Slow positron pulses are implanted at variable energies into a Si02 powder target to form positronium (Ps). By time-of-Aight methods we estimate the energy spectrum of the Ps that diffuses out of the target into the vacuum. As expected, higher implantation energies and lower target temperatures result in a colder Ps energy spectrum. However, even with 19-keV positron implantation energies, the Ps escaping from a target at 4.2 K is only about 2% thermalized. No evidence was found for a low-energy cutoff due to Ps localization between the powder grains.

We measure by time of Aight the energy distribution of positronium (Ps) emitted normally to a clean Al(111) surface that is bombarded with energetic (1.5 -4.5 keV) positron pulses. The data are consistent with the predictions of a free-electron model for nonadiabatic Ps emission with less than 10 ' of the Ps emitted normal to the surface in a monoenergetic peak at the maximum allowable energy. The Ps work function at 300 K is ( -2. 536+0.035) eV in agreement with the ( -2. 519+0.027) eV obtained from the sum of the electron and positron work functions minus the binding energy of Ps.

No abstract

We have measured the energy spectrum of positronium (Ps) thermally desorbed from Al(lll) surfaces. For sample temperatures as low as 84 K and Ps perpendicular energies between 5 and 50 meV the spectrum is a beam Maxwellian, i.e., a simple exponential. We argue that the surface must be essentially a blackbody for Ps emission, and that we thus have the first example of a system that fails to exhibit the expected perfect reflection of a very slow particle from a cold surface. We suggest that the effect is associated with the breakdown of the perturbation expansion for the Ps reflection probability.PACS numbers: 68.10.Jy, 36.10.Dr One quantity characterizing the interaction of a relatively slow neutral particle with a solid surface at a temperature 7 is the probability S(k,T) that the particle will stick when it impinges upon the surface with momentum k. A question of fundamental importance to quantum many-body theory is: What is the probability SQ = S(0,0) in the limit of zero velocity and zero 7? Given a very weak short-ranged coupling to some type of inelastic channel such as phonon or electron-hole pair emission, it is possible to show that So -0 due to the vanishing of the scattering particle's wave function at the surface. 2 Higher-order virtual processes, which may be characterized by a dimensionless coupling constant A, renormalize the wave function near the surface, increase the effective mass of the particle, and thus enhance the possibilities for sticking. It has been suggested that for A of order unity, where the perturbation series breaks down, the inelastic channels will overcome the universal tendency of a particle to quantum reflect from a potential whose characteristic size is much less than a de Broglie wavelength. [3][4][5][6][7] The basic questions about quantum sticking are as follows: (1) May -So ever be greater than 0 for a realizable system? (2) If so, may So have values between 0 and 1 or may it only have the extreme values 0 and 1? (3) In the latter case, might it be possible to vary the conditions of a sample surface so as to exhibit the sudden transition from So = 0 to So = l suggested by a recent calculation? 8 ' 9Possible systems for experimental study of the quantum-sticking problem would include scattering from solid surfaces using atoms with de Broglie wavelengths \-^ 100 A so that the range of the surface potential and the corrugations of the surface become negligible. The scattering of atomic H and He from He films 10 " 12 are well-known examples of studies that yield the extrapolated result -So^O with a nontrivial momentum dependence of S(k, 7). In these examples the surface potential is weak, so perhaps the lack of sticking is not surprising.An almost ideal atom in this regard would be positronium (Ps), which is as reactive as H, but has a light mass that allows us to attain X = 100 A at a relatively convenient kinetic energy of 7.5 meV. In this paper we present new measurements of the thermal desorption of Ps from which we extract information about So using detailed balance ar...