Electron Beam and Physical Vapor Deposition

Plasmaterials

Plasmaterials, Inc. is a leader in providing high purity materials for all types of thin film applications. We produce and market a full range of products for R&D and full scale production.

In the last BLOG article I described the simplest form of resistance evaporation. That involved heating a resistance filament in a vacuum environment which in turn heats an evaporant material that is either melted or sublimated. The vapor created in the process is then condensed onto a substrate, thus creating a resultant film.

Here I would like to discuss a slightly different form of Physical Vapor Deposition (PVD) in which a target anode material is bombarded by a stream of electrons that are generated from a tungsten filament. This is referred to as Electron Beam Deposition and is a widely used technology for depositing thin films in both the Research & Development community as well as for large scale production applications.

Normal Electron Beam deposition is done in a vacuum, usually initiating the process at levels less than 10-5 Torr. Once a suitable vacuum level is achieved a stream of electrons are emitted from a tungsten filament contained within the Electron Beam Source. This beam of electrons can be generated by a variety of means, including thermionic emission, field electron emission or ionic arc source, depending on the design of the source and the associated power supply. In all cases, the negatively charged electrons are attracted toward the positively charged anode material. The electron beam that is generated is accelerated to a high kinetic energy and aimed directly toward the material which is to be deposited onto a substrate. This energy is converted to heat through an atomic interaction with the evaporant material. To facilitate uniform heat distribution across the surface of the material being heated, a series of electro magnets can be employed in the Electron Beam Source to enable the operator to raster the beam across the surface in a uniformly distributed pattern. Rastering is generally controlled with a process controller so an infinite array of patterns can be generated to best accommodate the geometrical constraints associated with the deposition system as well as with the physical characteristics of the material being deposited.

For applications where multiple layers of resultant films are required, Electron Beam sources may be configured with multiple hearths. These are arranged in a turret design whereby each hearth can be mechanically rotated into position beneath the incoming stream of electrons. Commercially available sources are available with two, four or even six individual hearths of different sizes to accommodate a wide array of deposition criteria, geometrical configurations and resultant film variations.

Hearth sizes can vary in volume based on process requirements. For small Research and Development applications hearth sizes as small as one or two cc’s are available. For large scale production requirements, hearth sizes can go up to 150cc and larger. For continuous process applications, a wire feed system can be configured whereby a spool of wire is unwound into the hearth at a specified feed rate to keep the volume of molten material in the hearth continuously constant. A similar design is also available with a rod feed mechanism continually advancing the length of the rod material into the bottom of the Electron Beam source to constantly “refill” the hearth.

As mentioned above, the purpose of the stream of electrons being generated in the Electron Beam Source is to heat the material being deposited to a temperature above the threshold of the vapor pressure at a given background pressure. The vapor stream is then condensed onto the surface of the substrate. Rastering the beam helps to maintain a uniform temperature within the charge but there is still the issue of water cooling through the Electron Beam Source body. The source is cooled internally to minimize any radiant heat to the substrate and to protect the copper hearth from reacting with the molten source material. Unfortunately the water cooling around the perimeter and bottom of the hearth cools the source around the outer surface of the melt while the bombarding electrons are heating the source material on the top surface only. This creates a thermal gradient throughout the charge – hotter in the central portion and cooler around the outer edges. Using a crucible liner of suitable composition will help to equilibrate the temperature within the charge by isolating the molten material from the water cooling of the Electron Beam Source. Additionally, by choosing the correct liner material, the molten material will not react with the copper hearth causing contamination to the resultant films and damage to the source. For help in selecting the proper size and composition of a suitable liner for any given deposition material, please contact your Plasmaterials, Inc. Sales Engineer at in**@pl***********.com for advice.

There are some serious shortcomings associated with Electron Beam Evaporation however. It is essentially a point source, meaning that the vapor stream is initiated from a limited point source or surface area. Depending on the actual hearth size of the Electron Beam Source, the entire vapor stream is initiated within the pocket of the source. Depending on how the beam is rastered above the evaporant material, the material stream will form a conical cosine distribution from the top of the hearth outward toward the substrate. To help facilitate film thickness uniformity the substrate-to-source distance is usually quite a bit further than that of a sputtering source. The typical substrate-to -source distance is around 15” or so, but may depend on the geometrical restraints contained within a specific system. For large area substrates, it may be necessary to utilize more than one Electron Beam Source simultaneously. In these instances, multiple sources are used simultaneously with individual charges of the same deposition material. The sources are usually set in a line with the separation distances between the hearths set in such a way as to maximize uniformity of the resultant film. This configuration is quite often used for roll coaters where a continuous substrate or a web is passed in a line of site over the array of Electron Beam Sources. Care must be taken to work out the proper spacing as well as the deposition parameters to achieve consistently accurate and uniform film thicknesses.

Unlike sputtering, where the individual atoms are arriving at the substrate surface a very high velocity and momentum, a thermally created stream of vapor arrives at the substrate surface at a considerably lower velocity, but at a much greater arrival rate, i.e. the deposition rate for thermally evaporated material can be several orders of magnitude greater than sputter deposition rates. This is very beneficial for high volume production or thick film requirements but it does have some drawbacks. Because the kinetic energy of the arriving species during Electron Beam evaporation does not have the same momentum as sputtered species, the material tends to condense directly on the substrate surface. Atoms of sputtered material tend to penetrate into the substrate surface several atomic layers (or more) before losing momentum and then building up cohesive bonds within the nucleating structure and thin film growth. Sputtered films tend to provide better adhesive characteristics than thermally evaporated materials.

Like all thermal evaporation systems, the rate of deposition is dependent on the temperature of the material being deposited and the vapor pressure, a physical constant, of that material. For elemental materials there is a fixed vapor pressure for any specific background pressure (vacuum level) and material temperature. For alloys or composite materials though, there may be varying partial pressures associated with each constituent. This results in varying stoichiometries of the resultant films as one or more of the constituents being deposited are depleted differentially from the source material. Unless the material composition is a eutectic composition this could be a problem when trying to deposit a fixed composition over an extended period of time.