Index of content:
Volume 44, Issue 9, 01 May 1966
44(1966); http://dx.doi.org/10.1063/1.1727211View Description Hide Description
The effects of state, temperature, and chemically reactive solutes on the radiolysis of n‐hexadecane in condensed phases have been examined. All product yields are reduced by radical scavengers in liquid‐staten‐hexadecane, but only dotriacontane isomers are reduced in the solid. Iodine acts as a chemical scavenger and is more effective than 2‐methylpentene−1. A hot‐hydrogen‐atom mechanism, together with molecular elimination, accounts for the formation of most of the hydrogen in the irradiation of n‐hexadecane. Low‐molecular‐weight products are formed entirely by molecular reactions in the solid‐state radiolysis and, about equally, via molecular and radical processes in the liquid. Intermediate‐molecular‐weight products in the liquid radiolysis are 80% from combinations of radicals from chain scission with various hexadecyl radicals. The remaining 20% in the liquid and the total intermediate yield in the solid are from nonradical processes. The combination of hexadecyl radicals yields at least 40% and 70% of the dotriacontane isomers in the radiolysis of pure n‐hexadecane in the solid and liquid state, respectively. The remaining dimer may also be due to radicals which are formed in pairs, but ion—molecule reactions cannot be excluded. Temperature dependence of all products is small. Cage recombination of scission fragments and crosslinking of proximate radical pairs is enhanced in the solid state as compared with the liquid. Internal ···CH2–CH2··· rupture is equally probable at all positions, and seven times as likely as terminal CH3–CH2··· scission. Methyl C–H rupture is 40% as probable per bond as methylene C–H rupture.
44(1966); http://dx.doi.org/10.1063/1.1727212View Description Hide Description
Dilute solutions of several benzene derivatives in 3‐methylpentane solvent display three types of photocurrent when examined in their rigid, amorphous state at 77°K. There is the primary steady‐state photocurrent seen only when illuminating in the absorption region of the solute molecule. Following ionization in this ``primary'' spectral region (and never before) two new induced signals are seen. One is characterized by a photocurrent spike which is excited in the near infrared or higher energies. It is a transient signal which, once destroyed at any wavelength, can be seen again only after further primary irradiation. The second induced signal appears to have a threshold in the near‐ultraviolet region but at energies definitely below the onset of solute absorption. This signal is also basically transient. However, its decay is so slow compared to the decay of the spike signal, that it appears rather as a pseudo steady‐state photocurrent. It is found that oxygen exerts its influence primarily on the spike signal with a strong quenching effect. A volt—ampere study is reported. At the higher fields Poole's law is seen to hold. Ohm's law holds at lower fields. The parameter characterizing Poole's law is shown to be independent of solute and wavelength of primary excitation. It rather reflects any given preparation of the rigid solid. The Poole's‐law behavior can be removed by bathing the sample with near‐infrared light during the volt—ampere studies. The Poole's‐law behavior is also destroyed when the viscosity of the solid, at 77°K, is reduced by use of a two‐component solvent. A model is discussed which succeeds in rationalizing these observations. Basic to the model is that (1) the charge carrier is the electron, (2) the solid matrix contains sites (neutral) which trap electrons so firmly that their thermal mobilization is negligible, and (3) the photoionized system contains, homogeneously dispersed, the long‐range attractive Coulomb wells of the immobile cations—products of the photoionization, and (4) there is a very short‐lived mobile state of the electron. To explain the volt—ampere behavior and viscosity effects, a substructure is assigned to the mobile state of the electrons. It is proposed that this state consists of electrons trapped somewhat deeper than kT which are thermally activated into a condition of ``maximum mobility'' — the closest condition to a conduction band in this amorphous solid. Sufficiently intense fields aid in the activation of these electrons to the state of maximum mobility, while reduced viscosity permits thermal saturation of this state by diminishing the shallow trap depths.
