Lithium oxygen termination as a negative electron affinity surface on diamond

: a computational and photoemission study

Abstract

This thesis presents the investigation of the interaction between lithium and the diamond surface using a number of computational and experimental techniques, with the goal of producing a low workfunction material for electron emission. Because of its theoretically shallow donor state in the diamond bulk and its potential as an alkali metal to induce a negative electron affinity (NEA) effect on the diamond surface, lithium is an interesting candidate for doping diamond.

The reaction of diamond HPHT nanoparticle powders with lithium salts at temperature in an ambient gas have been investigated. Thermionic and field emission performance was substantially improved following lithiation and further improved after washing in acqua regia, but the cause of this improvement was unclear. Characterisation of the doped powders with electron microscopy, secondary ion mass spectroscopy and electron energy loss spectroscopy was inconclusive, but indicated that any changes were likely to be due to a surface efect rather than bulk doping, so the remainder of the investigation focused on the interaction between lithium and the diamond surface.

Using ab initio density functional theory calculations, the stability of lithium with the C(100) and C(111) surfaces was investigated. Although lithium on the bare C(100) surface gave a predicted NEA surface of between -1.07 and 2.7 eV, the surface was fairly weakly bonded. When the adsorption of lithium on the oxygenated C(100) surface was calculated, the surface complex was much more stable, with adsorption energies as high as 4.7 eV for both the 0.5 and 1 ML coverages of Li. In addition, the electron affnity of the surface was even lower, with the LiO surface with two oxygen and two lithum atoms per unit cell having an electron affinity as negative as -4.5 eV, with a workfunction shift of -3.9 eV.


A similar behaviour was calculated on the C(111) surface, albeit with a slightly different configuration due to the one fewer dangling bonds per unit cell available on this surface. The adsorption of lithium on the bare C(111) surface was even less stable, but the 0.5 ML adsorption of Li on the fully oxygenated surface had a similar NEA to the C(100) surface of -3.97 eV, with an adsorption
energy of 4.37 eV. These predictions for the two most prevalent surfaces on diamond indicate that lithium on the oxygenated diamond surface could present a stronger NEA surface than hydrogenated diamond but with a much greater stability than the equivalent CsO surface construction previously reported in the literature.

The viability of the LiO surface termination was explored experimentally using X-ray and ultraviolet photoemission (XPS and UPS), as well as low energy electron diffraction (LEED) and secondary electron emission using a scanning electron microscope. After lithium evaporation on the ozone-treated oxygenated surface and washing of the sample in deionised water, lithium was still observable using XPS, whereas on the hydrogenated sample with a similar treatment no such signal remained. Likewise, LEED patterns showed a signicant change after Li coating and washing, consistent with the predicted structure from the computational results.

The LiO coated surface had a high secondary electron emission yield when observed in a scanning electron microscope, comparable to the hydrogen terminated surface and indicative of an NEA. UPS measurements of the boron-doped C(100) surface showed a clear characteristic sharp NEA peak around 5.2 eV in kinetic energy, with a calculated workfunction of 2.8 + 0.1 eV and an NEA of -2.1 + 0.1 eV. After annealing to a series of temperatures, it was found that theUPS spectra remained fairly consistent until annealing to above 925C, when the spectra began to change, with XPS and UPS showing a graphitisation of the surface and a removal of surface oxygens after annealing to 1218C. This indicates that LiO on diamond is a strongly-bound, highly negative electron affinity surface. Experiments on the phosphorus doped C(111) surface were difficult to interpret due to large degrees of charging on the resistive surface, but an NEA peak did appear to be present.

The LiO surface termination was found to enhance the turnon and current density of field and thermionic emission, with as much as a forty-fold increase in thermionic emission current compared to a similar hydrogen-terminated sample. The conductivity of the surface treatment was also investigated, but on the macroscopic level the samples remained resistive after Li evaporation. Nonetheless this study has conrmed the predictions of a strongly bound highly negative electron affinity surface using LiO on diamond and this offers the potential for a number of applications.

Date of Award	15 Nov 2011
Original language	English
Awarding Institution	University of Bristol
Supervisor	Neil A Fox (Supervisor) & David Cherns (Supervisor)