Focused electron- and ion-beam induced processes

Focused electron- and ion-beam induced processing are well established techniques for local deposition and etching that rely on decomposition of precursor molecules by irradiation. These high-resolution nanostructuring techniques have various applications in nanoscience including attach-and-release procedures in nanomanipulation and fabrication of sensors (magnetic, optical and thermal) for scanning probe microscopy. However, a complete physical and chemical understanding of the process is hampered by the lack of suitable means to monitor and to access the numerous interrelated and time-varying process parameters (deposition and etch rate, yield, molecule flux and adsorption/desorption). This thesis is a first attempt to fill this gap. It is based on experimental and simulative approaches for the determination of process conditions and mechanical properties of deposited materials: Mass and force sensors: The use of tools merging micromechanical cantilever sensors and scanning electron microscopy was demonstrated for in situ monitoring and analysis. A cantilever-based resonant mass sensing setup was developed and used for real-time mass measurements. A noise level at the femtogram-scale was achieved by tracking the resonance frequency of a temperature stabilized piezoresistive cantilever using phase-locking. With this technique the surface coverage and residence time of (CH3)3PtCpCH3 molecules, the mass deposition rate, the yield, and the material density of corresponding deposits were measured. In situ cantilever-based static force sensing and mechanical modal vibration analysis were employed to investigate the Young's modulus and density of individual high aspect ratio deposits from the precursor Cu(hfac)2. Precursor supply simulations and experiments: A prerequisite to understand and quantify irradiative precursor chemistry is the knowledge of the local flux of molecules impinging on the substrate. Therefore, Monte Carlo simulations of flux distributions were developed and gas flows injected into a vacuum chamber were analyzed experimentally for the precursors Co2(CO)8, (hfac)CuVTMS, and [(PF3)2RhCl]2. The process parameters extracted from the mentioned approaches are valuable input for numerical focused electron- and ion-beam induced process models (Monte Carlo, continuum). We evaluated the precursor surface diffusion coefficient and the electron impact dissociation cross-section by relating deposit shapes to a continuum model.

Etude par STM de la déposition d'agrégats d'or sur TiO2 et mesure de l'émission électronique secondaire induite par l'impact d'agrégats sur une surface

Pierre Convers

This thesis work focuses on impact and diffusion processes as well as equilibrium positions of a metal deposited on a surface. The metal is deposited in the form of clusters containing n atoms in a controlled way. Gold or silver clusters cations (Au+n and Ag+n ) are used. Their size (n ∈ [1, 9]), charge state and deposition energy (E = 10 - 4500 eV) are well defined. The surface being at room temperature during the deposition, can be annealed and transferred in a scanning tunnelling microscope (STM) after deposition. The variable temperature microscope was developed during this thesis. The approach of the tip is carried out by an axial motor, which renders the microscope very stable. The first part of this work treats of the evolution of gold deposited on a TiO2(110) surface. Position and size of gold islands created by cluster or atomic depositions are determined by STM. Compared to atomic deposition at the same energy (10 < E < 100 eV), the cluster deposition produces smaller islands with a large islands density after annealing at 800 K. Regardless the deposition conditions, gold atoms diffuse on the surface already at 300 K. The impact energy is the key parameter to reduce the diffusion by cluster or atomic implantation or defect creation. Upon a high energy deposition, part of the deposited gold is invisible to the STM. The fate of these gold atoms is unclear : they are either buried under the surface or ejected in vacuum at the impact. The control of the islands size should permit the creation of a stable system with an optimum catalytic activity (i. e. CO combustion). Research groups have shown that this activity is strongly influenced by the islands size. It is equal to zero if the diameter is greater than 5 nm. The second part of this work is devoted to the electron emission γ induced by the impact of silver clusters Ag+n on a Pt and HOPG surface. Measurements are made by varying the impact energy (and so the speed v) of size-selected clusters on a defined surface. Impact velocity ranges from 104 to 105 m/s, which is lower than the classical threshold. Nevertheless every cluster size produces an electron emission. A potential emission (γ(v=0) ≠ 0) is observed for the monomer on both surfaces, which is caused by excited ions in a metastable state. The γ(v) curves are increasing, and no mesurable oscillation are observed. This does not confirm earlier work by Meiwes-Broer et al. These authors found oscillations in γ on similar systems relating them to the electronic structure of the clusters and the surface. A model based on heating of the electron gas is developed. This model gives good agreements with the γ(v) curves. The electronic temperature is estimated to 3000-8000 K. Similar behavior on both surfaces (Pt and HOPG) depending of the cluster size is shown. This behavior is probably related to the geometrical structure of the clusters. Finally, a more pronounced molecular effect is observed for the HOPG : the number of electrons emitted by the impact of a Ag+n is up to 7 times higher than the number of electrons emitted by impacts of n independent atoms. This value is smaller than 2 for the Pt surface.