Photoconductivity in Rigid Organic Solutions. II. Initial‐Rate Studies of Charge‐Carrier Production and the Two‐Photon Requirement44(1966); http://dx.doi.org/10.1063/1.1727213View Description Hide Description
The initial rises of photocurrents in certain rigid organic solutions are examined both as a function of light intensity and wavelength. The quadratic nature of the intensity dependence is taken to signify a two‐photon mechanism for charge‐carrier production. The initial rate of charge‐carrier production is linear in solute concentration and is not sensitive to O2 levels. Double‐beam experiments allow the isolation of each of the two one‐photon steps and the wavelength dependence of each step is studied. The first photon step shows specific vibronic activity for transfer into the intermediate state. The relative quantum yield rises sharply over the first absorption band of the solute molecule and then abruptly drops as the second absorption band is entered. The second step is relatively wavelength independent once its near‐ultraviolet threshold is reached. A dark lifetime of 4.5 sec is observed for the intermediate state. The triplet state of the solute molecule cannot be implicated. Additional initial‐rise studies reveal that of the two induced photocurrent signals one, the so‐called ``365'' pseudostationary signal, arises from a state of the system which is one photon produced and one photon ionized; the other, referred to as the spike signal, arises from a state which is two photon produced and one photon ionized. The general problem of creating full charge separation in low‐dielectric‐constant media is briefly reviewed to show how the two‐photon requirement is entirely to be expected. The results are treated within the framework of a straightforward kinetic scheme. Possible mechanisms, including a proton‐transfer intermediate, are briefly but inconclusively examined.
44(1966); http://dx.doi.org/10.1063/1.1727214View Description Hide Description
The generalized Breit—Pauli Hamiltonian is used to give a systematic treatment of magnetic and other relativistic intermolecular energies through O(α2) (where α is the fine‐structure constant) for intermolecular separations, R, sufficiently large that the charge distributions of the two molecules do not overlap, but sufficiently small that R<λ/0=(αΔε)−1, where Δε is the excitation energy of the first allowed transition of one of the molecules.
The theory is discussed in general and many types of interactionenergies are obtained which depend on the spin and orbital angular‐momentum states of the molecules. The interaction of two nondegenerate atoms (L=0, S=0) is considered specifically. Of particular interest is an interaction‐energy term which varies as α2/R 4.
44(1966); http://dx.doi.org/10.1063/1.1727215View Description Hide Description
The Casimir and Polder retarded dipole—dipole energy of interaction between two ground‐state (non‐degenerate) atoms is expressed in terms of sine and cosine integrals. This result should be accurate for all interatomic separations R. In the range of moderately large separations (discussed in the preceding paper), where the charge distributions do not overlap and where R is small compared to λ/0=(αΔε)−1, the Casimir and Polder results can be expanded in the formThis expansion is only accurate for R/λ/0<0.6. Here α is the fine‐structure constant. The R −6 C 6 term is the usual London dispersion energy. The α2 R −4 W 4 term was obtained in the preceding paper by taking the expectation value of the Breit—Pauli Hamiltonian using the wavefunction for the two‐atom system corrected for the classical electrostatic dipole—dipole interactions. Thus, at least in the dipole—dipole approximation, the Breit—Pauli Hamiltonian gives the energy of interaction accurate through O(α2). For large values of R/λ/0, the interaction energy is expanded in the seriesThis large R expansion is only accurate for R/λ/0>5.0.
44(1966); http://dx.doi.org/10.1063/1.1727216View Description Hide Description
High‐resolution spectra were taken of the 1–0 absorption band of HCl trapped in argon, krypton, and xenon matrices in the temperature range between 6° and 50°K. The influence of impurities on such spectra was studied by introducing small quantities of a different noble gas into the matrices.
A new spectral line, believed to be due to the combination of a ΔJ=0 transition with the constrained translational mode, is reported. Absolute intensities of the 1–0 band were measured and found to be 15 000, 18 500, and 18 000 darks in argon, krypton, and xenon, respectively. The spectra of HCl polymers are discussed.
