analogous manner. No maxima were found at 2725 A and 2500 A for bromine dissolved in chloroform and cyclohexane, respectively, in disagreement with Aikin, Bayliss, and Rees.
23(1955); http://dx.doi.org/10.1063/1.1742321View Description Hide Description
Solvent shifts of ultraviolet absorption bands (relative to the gas phase) have been shown to be much less for perfluoroheptane than for n‐heptane. The oscillator strength of the 2400 A band of styrene has been measured in the gas phase and in perfluoroheptane and n‐heptane solution. Small increases were observed in solution, which are, however, less than those calculated theoretically.
23(1955); http://dx.doi.org/10.1063/1.1742322View Description Hide Description
The magnetic susceptibility of graphite bromide residue compounds has been measured by a Faraday method. The addition of bromine decreased the diamagnetism of graphite. Fair agreement was obtained between the measureddiamagnetism and values calculated by assuming an undistorted graphiteband structure depleted in electrons due to the dissociation and ionization of two out of eleven bromine molecules.
It has also been shown that the transfer of electrons from the graphite bands to bromine can account for the observed pressure dependence of the graphite bromine equilibrium.
23(1955); http://dx.doi.org/10.1063/1.1742323View Description Hide Description
A new technique for measurement of the attachment coefficient of electrons to molecules has been devised. The technique involves production of a beam of electrons in an attaching gas. The transient current flowing after the initiation and termination of the electron beam is interpreted as a flow of negative ions produced by electron‐molecule collisions. Analysis of the transient current which may last several milliseconds yields a value for the mobility of the negative ions thus formed. The ratio of the magnitude of the electron beam current to the negative ion current yields a direct measure of the attachment coefficient.
Preliminary results obtained with SF6 show a very large attachment coefficient. Values of α/p (the ratio of attachment coefficient to pressure) are considerably larger in SF6 than those obtained for O2 in studies elsewhere. As might be expected the mobility of negative ions formed in SF6 is much lower than that observed for other gases. This is attributed to the very large collision cross section resulting from (1) the interaction between the negative ion and neutral gas molecules comprising polarization attraction and hard sphere repulsion and (2) resonance or exchange interaction.
Intramolecular Charge Transfer Spectra Observed with Some Compounds Containing the Nitro or the Carbonyl Group23(1955); http://dx.doi.org/10.1063/1.1742324View Description Hide Description
A new method, which proved to be useful for the theoretical study of electron migration for the monosubstituted benzene molecule, is employed to interpret strong absorption bands observed with various unsaturated compounds containing the nitro or the carbonyl group. Consequently, it is shown that these absorptions correspond to the transition between two energy levels produced by the interaction of the highest occupied level of the electron donating group with the lowest vacant one of the electron accepting group, and that theoretical values of transition energies and oscillator strengths evaluated by use of the present method are in very good agreement with the observed values. Furthermore, the present interpretation of these absorptions is discussed in comparison with the earlier resonance viewpoint.
23(1955); http://dx.doi.org/10.1063/1.1742325View Description Hide Description
Measurements of the ionic conductance of several metallic azides have shown an unusually low activation energy for the structure sensitive conductance in potassium,calcium, and strontium azides. This is here associated with the mobility of surface lattice defects.
Perturbation Method for the Calculation of Molecular Vibration Frequencies. II. Generalization of the Theory23(1955); http://dx.doi.org/10.1063/1.1742326View Description Hide Description
The perturbation theory developed previously for calculating frequencies by Wilson's F and Gmatrix method is now generalized in two respects: the effects, on the one hand, of degeneracy amongst the approximate frequencies and, on the other, of interaction constants coupling the two classes of vibration are considered.
Perturbation Method for the Calculation of Molecular Vibration Frequencies. III. Skeletal Stretching Vibrations of Normal Paraffins23(1955); http://dx.doi.org/10.1063/1.1742327View Description Hide Description
The perturbation theory developed in a previous paper is applied to the skeletal vibrations of the zigzag configuration of the n‐paraffins. By treating the CCC bending force constant as small compared with the CC stretching force constant the first‐order correction to the stretching frequencies is obtained in closed form. It is compared with the corresponding term in the Kirkwood‐Pitzer formula for the frequencies of long‐chain paraffins.
Hydrodynamic Frictional and Diffusion Coefficients in Idealized Liquids. I. The Collisional Contribution23(1955); http://dx.doi.org/10.1063/1.1742328View Description Hide Description
The approximate method of calculation of the collisional contribution to transport properties in liquids presented in an earlier paper is now extended to the calculation of the similar contribution to the frictional and diffusion coefficients. For the special case where the diffusing molecule is large compared to its neighbors, the calculated collisional contribution to the frictional coefficient is in good agreement with Stokes' law and the Langevin equation of colloid statistics is valid. However, for small diffusing molecules, the calculated frictional coefficient is not independent of the velocity of the given molecule. Numerical diffusion coefficients using molecular parameters obtained from sonic velocity data are about three times the experimental values as might be expected from the corresponding behavior of the calculated viscosity coefficient contribution previously reported. In spite of the failure of the Stokes equation for the frictional coefficient of small molecules, the Stokes‐Einstein equation represents fairly satisfactorily the case of self‐diffusion. This is due to the averaging over all velocities of the diffusing molecule.
Heat of Formation and Entropy of Titanium Tetrachloride from an Investigation of the Equilibrium: TiO2(s)+4HCl(g) = TiCl4(g)+2H3O(g)23(1955); http://dx.doi.org/10.1063/1.1742329View Description Hide Description
The equilibrium, TiO2(rutile)+4HCl=TiCl4(g)+2H2O(g), was studied in the temperature range of 800—1400°K by means of a flow method in which HCl was passed over a sample of rutile. The heat of reaction was found to be 12.9 kcal/mole in this temperature range. This leads to a value of —182.9±0.5 kcal/mole for ΔHf of TiCl4(g) and —192.5±0.5 kcal/mole for ΔHf of TiCl4(l). The entropy of TiCl4(g) at 1100°K is found to be 124 cal/°/mole.
23(1955); http://dx.doi.org/10.1063/1.1742330View Description Hide Description
The monohydrates of the hydrogen halides exist as ionic crystals. The infrared spectra of OH3F, OH3Cl, OH3Br, OH3I, and OD3Cl at —195°C showed the four fundamental frequencies expected for a symmetrical pyramid at about 1100 cm—1 (ν2), 1700 cm—1 (ν4), 2600 cm—1 (ν1 or ν3), and 3200 cm—1 (ν1 or ν3). The high‐frequency bands, which are very broad at —195°C, did not sharpen on cooling to 8°K, indicating that their breadth is caused by some frozen‐in disorder.
23(1955); http://dx.doi.org/10.1063/1.1742331View Description Hide Description
It is well known that the energy intervals separating ionic and covalent states, as given by the conventional ASMO method, are too large. In order to avoid this difficulty, a modification of the nonempirical atomic orbital method is proposed, and is applied to the ethylene molecule. The principal point of the modification is to use different atomic orbitals for covalent and ionic structures of the molecules in order to assess more correctly the energies of the ionic structures. The results of the calculation are fairly satisfactory in that they are as good as those of the semiempirical method by Moffitt. Some errors due to assumptions and approximations made in this calculation are discussed.