EPFL2005

Physico-chemical study of the focused electron beam induced deposition process

Tristan Bret

The focused electron beam induced deposition process is a promising technique for nano and micro patterning. Electrons can be focused in sub-angström dimensions, which allows atomic-scale resolution imaging, analysis, and processing techniques. Before the process can be used in controlled applications, the precise nature of the deposition mechanism must be described and modelled. The aim of this research work is to present a physical and chemical description of the focused electron beam induced deposition process. As an introduction, a review of the literature, up to the present day, shows how this process was first identified as the origin of contamination in electron microscopes. Then the modifications made to a scanning electron microscope for deposition experiments are described. Gas supply systems and gas cryo-trapping devices were set up. Electrical and optical systems were constructed for an in-situ monitoring of the process. The experiment procedure was carried out in several phases, with the first being the deposition of flat films of carbon, copper-containing carbon composites, and pure gold for the study of the physics of the process. By measuring the fraction of the probe current that was absorbed in the sample, an accurate description could be made of the propagation of electrons. The backscattered electrons clearly influenced the deposition rate of Cu-containing films. A physical model for scanning deposition was created. The second phase is the deposition of metal-matrix composite tips from a stationary beam. Forward electron scattering was discovered to be responsible for tip formation and shape. A physical model of electron trajectories is included. The third phase is the construction of three-dimensional micro-structures by moving the beam during deposition. The deposited shapes result from an electron range that is larger than the structure size. Further investigation into the chemistry of the process involved the analysis of carbon films by micro-beam techniques. Regardless of the precursor used, the resulting films had a composition as C9H2±xO1±x (x

EPFL2005

Oxygen assisted focused electron beam induced deposition of silicon dioxide

Alexandre Perentes

Focused Electron Beam Induced Deposition (FEBID) is a rapid prototyping technique for the investigation, production and modification of 2 and 3-D nanostructures. The process takes place at room temperature, in the high-vacuum chamber of a Scanning Electron Microscope (SEM). The principle is based on the local FEB decomposition of molecules injected in the chamber and adsorbed on a surface. These molecules, called the precursors, are specifically chosen to contain the element of interest to be deposited. However, the electron induced decomposition results in contaminated materials containing additional parasitic elements from the precursor molecule. These contaminants are incorporated with the element of interest, and downgrade the properties of the materials and prevent therefore their use for many applications. Gas-assisted FEBID, the subject of this thesis, is a promising evolution and solution to the problem that was never extensively studied in literature. It consists in injecting a reactive gas simultaneously to the precursors, which will react in real time with the deposited contaminants to form volatile products, leaving ideally only the interesting material on the surface. This work focused on the Oxygen assisted FEBID of Silicon dioxide (SiO2) and on the understanding of the deposition process. SiO2 it is a high performance material for nano-optic and electronic applications, widely used in research and industry. Modifications were brought to an existing SEM to allow the simultaneous injection of two different and controlled gases, and to allow the in-situ control of growth of optical materials. Si-precursors of three families (alkoxy-, alkyl- and isocyanato-silanes) were compared. The FEBID material could be gradually converted from contaminated Si-materials to pure SiO2, above a [O2]/[precursor] ratio specific to each precursor chemistry and precursor flow. The SiO2 deposited was stoichiometric, C and OH–free, had an estimated density of 2.2, was amorphous and had optical properties close to that of fused silica. Optical structures for nano-optics and plasmonics were produced. Compared to traditional FEBID, O2 assisted FEBID required much lower energy to induce a decomposition reaction, and secondary electrons achieved 90 % of the deposition process in our setup. The electron efficiency could be multiplied by 20 compared to conventional FEBID. O2 assisted FEBID was insensitive to electron density in our setup. The conditions for no C and large deposition rates during O2 assisted FEBID were high molecule reactivity to O2, low sticking coefficient, low acceleration voltage and dwell time (< 6 µs), and large replenishment time (> 60 µs). As function of the precursor chemistry and flow, additional O2 influenced the deposition rate, which could be increased it by a factor of 7. The O2 and precursor molecules co-adsorption resulted in a surface coverage competition, which determined the deposition efficiencies. Limitations due to the residual water in the SEM were demonstrated and characterized. They appeared as Oxygen incorporation in the deposited materials, which influenced the deposition process. Residual water impinging flow was demonstrated to be responsible of the FEBID process deposition efficiency when no Oxygen was injected.

EPFL2007

Etude par STM de la déposition d'agrégats d'or sur TiO2 et mesure de l'émission électronique secondaire induite par l'impact d'agrégats sur une surface

Pierre Convers

EPFL2005

Physico-chemical study of the focused electron beam induced deposition process

Tristan Bret

EPFL2005

Oxygen assisted focused electron beam induced deposition of silicon dioxide

Alexandre Perentes

EPFL2007