44(1966); http://dx.doi.org/10.1063/1.1727217View Description Hide Description
A low‐temperature infrared cell capable of maintaining temperatures between 7° and 50°K is used for the study of the vibration—rotation spectra of HCl and DCl in solid rare gases. Spectra at 0.2% to 2% concentrations in solid krypton and xenon are obtained. Spectra in argon have already been reported by other workers. Fine structure in the region of the (1–0) vibrational transitions is assigned to monomeric and polymeric species of HCl and DCl. The polymer lines can be assigned to dimers and trimers. A detailed description of these interesting features will appear in Paper II of this series. The monomer lines are assigned to vibration—rotation transitions of HCl and DCl. In addition to a J‐independent band‐origin shift, the rotational levels of HCl exhibit J‐dependent perturbations of 5 to 10 cm−1; the J‐dependent perturbations of the DCl rotational levels lie between 1 and 3 cm−1. Neither hindered rotation nor libration of HCl and DCl in the rare‐gas crystals provides a suitable description of these perturbations. A model suggested by Friedmann and Kimel based on the interaction of the rotational motion with the localized lattice vibrational (``translational'') motion of HCl and DCl at substitutional sites in the rare‐gas crystals is used to interpret the J‐dependent perturbations. A Hamiltonian, which treats the ``translational'' frequencies of HCl and DCl and the strength of the translation—rotation interaction as parameters, is constructed. The Hamiltonian matrix is truncated at a convenient size, and the perturbed energy levels are found by direct diagonalization of the truncated matrix. For a heteronuclear diatomic rotor in a cubic lattice, the translational—rotational states are shown to form a basis for the irreducible representations of a group isomorphous with D 2h . This property allows considerable factorization of the Hamiltonian matrix. The values of the ``translational'' frequencies which give the best fit to the observed rotational spacings are 36 cm−1 in xenon, 57 cm−1 in krypton, and 31 cm−1 in argon. The model in the case of argon is likely to be fuzzy because of the near mass resonance between argon and HCl or DCl. The strength of the translation—rotation interaction as measured by the distance between the center of mass and the center of electrical symmetry in the molecule is 0.13 Å for HCl and 0.096 Å for DCl. A mechanism for the induction of a Q branch in the vibration—rotation spectra of monomeric HCl and DCl in the solid rare gases is discussed in Paper III of this series.
44(1966); http://dx.doi.org/10.1063/1.1727218View Description Hide Description
A systematic method is presented for calculating the kinetics of a fluid for small time intervals in terms of the motion of a small number of particles. The method is based on a ``cluster expansion'' of the time evolution operator. The theory applies equally well at gas or liquid densities. As an example, the two‐particle form of the theory is used to estimate the self‐diffusion coefficient of the argon and hard‐sphere liquids and the viscosity and thermal conductivity of the hard‐sphere liquid. One new result is the exact calculation of the delta‐function singularity of the time‐dependent correlation functions describing ``collisional'' transport in the hard‐sphere liquid.
44(1966); http://dx.doi.org/10.1063/1.1727219View Description Hide Description
New measurements of the K x‐ray emission spectra from the metal atoms in MnO4 −, CrO4 2—, and VO4 3— ions are presented and discussed in terms of dipole transitions from occupied orbitals to the 1a 1, i.e., the metal 1svacancy. Observed emission lines are attributed to transitions from the 3t 2(metal 3p), the 4t 2 (oxygen 2s), and the valence 6t 2 (predominantly oxygen 2pσ) orbitals. Simple theory accounts semiquantitatively for the relative intensities and energy positions of the lines. The K‐absorption spectra of the metal atoms in these ions are considered in conjunction with the emission data. The most prominent absorption peak in each spectrum is attributed to transitions to the 7t 2 orbitals, which apparently consist mainly of metal 4pwavefunctions. Also identified in each spectrum is the absorption edge associated with the onset of photoionization of the K shell. The energy differences between the emission lines and the K‐absorption edge are taken to be the molecular orbital ionization potentials. There is not good agreement between the present deductions from the x‐ray spectra and the results of recent semiempirical calculations of the electronic structure of the MnO4 − ion.
44(1966); http://dx.doi.org/10.1063/1.1727220View Description Hide Description
Volume magnetic susceptibilities have been measured by the Gouy method for molten potassium chloride and for 7 and 15 mole % solutions of potassium metal in molten potassium chloride, between 800° and 1000°C. All these liquids are diamagnetic, but the potassium solute appreciably reduces the diamagnetism of the solvent. Making some assumptions about the densities (unmeasured) of the solutions and about the additivity of the ionic diamagnetism, it is possible to compute molar paramagnetic susceptibilities for the potassium valence electrons. These values can be very roughly accounted for by the equation for conduction‐band electrons, with the effective electron mass approximately equal to the free electron mass. Alternatively, one can account for the data by localized electronmodels. One such model, based on a number of assumptions, is presented in the discussions. It leads to the conclusion that there exists a slight energy of association, of the order of kT or less, of potassium solute atoms into nearest‐neighbor clusters.
