Manufacturing Critical Optical Coating Materials:
Silicon and Its Oxide
Within larger industries such as metals, semiconductor and solar, specialty optics thin film engineering is a small but discriminating piece. Whether through sputtering or evaporation processes, engineers strive to optimize thin film properties at the most reasonable cost. In comparison to other metals, Silicon has complex origins and remains fragile and temperamental. Beyond the base material challenges, deposition of high quality PVD SiO2 thin films, whether by reactive sputtering or e-beam evaporation of the oxide, is further constrained by limitations and defects consistent with those techniques. Engineers often focus on predominant limiting factors such as arching, spitting or post-deposition annealing, but other upstream risks often go unnoticed. If only due to their enticing lower cost or their proximity to the bigger semiconductor and solar markets, Silicon and Silicon Oxide can be seen as materials that require more detailed specifications for intense photonics.
The electron beam process has the advantage of offering flexibility in depositing a host of even nominally stable materials. Whereas, the reactive sputtering process offers superior control of growing film while minimizing the risk of sub-stoichiometric compounds, particles and spitting, although at a compromised rate. Outside of the DUV and M/LWIR, oxides dominate the materials landscape due to the large number of compounds they make possible and the diverse manufacturing processes employed to produce a host of products covering numerous platforms. In all cases, it is possible to over-engineer the source material while trying to identify the critical fit for the deposition technology and application. The following are some examples of important upstream processes with downstream influence.
Pure and doped Silicon is critical for reactive sputtering and IR optics. Driven by VLSI (integrated circuit) and solar markets, two of the numerous processes include crystal pulling and recrystallization techniques. Figure 1. The Czochralski method produces a single crystal pure Si round ingot by placing a seed crystal on the bottom of a vertical arm barely in contact with molten Si in the crucible. The arm is raised slowly as the crystal grows underneath at the rotating interface between the crystal and the melt. The Bridgman–Stockbarger technique grows a large single crystal ingot by slowly lowering smaller polycrystalline material in a sealed cylindrical ampoule (conical lower end) into molten Silicon. To enhance sputtering behavior, Boron is added to the feedstock in a boule or crucible which is pulled or directionally solidified into a larger polycrystalline ingot. Float Zone Silicon transforms a polysilicon rod into a larger diameter single crystal ingot by passing the rod through a RF coil annulus.
Figure 2. High purity Si powder can come from silane (SiH4) in a fluidized bed reactor process or polycrystalline rod as shown in Figure 3, from the CVD Siemens process (HSiCl3).
Intuitively, you can imagine each different process has slightly different amounts of resident oxygen or hydrogen from fabrication that can be of critical importance to radiation stability of thin films or arcs during reactive deposition. Figure 4. Shows how dramatically different these two CVD processes are when compared to classical crystal pulling. In addition to concerns about process contaminants, there are real physical dangers to targets from water pressure, bonding, and backing plate composition. The general fragility of silicon, compared to other metals, can result in formidable production challenges beyond these complex details that are upstream of the targets or evaporation materials.
What pure silicon is to reactive sputtering, SiO2 is to electron beam deposition. The low index of choice for UV, VIS and NIR, it has the amazing capacity to remain SiO2 before, during and after deposition. The downside of this ability is its rather limited melting characteristics, which curtail the rate and stability of a pocket of granules. Substrate transmission, SiO2 particle ejection, hydroxyl radicals absorption and radiation hardness are perhaps the only instances where the rather ubiquitous dry CVD SiO2 or natural quartz (Figure 5), doesn’t meet the needs of specialty optics. In these cases, fused silica is adequate but may further challenge the rate and stability of the pocket. During the Siemens process, for every amount of Si produced, 4x that amount of the SiO2 precursor (SiCl4) is formed. It is then reacted with water, acids and bases to produce high quality SiO2 materials of different grades. For substrates and sputtering targets, natural quartz or CVD SiO2 are melted (fused) into amorphous glass (super cooled liquid). For the best radiation hardening, vacuum outgassing and hydrogen absorption lines, natural silica quartz is drawn as a conventional lamp tube and crushed. For sputtering targets and windows, the precursors, distillation and melting technology are employed to deliver custom DUV, UV and NIR performance. In recent years, the finite details of these materials have been increasingly important for SiO2 thin films. Electron beam processes tend to be the most forgiving for deposits to a certain point – limited by the density of the film and the stability of the coating run.
Navigating the challenges of deposition processes begins with how the targets and evaporation materials are manufactured. The techniques themselves can be exploited to benefit process control and adaptability of platforms to provide consistent and valuable options for photonics. Success in the divergent and certainly complex fields of communication, mobile electronics, remote sensing and medical device markets requires synergy between materials, thin film tool suppliers and thin film engineers around the world.
At Materion, we offer advanced materials, deposition processes and technological capability to resolve these challenges. Contact us today to put Materion's expertise to work on a solution for your optical coating material challenges.