We report high-resolution electron-energy-loss spectra at 14.3 and 19 eV incident electron energy for electronic excitations in multilayer films of amorphous ice. The lowest electronic transition has its threshold at 7.3 eV and its maximum at 8.5 eV. It is followed by a broad shoulder around 10.4 eV and a rise above 11.6 eV.

Electron-impact-excitation cross sections for ${\mathrm{Zn}}^{+}$ are calculated in a multiple-LS-term close-coupling approximation. Total cross sections for the 4s→4p transition are compared with crossed-beams [Rogers et al., Phys. Rev. A 25, 681 (1982)] and merged-beams [Smith et al., Phys. Rev. Lett. 67, 30 (1991)] experiments. Taking into account cascade effects, the close-coupling calculations and the crossed-beams measurements are found to be in reasonable agreement over a wide energy range. When only the forward-angle part of the cross section is calculated, good agreement with the merged-beams measurements is found near the excitation threshold, but at the higher energies the experimental value is somewhat lower than the theory predicts.

The phase-shift equations are derived in a form that is convenient for the analysis of both the radial Schrödinger equation and the problem of wave propagation in nonhomogeneous media. In the case of bound quantum states, the conservation condition is discussed for the sum of the classical action (or eikonal) and the phase. The expression for the reflection coefficient describes exactly two opposite limiting cases: the Fresnel reflection and the short-wavelength limit. In the case of wave propagation in symmetric media, both transmission and reflection coefficients are given as elementary functions of the phase shifts only.

The simultaneous optimization of the bound and continuum multiconfiguration Hartree-Fock orbitals for an electron scattered by an atom leads to a set of bound orbitals for the target atom that completely takes into account the polarization of the target by the electron. The polarized orbitals that are generated for zero kinetic energy and angular momentum of the continuum electron form an excellent basis for the calculation of the static dipole polarizability. We have done this for the ground states of the inert-gas atoms helium, neon, and argon. Our results agree extremely well with experimental measurements, indicating that this ab initio approach provides very good wave functions with which to determine highly accurate values of the polarizability.

This paper reports measurements of absolute differential cross sections for electron capture and loss for fast hydrogen atoms incident on ${\mathrm{H}}_2$, ${\mathrm{N}}_2$, ${\mathrm{O}}_2$, Ar, and He. Cross sections have been determined in the 2.0- to 5.0-keV energy range over the laboratory angular range 0.02°–2°, with an angular resolution of 0.02°. The high angular resolution allows us to observe structure at small angles in some of the cross sections. Comparison of the present results with those of other authors generally shows very good agreement.

Measurements of the cross sections for the electron-impact dissociative excitation of ${\mathrm{H}}_3^{+}$ as a function of electron energy are reported. Sharp structures are observed close to the excitation thresholds for the upper states of the ion, and these are interpreted as Feshbach resonances. Direct excitation appears to play a subordinate role at the energies studied. The implication of these findings applied to previous measurements of the dissociative recombination of ${\mathrm{H}}_3^{+}$ is discussed.

In this work we have made an intense theoretical effort to determine accurate K-Auger transition energies, Auger rates, and line intensity ratios for Be- and B-like multiply charged neon ions using several different atomic models. Theoretical data on Li-like states in Ne viii are also reported. In particular a detailed study of Auger line intensity ratios is presented. For several Be- and B-like states strong cancellations in the Auger matrix elements are predicted. These rates are quite sensitive to the atomic wave functions. We have also deduced experimental Auger intensity ratios for Be and B-like Ne ions from previous data. In order to classify our high-resolution projectile spectra, theoretical Auger energies and rates for various terms of the electron configurations 1s2${\mathit{s}}^2$2${\mathit{p}}^{\mathit{n}\mathrm{-}1}$, 1s2${\mathit{p}}^{\mathit{n}+1\mathrm{}}$, and 1s2s2${\mathit{p}}^{\mathit{n}}$ (n=1–3) have been calculated relativistically, using the saddle-point technique, the multiconfiguration Dirac-Fock model, and the 1/Z expansion method. Several previously unknown line identifications of Be-like states in Ne vii and B-like states in Ne vi are reported.

