The Naval Research Laboratory (NRL) has developed a processing system based on an electron beam-generated plasma. Unlike conventional discharges produced by electric fields (DC, RF, microwave, etc.), ionization is driven by a high-energy (∼ few keV) electron beam, an approach that can be attractive to atomic layer processing applications. In particular, high electron densities (10 10 -10 11 cm −3 ) can be produced in electron beam generated plasmas, where the electron temperature remains between 0.3 and 1.0 eV. Accordingly, a large flux of ions can be delivered to substrate surfaces with kinetic energies in the range of 1 to 5 eV. This provides the potential for controllably etching and/or engineering both the surface morphology and chemistry with monolayer precision. This work describes the electron beam driven plasma processing system, with particular attention paid to system characteristics and the ability to control the generation and delivery of ions to the surface and their energies. Electron beam generated plasmas are produced by injecting a highenergy electron beam into a gas background, which will ionize, dissociate, and excite atoms or molecules as it traverses the gas volume. While the basic inelastic processes that lead to species production are the same as those in discharge plasmas, the use of energetic electron beams to drive production results in plasmas that have very different properties than conventional discharges. Some of these properties are attractive for plasma-based atomic layer processing applications where, whether etching, depositing, or chemically modifying materials, fine control over the flux and energy of ions is needed. In the case of atomic layer etching, perhaps the single most important need is to tightly control the kinetic energy of ions incident to the processing surface so as to avoid damage while maintaining a reasonable etch rate. [1][2][3] In this regard, a significant advantage of beam-driven plasmas is the inherently low electron temperature T e , which is typically a fraction of an eV in most gas mixtures of technological interest. Importantly, this is true regardless of the plasma density. Thus, one can produce a large fluence of reactive ion and neutral species where the kinetic energy of the ions is as low as a few eV.The interest in electron beam generated plasmas for materials processing can be traced back at least 4 decades. In the early 1970s Bunshah 4 described an electron beam vapor deposition system that utilizes an electron beam to both vaporize the metal and "activate" the background gas. It was also noted by Dugdale 5 that high-energy electron beam systems developed for welding could be employed for "soft vacuum vapor deposition" in a manner similar to that of Bunshah. In the 1980s, Collins and co-workers at Colorado State University, developed electron beam produced plasmas for plasma enhanced chemical vapor deposition of SiO 2 . 6,7 This system used a sheet-like beam of multi-keV electrons injected parallel to the growth substrate. A similar configurat...

This paper, the first in a series of two, presents experimental results demonstrating the control of electron and ion energy distribution functions in electron beam generated processing plasmas by adding trace concentrations of N 2 to an Ar background. Measurements of the electron energy distribution function, f 0 (E), are performed using a Langmuir probe while measurements of the Ar ion energy distribution function are performed using an energy-resolved mass spectrometer. The experimental results agree with modeling results, described in part II of this work, which indicate that inelastic electron collisions with nitrogen molecules provide an energy sink that can be exploited to control the electron energy distribution function.

Numerical simulations of neutron production from deuterium-lithium nuclear fusion reactions have been performed. A set of differential cross sections for the 7 Li(d,xn) reaction for incident deuteron energies of up to 50 MeV is assembled. The angular distribution of neutrons from a thick lithium target is simulated and benchmarked against experimental data. Two-stage neutron production from laser-target experiments has been studied as a function of laser intensity and energy. During the first stage a well collimated deuteron beam is generated using a high-intensity ultrashort pulse laser. During the second stage it is transported through a lithium target using a 3D Monte-Carlo ion beam-target deposition model. The neutron yield is estimated to be ∼10 8 neutrons J −1 laser energy. Some 10 10 neutrons can be expected from a ∼100 J petawatt-class laser. For incident deuteron energies above 1 MeV the proposed scheme for neutron production from d-Li reactions is superior to that from d-d reactions, producing a collimated beam of neutrons with higher neutron yield.

The effect of nitrogen addition on the emission intensities of the brightest argon lines produced in a low pressure argon/nitrogen electron beam-generated plasmas is characterized using optical emission spectroscopy. In particular, a decrease in the intensities of the 811.5 nm and 763.5 nm lines is observed, while the intensity of the 750.4 nm line remains unchanged as nitrogen is added. To explain this phenomenon, a non-equilibrium collisional-radiative model is developed and used to compute the population of argon excited states and line intensities as a function of gas composition. The results show that the addition of nitrogen to argon modifies the electron energy distribution function, reduces the electron temperature, and depopulates Ar metastables in exchange reactions with electrons and N2 molecules, all of which lead to changes in argon excited states population and thus the emission originating from the Ar 4p levels.

In this work, the second in a series of two, a spatially averaged model of an electron beam generated Ar-N 2 plasma is developed to identify the processes behind the measured influence of trace amounts of N 2 on the development of the electron energy distribution function. The model is based on the numerical solution of the electron Boltzmann equation self-consistently coupled to a set of rate balance equations for electrons, argon and nitrogen species. Like the experiments, the calculations cover only the low-energy portion (<50 eV) of the electron energy distribution, and therefore a source term is added to the Boltzmann equation to represent ionization by the beam. Similarly, terms representing ambipolar diffusion along and across the magnetic field are added to allow for particle loss and electrostatic cooling from the ambipolar electric field. This work focuses on the changes introduced by adding a small admixture of nitrogen to an argon background. The model predictions for the electron energy distribution function, electron density and temperature are in good agreement with the experimentally measured data reported in part I, where it was found that the electron and ion energy distributions can be controlled by adjusting the fraction of nitrogen in the gas composition.