44(1966); http://dx.doi.org/10.1063/1.1727221View Description Hide Description
Wavefunctions for VO have been calculated by the method of Roothaan to establish the electronic ground‐state symmetry. The methods and basis functions used here are similar to those employed in previous studies of ScO and TiO. The present ab initio calculations show that the ground state of VO is 4Σ−, or possibly 2Σ− if configuration interaction is very important. These term types are associated with the configuration σδ2 and are not in conflict with present experimental evidence. The band systems of the molecule have not been adequately classified, but they are suggested to involve quartet and doublet terms. These theoretical ground‐state assignments are contrary, however, to the expected 2Δ r (σ2δ) term based on analogy with experimental results for NbO and TaO. It is not surprising, however, that a reversal in ground‐state configurations occurs in this family, because a similar reversal is known for the monoxides of the titanium family and for the ground‐state atoms of the vanadium family.
44(1966); http://dx.doi.org/10.1063/1.1727222View Description Hide Description
44(1966); http://dx.doi.org/10.1063/1.1727223View Description Hide Description
Raman spectra of molten MgCl2, MgCl2–KCl, and MgCl2–2KCl were recorded, at 730°, 525°, and 475°C, respectively, using a Cary 81 Raman spectrophotometer. Liquid MgCl2 exhibits three Raman lines with frequencies 269, 214, and 142 cm−1, which are explainable on the basis of the presence of MgCl6 4— ions with octahedral configuration. Liquid MgCl2–KCl exhibits four Raman lines with frequencies 451, 250, 208, and 158 cm−1; and their polarization characteristics indicate the existence of MgCl3 − ions with pyramidal configuration C 3v . The Raman spectrum of liquid MgCl2–2KCl is almost similar to that of liquid MgCl2–KCl. Force constants for the various Mg–Cl ionic species were calculated using generalized valence forces.
44(1966); http://dx.doi.org/10.1063/1.1727224View Description Hide Description
The 3Σ u −←3Σ g − transition of S2 has been observed in absorption in Ne and in fluorescence in Ne, Ar, Kr, Xe, and N2 matrices. The strongest bands observed in fluorescence belong to the (0, v″) progression as the result of rapid vibrational deexcitation of the 3Σ u − state. T 00=31 680, 31 350, 30 900, 30 780, and 31 480 cm−1, respectively, for S2 in the five matrices. Unlike the other matrices, argon yields very sharp fluorescent bands, having a width of 15 cm−1. The frequency of the forbidden infrared vibration transition has been calculated from the fluorescent data, and is found to differ from the observed infrared spectrum.
44(1966); http://dx.doi.org/10.1063/1.1727225View Description Hide Description
Excitation of the SPG (second positive group: C 3π u →B 3π g ) by fission fragments and by 6.1‐MeV alpha particles was studied with a pulse technique. Both types of particles exhibit nearly the same efficiency for excitation from which it was concluded that in either case secondary electrons are responsible for the production of the triplet states.
44(1966); http://dx.doi.org/10.1063/1.1727226View Description Hide Description
The narrow emission lines of Nd3+, Er3+, and Yb3+ have been observed from rare‐earth dopedCdS. These activators formed the principal radiative recombination centers in selected crystals. Decay‐time measurements gave fluorescent lifetimes of ∼100 μsec for CdS:Yb, ∼300 μsec for CdS:Nd and ∼3000 μsec for CdS:Er. Both the emission spectra and lifetime measurements showed that the rare‐earth ions occupied more than one type of site in the CdS lattice. No emission was observed from a rare‐earth ion metastable level greater than 1.53 eV above the ground state, even though the CdSband gap was 2.43 to 2.52 eV in the temperature range of the experiments. This is taken as evidence for rare‐earth‐ion‐acceptor‐type defect pairing.
