Periodic intensity variations in the measured Compton profile anisotropies of ordinary ice $\mathrm{I}h$ correspond to distances of 1.72 and 2.85 Å, which are close to the hydrogen bond length and the nearest-neighbor O-O distance, respectively. We interpret this result as direct evidence for the substantial covalent nature of the hydrogen bond. Very good quantitative agreement between the data and a fully quantum mechanical bonding model for ice $\mathrm{I}h$ and the disagreement with a purely electrostatic (classical) bonding model support this interpretation and demonstrate how exquisitely sensitive Compton scattering is to the phase of the electronic wave function.

A theory for calculating the momentum distribution of annihilating positron-electron pairs in solids is presented. To test the theory, momentum distributions are measured by the Doppler broadening of the annihilation radiation for several bulk metals and semiconductors, as well as for semiconductor alloys and for positrons trapped at vacancies in semiconductors. The theory is based on a two-particle description of the annihilating electron-positron pair. Then, the electron-positron correlation effects, i.e., the enhancement of the electron density at the positron, depend on the electronic state in question. The theory is suited for calculating the high-momentum part of the annihilation spectrum that arises from the core electrons and which can be measured by the Doppler broadening using coincidence techniques. The ideas of the theory are justified by a good agreement between theory and experiment in the case of positron annihilation in undefected bulk lattices. Moreover, the comparison of the theoretical and experimental spectra for alloys and vacancy defects tests the theoretical description for the positron distribution in delocalized and localized states, respectively. © 1996 The American Physical Society.

First-principles calculations of positron-annihilation characteristics in solids are usually based on the local-density approximation (LDA) for positron-electron correlation. The LDA systematically overestimates the annihilation rate. As a remedy we introduce a generalized gradient approximation (GGA). Our results for several metals and semiconductors show that the GGA systematically improves the predictive power of positron lifetime calculations over those based on the LDA. We compare also the resulting positron energy levels in solids with data from slow-positron experiments.

The generalized gradient correction method for positron-electron correlation effects in solids [B. Barbiellini et al., Phys. Rev. B 51, 7341 (1995)] is applied in several test cases. The positron lifetime, energetics, and momentum distribution of the annihilating electron-positron pairs are considered. The comparison with experiments shows systematic improvement in the predictive power of the theory compared to the local-density approximation results for positron states and annihilation characteristics. © 1996 The American Physical Society.