Generation and applications of a broad atomic oxygen beam with a high flux-density via collision-induced dissociation of O2
Zhiqiang Hana,b, Liying Songa,b, Po-Wan Shumb,d*, Woon-Ming Lau c,a,b,*
a Beijing Advanced Innovation Center for Materials Genome Engineering, Center for Green Innovation, School of Mathematics and Physics, University of Science and Technology Beijing, Beijing 100083, Chinab Shunde Innovation School, University of Science and Technology Beijing, Foshan, Guangdong 528399, China
c School of Chemistry and Chemical Engineering, Linyi University, Linyi, Shandong, 276000, China
d The Sun Age (Guangdong) New Energy Ltd, China
Corresponding Author
*Woon-Ming Lau. Email:leolau@lyu.edu.cn
*Po-Wan Shum. Email:samshum@thesunage.com
Abstract
We detail the generation of a pulsed atomic oxygen (AO) broad beam with a high flux-density via collision-induced dissociation of O2 to support practical industrial exploitation of AOs, particularly for facilitating 2-dimenstional oxidation/etching at a fast rate of one-monolayer per second in an area ≥1000cm2. This innovation fuses the following interdisciplinary concepts: (a) a high density of O+ can be produced in an electron-cyclotron-resonance (ECR) O2 plasma; (b) O+ can be extracted and accelerated with an aperture-electrode in the plasma; (c) O+ with adequate kinetic energy can initiate a cascade of gas-phase collisions in the presence of O2; (d) collision-induced dissociation of O2 yields AOs with adequate kinetic energy which can cause additional collision-induced dissociation of O2. Computational simulations of such collisions, with both ab initio molecular dynamics and direct simulation Monte Carlo methods, are used to guide the experimental generation of the proposed AO-beam. We experimentally demonstrate the highest known AO mean flux-density of about 1.5×1016 atoms cm-2 s-1 in a broad-beam, and use it to oxidatively modify a self-assembled molecular layer of siloxane on a silicon wafer. In addition, we also demonstrate the growth of Al2O3 through an AO-assisted atomic layer deposition process at a room temperature.
KEYWORDS. Collision, Molecular dynamics, Monte Carlo, Oxidization, Atomic layer deposition
Introduction
Atomic oxygen (AO), generated by the dissociation of molecular oxygen, exhibits high chemical reactivity and has been extensively employed in investigating material oxidation and corrosion mechanisms, as well as in developing novel oxidation processes in the chemical industry.[1-6] Particularly, the generation of a broad beam of AOs for the controlled growth of a metal-oxide monolayer and the controlled oxidative-etch of an organic/carbonaceous device-constituent is highly desirable to fuel the emergency of 2-dimensional structures with a monolayer accuracy in advanced electronics, optoelectronics and other functional devices.[7-9] For example, atomic layer deposition (ALD) of metal-oxide has indeed become a crucial technique in many device-manufacturing processes but its practical exploitation has been stalled by the difficulty in quickly purging the water vapor which is used to oxidize the metal-oxide precursor.[10-13] In principle, AO is more active than water vapor and AO can be purged quickly. In the engineering design of an AO-beam generator for practical manufacturing, the flux-density of AOs should be considerably higher than 1015 atoms cm-2 s-1 in order to support the growth/etch of a monolayer-substance per second, with a broad beam covering an area of not less than 300cm2. An option of pulsing the AO-beam on and off every second further enables the precise control of monolayer-wised oxide-growth and etching per second. Such an AO-beam generator is ideal but has been generally conceived to be technically infeasible.
Historically, the development of an AO broad-beam was first quested for the facilitation of space-exploration. In brief, when space-crafts and artificial satellites operate within low Earth orbit (LEO), chemical reactions causing detrimental damage of such space-facilities occur because the presence of O2 in LEO dissociates into AO by solar ultraviolet irradiation and AO colliding with space-facilities traveling at a sonic speed is violent and reactive. More specially, the kinetic energy of such AO (about 5eV) indeed matches the molecular chemical bond strength (3-5eV) of typical materials and an AO flux of 1015 atoms cm-2 s-1 is sufficient to cause collision-induced chemical-modification of one monolayer per second.[14] As such, the research in this specialized field has fostered the development of AO-beam equipment with controllable kinetic energy up to 5eV and high throughput (1015 atoms cm-2 s-1). This, in turn, can facilitate the investigation of novel oxidation mechanisms and applications by regulating the kinetic energy and throughput of AOs.