44(1966); http://dx.doi.org/10.1063/1.1727227View Description Hide Description
The approximate self‐consistent molecular orbital theory with complete neglect of differential overlap (CNDO) presented in earlier papers has been modified in two ways. (a) Atomic matrix elements are chosen empirically using data on both atomic ionization potentials and electron affinities. (b) Certain penetration‐type terms, which led to excess bonding between formally nonbonded atoms in the previous treatment, have been omitted. The new method (denoted by CNDO/2) has been applied to symmetrical triatomic (AB2) and tetratomic (AB3) molecules, for a range of bond angles. The theory leads to calculated equilibrium angles, dipole moments, and bending force constants which are in reasonable agreement with experimental values in most cases.
44(1966); http://dx.doi.org/10.1063/1.1727228View Description Hide Description
A linear FE model with electron—electron and electron—core interaction appeared to be fairly accurate in reproducing the experimental data for (i) lowest excited singlet—triplet separation in polyenes, (ii) twisting frequencies of ethylene, and (iii) location of the absorption maxima in cumulenes. For calculations of (i) and (ii), electron—electron interaction has been assumed to play the dominant role. With the same assumption made, cis—transisomerization of polyenes was found to be easier with long chains than with short. It also is easier about a central double bond in the chain than about an outer double bond. In the case of a single double bond, an important part of the energy path in isomerization is that for the triplet state; however, the role of triplet states in isomerization should diminish rapidly with an increase in the number of double bonds in the chain. With ethylene as the example, the effect of pi‐electron correlations in the FE model was briefly investigated, and the correlation for the perturbation calculus of an order higher than the first was found to be in the FE model much less important than in the LCAO theory. No empirical parameters other than the fundamental atomic constants and the length of the FE path of a given molecule were used in the calculations.
44(1966); http://dx.doi.org/10.1063/1.1727229View Description Hide Description
The rate of evaporation from surfaces of Al2O3, Ga2O3, and In2O3 has been studied through a combination of thermal‐imaging and mass‐spectrometric techniques.
A discontinuity is observed in the rate of evaporation of Al2O3 and Ga2O3 at the melting point TM. The evaporation coefficient α v of the solid at the melting point is equal to 0.3±0.05 for all gaseous species over Al2O3, Ga2O3, and In2O3. The rate of evaporation from the surface of the liquid at the melting point is the equilibrium rate (α v =1) for all gaseous species. The evaporation coefficient α v over the solid is very nearly temperature independent for monatomic vapor species but depends strongly on temperature for polyatomic species.
It is shown that the data for the different systems can be directly correlated when they are plotted as logα v vs TM/T, where TM is the melting temperature. This correlation also applies to other systems and appears to be useful in predicting evaporation behavior in related systems.
The equilibrium rates of evaporation are used to calculate: D 0°(AlO) = 116.0±3.0 kcal mole−1, D 0°(Al2O) = 241.6±7.0 kcal mole−1, ΔH° v,298(Ga) = 64.2±2.0 kcal mole−1, D 0°(GaO) = 90.2±3.5 kcal mole−1, D 0°(Ga2O) = 206.8±7.0 kcal mole−1, D 0°(In2O) = 179.2±4.0 kcal mole−1.
44(1966); http://dx.doi.org/10.1063/1.1727230View Description Hide Description
The velocity of sound in NH3 at 300°K has been measured at a frequency of about 1 MHz over the pressure range from 6 atm down to ½ atm, employing the optical diffraction method. All measuredvelocities were lower than the low‐frequency ideal‐gas limit of 436 m sec−1. Interpreting this result as evidence of nonideality and not dispersion, we fitted the data to the formula V 2=Vi 2(1+sp) and obtained Vi 2=(19.122±0.034)×104 m2 sec−2 and s=(−13.9±0.05)×10−3 atm−1. The ideal velocity squared, Vi 2, agrees with the spectroscopic value to within 0.05% and shows that no loss of vibrational heat capacity occurs up to f/p=2.1 MHz atm−1.
The nonideality correction term s is used to recalculate nonideality corrections to the data of previous workers, and it is suggested that the vibrational relaxation time in NH3 may be of the order of 6 nsec at 300°K.