The neutralization of highly charged ions during interaction with metallic surfaces is accompanied by the ejection of a large number of secondary electrons. Recent experiments demonstrate two main contributions to this electron ejection process: one from the region below the surface and the second from the above-surface portion of the trajectory. We present a theoretical analysis of the neutralization dynamics above the surface, prior to impact, based on the classical over-the-barrier model. The theory incorporates resonant multielectron capture of conduction electrons, resonant loss into unoccupied states of the conduction band, and intra-atomic Auger deexcitation. The effective barrier potential includes quantum corrections to the classical image potential. The effect of below-barrier (‘‘tunneling’’) transfer is investigated. The solution of a coupled system of rate equations allows the approximate determination of the n-shell populations, the projectile charge state, and the total number of Auger electrons. The calculation describes the transient formation of ‘‘hollow’’ atoms. We find satisfactory agreement with recent data for K Auger yields by Meyer et al. [Phys. Rev. Lett. 67, 723 (1991)].

Electron excitation to H n=2 and n=3 states in 1–300-keV/u ${\mathrm{He}}^{2+}$-H collisions is studied within the semiclassical close-coupling method with atomic-orbital basis sets. Calculated excitation cross sections tie in smoothly with results from high-energy theories above 100 keV/u; they also agree with recently published H n=3 to n=2 line-emission cross sections [A. Donnelly, J. Geddes, and H. B. Gilbody, J. Phys. B 24, 165 (1991)]. An interesting aspect of this study is that low-energy-excitation cross sections show structures that may be explained by the complex excitation mechanism. These structures are confirmed in as yet unpublished H2p-1s line-emission cross sections. No other information exists for energies below 10 keV/u.

The elastic scatterings of electrons and positrons from rubidium are calculated at intermediate energies in the Born, Glauber, modified Glauber (MG), and eikonal-Born-series approximations. The results of differential cross sections (DCS) for elastic scattering obtained within these various methods of approximation are presented with discussion. The modified Glauber results predict the existence of structures for the DCS, similar to those of scattering from potassium and sodium. Furthermore, the structures in this case persist even at a scattering energy as high as 500 eV. The DCS in the modified Glauber approximation at small scattering angles are used to estimate the ‘‘effective’’ total cross sections for positron and electron scatterings from rubidium. The effective MG total cross sections are found to be in better agreement with experimental data.

K-shell x-ray-production cross sections are reported for elements with K-shell x-ray energies between 277 eV (C) and 1487 eV (Al). The x-ray measurements were made with a windowless Si(Li) detector that was calibrated for efficiency by comparing bremsstrahlung spectra from electron bombardment of thin foils of aluminum, silver, and gold with theoretically determined bremsstrahlung spectral distributions. The x-ray-production cross-section measurements are compared to first-order Born and perturbed-stationary-state with energy-loss, Coulomb deflection, and relativistic corrections (ECPSSR) ionization theories using single-hole fluorescence yields. The ECPSSR and first-order Born theoretical predictions are, in general, in close agreement with each other and both generally fit the data quite well.

A method is proposed for simulating resolved transition arrays of ionized atom spectra in full intermediate coupling. The wave number and intensity of each line in an array are picked at random from separated but correlated distributions. Even though each line is not exactly reproduced, this procedure yields the correct following characteristics of the supposedly symmetric array: total intensity; second and fourth moments of the distributions of unweighted wave numbers, of intensity-weighted wave numbers, and of transition amplitudes; numbers of lines and sums of intensities in consecutive narrow energy ranges. All the parameters of the distribution are obtained by means of compact formulas, or tabulated. Applications to the arrays 4${\mathit{d}}^4$-4${\mathit{d}}^3$5p of ${\mathrm{Pd}}^{6+}$ and 4${\mathit{d}}^7$5s-4${\mathit{d}}^7$5p of ${\mathrm{Cd}}^{4+}$ are presented. Comparison with the explicit results of the Slater-Condon method shows good agreement. It is proposed to use this method for fast and reliable computation of Rosseland means and other opacity properties.

