How to Optimize E-Beam Use.
From the 1950s when e-beam evaporation systems became commercially available, and through the 1970s, much time and effort was spent on solving the inherent design problems that e-beam sources experienced. For example:
* The e-beam emitter would quickly become coated from being too close to the evaporant and electrically short due to the 180 [degrees] beam-deflection system.
* Positive ions formed at the evaporant surface accelerated back toward the emitter, bombarding the emitter and drastically reducing its operational lifetime. This situation also caused the runaway phenomena of the emission current of the early e-beam source power supply designs due to ion current flow.
After overcoming the above problems to a certain degree of satisfaction with the advent of the 270 [degrees] beam deflection, e-beam coating technology moved rapidly ahead with the onset of emerging technologies-microelectronics, optics, and vacuum metallization, just to name a few.
The 1980s witnessed a lull in e-beam technology advances. The major design deficiencies of most e-beam Sources that exist today are many and are heavily weighed. Two of the most predominant are beam curl and inconsistent beam density. Beam curl is the result of having an insufficiently designed magnetic field.
A charged particle, an electron in this case, in motion can be deflected by a magnetic field. The magnetic field in e-beam sources is designed to focus a beam of electrons while deflecting it along a 270 [degrees] arc, enabling the beam to diverge into the crucible. The curvature radius of the beam's trajectory is completely dependent on the acceleration voltage and shape and strength of the magnetic field. The ideal e-beam source should provide an impact of the beam onto the top surface of the crucible at a 90 [degrees] angle. As the beam impacts the crucible materials' surface at 90 [degrees], maximum energy transformation results, yielding maximum efficiency. The deficiencies of the magnetic field are such that once the beam enters the crucible area, it continues its arc rather than taking on the desired vertical trajectory.
An e-beam source that exhibits beam curl will cause the emitted vapor from the crucible to shift over time--a period as short as three or four layers of coating. The end result is a finished coating with unpredictable, unrepeatable, and variable thickness on a given substrate.
The typical e-beam source has a very inefficient magnetic system. The traditional e-beam source magnetic system causes the electron to travel through multiple focal points. During beam sweeping, the e-beam's power density changes unpredictably from crucible edge to crucible edge. For example, in the center of the crucible, the beam is normally triangle-shaped. In the 12, 3, 6, and 9 o'clock positions, however, one can visually observe shapes ranging from exaggerated "flattened triangles" to "crescent-moon" shapes. In terms of coating process limitations that are caused by everchanging beam power densities on the crucible, the end user will experience non-uniform temperatures across the evaporant material. Non-uniform heating of dielectric (subliming) materials will cause the e-beam to focus on one point of the material. The material is eroded only locally, forming a tunnel.
Tunneling severely limits the vapor distribution. In addition, because of this behavior, material utilization is very poor. In multi-crucible or carousel-type crucibles, as much as 60% of the material may be wasted simply because the beam-density differences within the crucible are too great to overcome. Varying deposition rates across the crucible is another bothersome symptom. Some evaporant materials, such as hafnium dioxide (Hf[O.sub.2]) are susceptible to "spitting" because of inconsistent surface temperature profiles. Thus, alternate methods of vacuum depositing these evaporant materials were developed and incorporated.
The future in coating technology lies in improved process control and innovative monitoring techniques. E-beam sources and related control electronics designed with the end user's problems in mind (greater control and versatility), makes advanced monitoring more or less straightforward and less of a guessing game.
When thin-film designers and coating process engineers design and develop new coatings, there are many process variables that must be accounted for. Coating uniformity, soak times, process gas mixtures and partial pressures, deposition rate, substrate temperatures, and beam sweep patterns are just a few of the many variables that worry the end user who obviously desires precise and repeatable coatings run to run.
An ideal e-beam source would be one that exhibited process-favorable beam characteristics and control electronics that allows flexibility, precision, and speed to handle current high-performance coatings.
The optical coating needs for telecommunication systems and high-power lasers are at their greatest and most demanding. In telecommunications, optical filters and narrow bandpass filters for wavelength division multiplexing (WDM) call for stacked arrays of up to 100 tightly specified coating layers. These layers typically take, on average, 20 hrs to complete. Because of the prolonged coating runs, only the latest in e-beam deposition technology can perform these coatings, which are mainly oxide materials, repeatedly.
