|
Low vacuum plasma deposition of epitaxial thin films of multicomponent compounds. (Research proposal) Leonid Sakharov Introduction. The plasma deposition of the thin films of novel multicomponent compounds with special promising for the electronic properties as superconductivity, ferromagnetism, luminescence [1] est. is potentially much more productive and lower cost technique comparable with the pulse laser deposition. One of the most challenging problems of the plasma deposition is maintaining a stoichiometry at the large area of exposed substrate. A transport of material sputtered from the target to the substrate is usually described in terms of two stage process: thermalization of high energy ions as a result of colliding with the molecules of sputtering and reaction gas and then chaotic migration until reaching a solid surface where solidification is happen (in crystalline of amorphous form). The obvious approach to achieve stoichiometry transportation from the multicomponent target is a placing a substrate on the distance before thermalization had occurred. Unfortunately a secondary sputtering (etching) of the substrate materials is degraded a quality of the deposited films by polluting them with the substrate materials and producing radiation type defects. Because of different profiles in the distances from the target where thermalization for different ions are happened there will be variations for them in relative rates of reaching of solid surfaces (substrates, chamber walls, returning back to sputtering device). As result the transportation coefficients for different atoms of target has to suffer a significant variations along substrate and stoichiometry deposition of the large enough areas could be impossible in a general case without special modification of the technology scheme. In the ideal situation all thermalized atoms should be transported by outer force to the substrate.
A role of such force can stage a controlled flow of the gas in the vacuum chamber.
As a criteria of the significant influence of the gas flow on the transportation coefficients
of the thermalized atoms could be considered a ratio between velocities of gas flow and average
thermal velocity of atoms that can be calculated by formula:
where k- Boltzman constant, T - temperature, M - mass of the diffusing atom. A good estimation of the value of VT is close to sound velocity and have numerical value around 102 m/s. In the article [2] there were shown that the rate of flow of reactive gas (Ar) via the magnetron is a critical factor for achieving high quality superconductor YBCO film. The estimation of the gas velocity flown through magnetron ring made on the base of known gas pressure, square of the ring, volume of the vacuum chamber and rate of changing pressure on vacuum chamber after switching off the pump demonstrate that the critical value of Ar flow via magnetron to get quality superconductor films is indeed in good agreement with this value - 102 m/s. This qualitative observation demands systematic experimental and theoretical exploration to be scaled up to the design industrial installations level. The experimental scheme. In the proposed research the effect of controlled gas flow on the transportation of the sputtered ions is proposed to be investigated on the quantity level. The plasma sputtering system such as a magnetron or other with similar characteristics will be modified to permit a reliable control and registration of the gas characteristics in the area of materials transportation from target to the substrate. Two principal different geometry of controlled gas supply are planed to be investigated. In first one the sputtering gas (Ar) will be flow out directly from the magnetron area in the direction of spattered from the target material. In the other scheme additional gas source perpendicular to the sputtering direction will be added. A distribution of the transportation coefficients of sputtered materials along the working space will be registered by the measurement of thickness and chemical composition of the films deposited on the substrates. The result of one experiment will be a distribution of rate of deposition of given atoms along the substrate for the given condition of experiment: voltage of the plasma system, pressure, temperature, substrate position and rate of gas flow. Varying only rate of gas flow the influence of this parameter on the transportation coefficient can be revealed. Here is worth to note that in the spite of direct correlation of the gas flow rate with the pressure in the vacuum chamber it's possible to vary only gas flow under constant pressure. To do this a second independent source of reactive gas has to be set in reactive chamber far enough from experimental zone. For the given productivity of vacuums pump and stable pressure in the chamber the cumulative gas flow from two sources is a constant. Setting the same pressure in two steps using gas source from magnetron at first and then adding pressure to the constant for given the experimental series will permit to produce a series of deposited films with one controlled variable -gas flow rate in direction of the deposition. A priory one can presume that there has to be an extreme value of gas flow when additional increasing will not lead to noticeable change in transition coefficients of the sputtering material. From practical point of view a determination of this parameter for different chemical components can be most important result of the research to design an industrial scale plasma deposition installation. A temperature of the gas in the area around substrate surface can be also significant factor that could influence of the transition coefficients of different elements atoms. A forming of crystalline structure for most high temperature compounds is taken place in the interval 800 - 1000oC. Comparable with room temperature (or about) near plasma source average velocity of gas molecules can be estimated as 50% higher in area of heated substrate that is significant enough factor to plan special experiments to compare deposition on the substrates under different temperatures. A of high temperature vacuum heaters with in sintered in alumina ceramic near its surface Pt wires [3] with working temperature up to 1200oC could be useful element of experimental procedure. A theoretical modeling. A process of material transportation of sputtering material from target to the substrate can be numerical simulated on the base of the model of random collisions with the molecules of sputtering and reactive gases. Because of low relative pressure sputtered atoms it's reasonable to neglect their mutual interactions that significant simplify a model as a whole. For example for the case when geometry of the deposition can be approximated as a two infinite parallel surfaces target and substrate with specific figure of bombardment areas it's possible to reduce a volume of calculation by presuming that the final distribution of the atoms on the substrate will be integral superposition of number dot-lile sources represented sputtering area. In this case the Monte-Carlo simulation will be needed to apply only for the modeling of one dot like source of deposited material. In the spite of clear fundamentals laying under model of thermalization of high energy atoms spattered from the target and their diffusion migration to the solid surface a numerical modeling of this processes is not a trivial task as soon there is not exist a simple routine for calculation of result of collision of two atoms (molecules). The quasimechanical approach that takes into account random distribution of angles collided atoms used in [2] gave reasonable an agreement with experiment results. One of the important parameter of this model, a value of effective cross-section of atoms different elements, has to be defined from the experiment. The same problem is exist with the parameter of the distribution of the sputtered atoms along energy parameter. After achieving of the good quantitative correlation between experimental results of the distribution of atoms different elements along substrate with numerical modeling the virtual testing of results different configurations of plasma sources and additional gas ejectors supposed to be performed with the goal of finding the configuration allowing a deposition of the stoichiometrical content on the large area of a substrate. One of the prospect configurations could be a facet type multiple magnetrons surface with sputtered gas ejected in the direction of deposition that will be maintained on the platform moving with a amplitude that is more then distance between magnetrons. Analytical methods. An optical measurement of the thickness of films on the substrate can be most operative methods of determining rate of deposition in case of monocomponent (metal) targets. For the multicomponent deposition thickness of films will give an integral rate of the transport. An automated thickness mapping systems are standard equipment in the semiconductor industry [2, 3] can be utilized for this goal. A distribution of the chemical composition of the deposited films can be measured by the X-ray microprobe analysis and FRIR (Fourier transform infrared systems) [4]. As more time consuming methods a full profile mapping of concentrations on the substrate will be planned only for verification of the model developed on the base monocomponent deposition. Applications. A results of experimental data and numerical modeling has to improve quality of deposited films of multicomponent composition. In the proposed project the developed models will be applied to accelerate search of optimum conditions for the deposition of epitaxial films of superconductor oxides and photo-luminescent compounds. Literature:
|