Reactive Sputtering Process Stability Reinforced by Advanced Energy’s Power Delivery Technology

Fig. 1. Schematic diagram of industrial Drum Coater equipped with two cylindrical TiOx cathodes (left) powered with Ascent AMS and Ascent DMS power supplies (right). The purple color represents the argon plasma emission. An oscilloscope was used to capture instantaneous I-V readouts.

Where 𝑝O2 and 𝑝H2O are the partial pressure of the oxygen and water vapor, respectively. This parameter merges partial pressures of both incorporated gases into a single parameter, defined as a weighted combination of the oxygen and water partial pressures.

Fig. 2. Evolution of water vapor and oxygen partial pressures during deposition of TiO2 coatings. The vertical dashed lines represent the beginning and end of sample exposure time to the cathodes.

Controlling Deposition Rates of TiO2 Films

Fig. 3a shows the dynamic deposition rate (DDR) of TiO2 films deposited at 24 kW and a constant frequency of 20 kHz as a function of O2 flow. A slight increase in O2 flow resulted in a significant reduction in deposition rate, however, the data exhibited considerable scatter of up to 20% for nominally identical flow conditions. When the time-averaged O2 partial pressure during deposition was used instead of flow rate (Fig. 3b), the variability decreased approximately 10% for the same oxygen pressure. Incorporating the dimensionless parameter p* further consolidated the data, yielding a highly monotonic and predictable trend in DDR, (see Fig. 3c) with a linear fit indicating a decrease of 5±1 𝑛𝑚.𝑚𝑚𝑖𝑛⁄ per unit of 𝑝∗. This demonstrates that even moderate water desorption from chamber surface can noticeably shift the reactive sputtering operating point. This shift may lead to run‑to‑run variability in deposition rate, film density, and stoichiometry, directly impacting coating uniformity and production yield.

Fig. 3. Dynamic deposition rate of titanium dioxides films as a function of oxygen flow (a), oxygen partial pressure (b) and dimensionless pressure parameter (c). Absorbing films are shown in orange dots.

Spectroscopic ellipsometry was used to evaluate the optical properties of the deposited films. As shown in Fig. 4, the refractive index of the TiO2 films exhibited a clear and systematic dependence on p*, highlighting the strong influence of sputtering pressure conditions on film density and optical quality.

At lower p* values of approximately 1.3, the refractive index at a wavelength of 550 nm reached its maximum, exceeding 2.51, while films showed slight absorption in the visible range. This could be attributed to the presence of sub-stoichiometric TiOx or increased defect density associated with energetic low-pressure sputtering conditions. As 𝑝∗ increased, n decreased monotonically, which was in accordance with lower kinetic energy of arriving species at higher pressure, promoting a less dense microstructure with lower polarizability. All TiO2 films had thickness of approximately 290 nm, indicating that optimization of p* can enhance optical performance without the need for thickness compensation. The highest refractive index for transparent films in this study, n = 2.498, was obtained at 𝑝∗≈ 1.54.

Fig. 4. Refractive index of TiO2 thin films deposited at 20 kHz as a function of dimensionless pressure parameter p*. Absorbing films are shown in orange dots.

• Residual water released during deposition is a critical variable in batch coaters and if not properly managed, can shift the effective reactive operating point.
• Oxygen partial pressure is a dominant parameter for DDR, but water cannot be ignored. Utilizing AE’s Ascent AMS and DMS power supplies help to suppress plasma sensitivity to these impurities, enabling stable deposition conditions.
• The p* parameter provides a robust way to correlate reactive gas effects into a single predictive metric.
• AE’s pulsed-DC power system assists the user to achieve reproducible results even under unavoidable fluctuating reactive conditions.