A Monte Carlo technique for modeling spectra with many resolved lines of ions in intermediate coupling was presented in the preceding paper [Phys. Rev. A 44, 5707 (1991)]. We use the same concept here with a modification for the near-LS-coupling case. Rosseland and Planck means of transition arrays of Fe v and Fe vi are evaluated. Comparison with ‘‘exact’’ computations shows much better agreement than is found using the simple unresolved-transition-array approximation.

Collisional broadening of nS Rydberg states by K, Rb, and Cs atoms is studied using the adiabatic approach for the ${}_{}{}^3{}_{}{}^{}\mathrm{\Sigma }$ quasimolecular state and impulse approximation for all other contributions to the collisional width. The calculations are based on the use of the low-energy phase shifts for electron scattering by alkali metals obtained from ab initio close-coupling calculations and their extrapolation according to the modified effective range theory. We obtain a good agreement for the magnitude of the width with experimental data and satisfactory agreement for the oscillatory part of the width as a function of the principal quantum number. We find that the theory overestimates the amplitude of the oscillations. In the case of Cs as a perturber, the agreement with experimental data is good if we assume the existence of the low-energy ${}_{}{}^3{}_{}{}^{}$P resonance of ${\mathrm{Cs}}^{\mathrm{-}}$. However, recent ab initio bound-state and scattering calculations indicate the existence of the Cs ${(}^3$P) bound state rather than the resonance.

Excitation spectra of the CdKr excimer were produced in a molecular beam employing free-jet supersonic expansion, crossed with a pulsed dye-laser beam. Bands arising from B1←X$0^{+}$ and A$0^{+}$←X$0^{+}$ vibronic transitions were analyzed, yielding spectroscopic constants for the three states. Computer simulations of the bands yielded relative values of ${\mathit{r}}_{\mathit{e}}$, the equilibrium internuclear separations. Potential-energy curves constructed on the basis of the data and of the Morse potential provided a useful comparison with theoretically calculated curves.

In the analysis of high-resolution Doppler-free two-photon hyperfine-structure (hfs) measurements of levels in the Eu i term ${}_{}{}^6{}_{}{}^{}\mathrm{D}$ of the configuration 4${\mathit{f}}^7$6s6d of ${}_{}{}^{151}{}_{}{}^{}\mathrm{Eu}$ and ${}_{}{}^{153}{}_{}{}^{}\mathrm{Eu}$ the necessity occurs to take into account also hfs perturbations. For the magnetic hfs interactions of ${}_{}{}^{151}{}_{}{}^{}\mathrm{Eu}$, the following off-diagonal matrix elements ${}_{}{}^{151}{}_{}{}^{}$${\mathit{A}}_1$(JJ’) could be determined from the high-resolved two-photon line profiles: ${}_{}{}^{151}{}_{}{}^{}$${\mathit{A}}_1$(1/2,3/2)=5.70(3) GHz, ${}_{}{}^{151}{}_{}{}^{}$${\mathit{A}}_1$(3/2,5/2)=4.3(2) GHz, ${}_{}{}^{151}{}_{}{}^{}$${\mathit{A}}_1$(5/2,7/2)=2.9(2) GHz, ${}_{}{}^{151}{}_{}{}^{}$${\mathit{A}}_1$(7/2,9/2)=2.2(2) GHz. The high quality of the calculations is represented both in the reproducibility of the absolute hfs level energies and also in the hfs anomaly where a mean value deduced from the corrected magnetic hfs constants of -0.66(8)% can be stated.