However, the development of high-throughput AO-beam equipment is stalled by numerous technical limitations and difficulties.[14, 15] First, although AOs can be readily produced by pyrolysis, photodissociation and plasma-reactions, the kinetic energy of AOs thus produced is uncontrollable and typically in a tame region of a few meV. [16-20] Technically, hyperthermal AOs with kinetic energy up to 10eV can be generated by the well-established molecular-beam technology via high-pressure-jet acceleration of O2, with laser-dissociation of the accelerated O2.[20, 21] Practically, this costly method can typically give an AO-beam with a cross-sectional diameter not more than 1mm, and is thus suitable for fundamental research but not for engineering of industrial exploitations. In this context, industrial engineering has turned to the adoption of the electron-cyclotron-resonance (ECR) plasma technology which is known to facilitate the most intensive ionization-induced dissociation of molecules in gas phase, including the dissociation of O2 into AOs and oxygen ions.[22] Oxygen ions thus generated are extracted from the plasma and then accelerated by an electric field to regulate their kinetic energy. They are finally converted to AOs with the desirable kinetic-energy by collision and charge-transfer with a metal-plate mounted near the exit of the oxygen ion beam extracted from the ECR plasma. This technology has indeed been used to produce an AO-beam of about 5eV in kinetic energy, with a cross-sectional beam diameter of about 20cm. The AO-beam size is limited by the practical size of an ECR plasma source. This known demonstration of a practical broad AO-beam with a controllable kinetic energy has, nevertheless, the following deficiencies: (a) the AO flux density is limited by the ion flux density and even an ECR plasma can only support a maximum ion flux density of about 1x1015 ions cm-2 s-1 and a beam of about 300cm2; (b) ion neutralization by charge-transfer with a metal-plate inevitably leads to ion-sputtering and metal-contamination of the AO-beam; (c) the AO-beam must be used close to the ion-exit from the plasma in order to prevent a severe flux-density drop caused by broadening of the ion beam and AO-beam. The placement of a workpiece-specimen close to a plasma which gives strong ultraviolet radiation and excited atoms/molecules, however, risks detrimental damages of the specimen.
Here, an innovative and practical method is reported to overcome the deficiencies of this AO-beam technology. The present innovation articulates the science and technology of converting a broad ECR-based oxygen-ion beam with a flux of 1×1015 ions cm-2 s-1 to a broad AO beam with a flux of 1.5×1016 atoms cm-2s-1 via collision-induced dissociation of O2 into AOs, as shown in Figure 1 . In this AO generator, one ion injection with a kinetic energy of 300eV initiates many collision cascades in which on the average about 130eV of the 300eV injection-energy is used to drive about 24 O2dissociations each costing a dissociation energy of about 5.1eV. Since ion extraction can be easily pulsed, the resultant AO flux can also be engineered with a pulsing frequency not less than one on/off cycle per second. The objective is the facilitation of industrial exploitation of AO surface-modification and AO-assisted metal-oxide deposition, to realize the rising demands of layer-by-layer surface-modification/deposition with a controlled operation of one atom/molecular monolayer per second, with a scalable modification/deposition area of ≥1000cm2. In the following sections, the fundamental concepts of AO generation by gas-phase dissociative collisions initiated by oxygen ions extracted and accelerated from an oxygen plasma are explained by ab initio molecular dynamics (AIMD) of the kinematics in dissociative O+-O2, O-O2 and O2-O2 paired-collisions, particularly to pin-point the threshold kinetic energy required for O2dissociation in different collision-impact trajectories. With these critical attributes, direct simulation Monte Carlo (DSMC) method of the generation of a high-flux AO-beam by extracting O+ from a plasma into a low-pressure O2 atmosphere are detailed. The simulations are then translated into the engineering designs of a practical AO-beam generator, as shown in Figure 1 , with an AO flux density more than 1016 atoms cm-2s-1 and a beam ≥1000cm2 at the proximity of the generator exit, in a pulse-operation mode with a pulsing frequency as high as one on/off cycle per second. The generator comprises an array of 2-by-2 ECR plasma sources, with a total power of 800W. These functional features are demonstrated and validated. More specifically, the oxidative modification of a self-assembled molecular layer of alkyl silane on the surface of an oxidized silicon wafer is also explained. Further, the ALD of alumina with a fast rate of one-deposition cycle per 2 seconds is articulated. The supplementary information (SI) provides further elaboration on the details.
Simulations