For example, many optical coaters now specify coating uniformity requirements unheard of 2 years ago. By holding the emitted vapor distribution consistent over 20 hrs, plus or minus tenths percent uniformity is now possible over typical substrate sizes. First, the e-beam source magnetic field is manipulated locally; that forces the electron beam to travel vertically through the crucible, thereby eliminating the beam curl condition. Second, a magnetic field structure conducive to avoiding "hot spots" in the crucible is incorporated. High-peak power laser systems use mirror coatings extensively. Laser power output capability is determined by the damage thresholds in the coatings. Catastrophic failures at these power levels will take place in the coatings if the film is applied incorrectly.
The optical coatings referred to here are alternately applied high- and low-refractive index materials, such as Hf[O.sub.2] and silicon dioxide (Si[O.sub.2]). Hf[O.sub.2] was the material of choice since out of all of the high-index materials available, it has the highest damage threshold characteristics. Hf[O.sub.2] and Si[O.sub.2] are the two most difficult materials to evaporate out of an e-beam deposition source. Each presents particular difficulties.
The e-beam will form a tunnel through Hf[O.sub.2] if not heated uniformly. Furthermore, columnar-like growths or protrusions of Hf[O.sub.2] material will occur along the cooled crucible walls. This growth condition will prevent multi-pocket e-beam sources from rotating from pocket to pocket. The coating defects that cause the Hf[O.sub.2] films to fail are caused by ejected microscopic nodules or spitting from the crucible. The cause of this ejecta phenomena has been narrowed down to the stress relieving of the Hf[O.sub.2] material itself. As the beam sweeps across the crucible, the Hf[O.sub.2] undergoes a temperature-induced phase transition and spits occur upon each beam sweep cycle. To reduce the number of spitting events and hence the nodular defect growth, the phase transitions must be minimized. The method used lies in the use of high-frequency e-beam sweeping.
Time-variable deflection of the e-beam refers to a mechanism allowing the e-beam impact point to be varied over a period of time. In other words, the beam sweeps over the surface of the crucible. This sweeping action can be accomplished by placing electromagnetic coils around the perimeter of the crucible to produce time-variable magnetic fields. High-frequency sweeping on an e-beam source has matured to the point of being able to control the "softness" of the beam. Sweeping at frequencies around 200 Hz has allowed the user to electronically defocus the beam for processes that require "soft beams," such as Hf[O.sub.2] and Si[O.sub.2].
A useful low index quality, good abrasion resistance, favorable packing density, and optical clarity make Si[O.sub.2] a valuable film tool among optical coating designers. But its optimal use is severely limited by tunneling, vapor distribution shifts over time, and poor material utilization. In the past, both granular and solid Si[O.sub.2] disk forms have been evaporated, but to no avail. Some e-beam users have gone as far as evaporating straight silicon-monoxide in an oxygen background to achieve satisfactory films. The difficulty of evaporating Si[O.sub.2] lies in the fact that it is an insulating material and needs to be treated much differently than, say, a metal would be. Precise beam-control capability and thorough understanding of beam power densities as it relates to insulating materials must be fully understood before attempting to successfully evaporate Si[O.sub.2] in any shape or form.
A unique sweep pattern was developed to evaporate solid 15.24-cm dia Si[O.sub.2] disks. The purpose was to evaporate 75% of the disk so the erosion profile is flat and unchanging throughout the life of the disk. This technique buys two distinct advantages: First, one doesn't have to rely on the coating operator's experience and ability in Si[O.sub.2] evaporation, and second, stable and predictable vapor distributions can now be achieved from Si[O.sub.2]. A whole new world of possibilities and Si[O.sub.2] applications will open up because of this newfound capability.
WEB RESOURCES FOR E-BEAMS: www.mdcvacuum.com www.isi-seal.com www.svc.org
Harris is the e-Vap group manager at MDC Vacuum Products.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||electron beam deposition|
|Comment:||How to Optimize E-Beam Use.(electron beam deposition)|
|Publication:||R & D|
|Article Type:||Brief Article|
|Date:||Jun 1, 2001|
|Previous Article:||How to Seal Using Glass-Ceramics.|
|Next Article:||How to Disperse Particulates.|