It is shown that the introduction of configuration interaction through 4s4f-4p4d and 4p4f-4${\mathit{d}}^2$ mixing has a very significant effect on 3d-4f transitions in highly charged Ga-, Zn-, and Cu-like ions. This effect changes the width and, to a lesser extent, the mean wavelength of transition arrays as compared to the spin-orbit split array model. Because of this mixing, lines previously identified as dielectronic recombination satellites may have a non-negligible inner-shell excitation component. This should be taken into account in models for the intensities of these lines.

The x-ray spectrum of laser-produced uranium plasma has been obtained in the range 3.1–6.4 keV (1.9<λ<3.9 Å). Detailed analysis of the spectrum has been performed using the hullac code and the unresolved-transition-array formalism. This led to the identification of most of the features of the spectrum as transitions in the Ni-like ion (${\mathrm{U}}^{64+}$) and satellite transitions of neighboring ionization states. The identification is fully consistent with the previous analysis of the spectra emitted by laser-produced plasmas of lower-Z elements.

A recently proposed scheme for calculating K-shell ionization cross sections in atoms and molecules using the relaxed-core Hartree-Fock (RCHF) approximation is extended to the treatment of shakeup processes. As an example, this scheme is applied to the π-${\mathrm{\pi }}^{\mathrm{*}}$ satellites in the C 1s photoelectron spectrum of CO. These results for the (S0, S1) pair of π-${\mathrm{\pi }}^{\mathrm{*}}$ satellite states reveal strong interference effects between the ‘‘direct’’ (bound-free dipole integral) and ‘‘conjugate’’ (bound-free overlap integral) contributions in the transition moment at near-threshold energies. These effects are particularly dominant in the kπ subchannel and, for the triplet-coupled satellite (S1), lead to a drastically increasing cross section with decreasing photon energy. For the singlet-coupled satellite (S0) destructive interference leads to a comparatively small cross section near threshold. While these findings may explain the strikingly different behavior of the S0 and S1 satellite intensities in recent experiments, the present results also show distinct σ-type shape resonances in both satellite channels at about 10 eV photoelectron energy that are apparently at variance with the experimental evidence. An analysis of the RCHF transition moment via perturbation theory shows that already in first order some potentially important dynmical contributions are missing. In these studies a direct method based on the Schwinger variational principle and single-center expansion techniques is used to obtain the photoelectron orbitals.

Ab initio methods for calculating the binding-energy spectra of large molecules have traditionally been restricted to primarily either Koopmans’s theorem or the density-functional transition-orbital method. The limitations of the former are well known, and the density-functional ‘‘band-gap problem’’ has led to a further realization of intrinsic difficulties with the latter. An increasingly popular alternative to these two methods is to seek a simple approximation for the Green-function self-energy. The Green-function self-energy is the optical potential seen by a scattering particle (hole). As such, the dominant many-body effects contributing to the self-energy result from polarization of the charge density at energies below the first excitation energy of the target molecule (quasiparticle regime), as well as excitations of the target at higher energies. The physical importance of polarization effects is apparent in Hedin’s GW approximation, which treats the self-energy as a product of the Green function (G) and a screened interaction (W) that can be calculated (essentially) from the time-dependent linear response of the charge density. In the present paper, we examine the contribution of polarization to the usual second-order Green-function (GF2) approximation with respect to the calculation of outer-valence ionization potentials in small molecules. A simplified version (GW2) of the GW approximation is found to be an acceptable substitute for the GF2 approximation, provided a self-interaction correction is included to prevent an electron from polarizing itself.

Polarization effects are further analyzed using the Coulomb-hole and screened-exchange (COHSEX) and modified-COHSEX (M-COHSEX) approximations. A second-order version (M-COHSEX2) of the M-COHSEX approximation is used to examine the origin of the incorrect ordering by Koopmans’s theorem of the first three ionization potentials of the nitrogen molecule in terms of static polarization and retardation effects. Finite-basis-set errors are also explored. Although higher-order Green-function approximations must be examined before drawing final conclusions, we believe that the present work provides preliminary evidence that suitably modified versions of time-dependent density-functional, dielectric-function-based self-energy approximations can be useful for molecules.