Skip to content

Development of Minimal RF-magnetron Sputtering Machine

MetadataDetails
Publication Date2023-01-01
AuthorsSatoshi Fujii, Yuki Odo, Hiroshi Nishizato
InstitutionsHoriba (Japan), Toyohashi University of Technology
AnalysisFull AI Review Included

This analysis details the development and optimization of a minimal RF-magnetron sputtering system designed for depositing piezoelectric Aluminum Nitride (AlN) and Scandium Aluminum Nitride (ScAlN) thin films, critical for elastic wave devices.

  • System Footprint: A minimal RF-magnetron sputtering system was successfully integrated into a standard minimal fab tool enclosure (144 cm H x 30 cm W x 45 cm D), achieving a base vacuum of approximately 1 x 10-5 Pa.
  • Film Quality: The optimized process yielded high-quality, c-axis oriented AlN thin films on Si(001) substrates.
  • Crystallinity Metric: The best conditions resulted in a c-axis rocking curve Full Width at Half Maximum (FWHM) of 2.66°, sufficient for piezoelectric applications.
  • Uniformity: The deposited AlN film exhibited excellent thickness uniformity, measured at ± 0.95% across the substrate.
  • Device Application: A 5 GHz diamond Surface Acoustic Wave (SAW) one-port resonator was fabricated using the deposited AlN film on a single-crystal diamond substrate.
  • Device Performance: The prototype SAW resonator achieved a resonant frequency of 5.2 GHz and a Q-value of 1737.
  • Limitation Identified: The measured electromechanical coupling coefficient (K2) of 0.89% was 74% lower than the theoretical value, attributed primarily to insufficient AlN orientation and defects in electrode formation, suggesting the need for a dedicated substrate heating mechanism.
ParameterValueUnitContext
System Footprint (H x W x D)144 x 30 x 45cmMinimal fab tool enclosure size
Achievable Vacuum1 x 10-5PaBase pressure capability
Optimized RF Power100WAlN deposition condition
Optimized Process Pressure3mTorrAlN deposition condition
Optimized Gas Mixture30% N2 / 70% Ar%AlN deposition gas ratio
Substrate Temperature210°CAchieved via plasma heating (no dedicated heater)
AlN Film Thickness (Optimized)587.5nmMeasured center value on Si(001)
Thickness Uniformity± 0.95%Across the deposited film
AlN Crystallinity (FWHM)2.66degreesc-axis rocking curve on Si(001)
SAW Resonant Frequency5.2GHzDiamond/AlN one-port resonator
SAW Q-value1737-Measured performance
SAW K2 (Measured)0.89%Electromechanical coupling coefficient
SAW K2 (Theoretical)1.2%Theoretical value for 5 GHz design
AlN Film Thickness (SAW Device)0.70”mDesign parameter for 5 GHz SAW
SAW Wavelength (λ)2.0”mDesign parameter
Al Electrode Thickness (SAW Device)90nmDesign parameter

The minimal RF-magnetron sputtering process was optimized by evaluating film thickness distribution and c-axis crystallinity.

  1. System Configuration: A minimal fab tool enclosure was equipped with an aluminum chamber, turbo molecular pump, RF cathode (13.56 MHz), Q-mass, and a Particle-Lock Airtight Docking (PLAD) wafer transfer system for 0.5-inch wafers.
  2. Target and Gas: A one-inch pure 4N Al or Al-Sc alloy target was used in a constant-pressure environment with an Argon (Ar) and Nitrogen (N2) gas mixture.
  3. Optimization Parameters: The study varied three primary parameters: RF power (30-100 W), process pressure (3-10 mTorr), and target-to-substrate distance (TSD) (30-50 mm range for optimization, 730 mm used in final recipe—likely a typo in the source table, but used as reported).
  4. Thickness Evaluation: Film thickness distribution was measured using an interferometric film thickness meter. Uniformity was defined as: Uniformity = (Tmax - Tmin) / Taverage x 100 (%).
  5. Crystallinity Evaluation: Crystallinity and c-axis orientation were assessed using X-ray diffraction (XRD) 2Ξ-ω scans and rocking curve measurements.
  6. Optimized Recipe: The final optimized conditions for high-quality AlN were: 100 W RF power, 3 mTorr pressure (30% N2/Ar), and a substrate temperature of 210 °C (achieved through plasma heating).
  7. Device Fabrication: A 0.70 ”m AlN film was deposited on a single-crystal diamond (100) substrate. Aluminum (Al) Interdigital Transducers (IDTs) were patterned using Electron Beam (EB) lithography to form the 5 GHz one-port SAW resonator.

The minimal RF-magnetron sputtering system and the resulting high-quality AlN/ScAlN films are critical for several high-growth technology sectors:

  • 5G and 6G Communication: Essential for fabricating high-performance filters (SAW and FBAR) required for high-frequency operation in the millimeter wave band (e.g., 28 GHz).
  • RF-MEMS Devices: AlN and ScAlN piezoelectric thin films are core materials for next-generation radio frequency micro-electromechanical systems.
  • High-Velocity Resonators: The use of diamond substrates combined with AlN enables the creation of ultra-high frequency, high Q-factor resonators, suitable for demanding communication and timing applications.
  • Minimal Fab Ecosystem: Supports the production of a wide variety of electronic and semiconductor components in small quantities, facilitating rapid research, prototyping, and low-volume manufacturing.
  • Energy Harvesting: The piezoelectric properties of AlN/ScAlN films are applicable to general MEMS devices, including environmental power generation systems.
View Original Abstract

A minimal RF-magnetron sputtering system was developed for AlN and ScAlN thin film deposition in elastic wave devices, and the sputtering conditions for this system were optimized. X-ray diffraction measurement results after optimization showed that the c-axis rocking curve of thin AlN films on Si(001) substrates had an FWHM of 2.66 under the best conditions, and the thickness distribution measured using optical thickness measurement was less than 1 %. A diamond surface acoustic wave (SAW) resonator with an electrode of 2 ÎŒm wavelength (0.5 ÎŒm comb electrode width) was fabricated by depositing a 700-nm thick AlN film on a single-crystal diamond (100) substrate using this apparatus. The measured resonance characteristics yielded a resonant frequency of 5.2 GHz, an electromechanical coupling coefficient (K 2 ) of 0.89 %, and a Q-value of 1737. However, the K 2 value was 74% smaller than the theoretically obtained one. This deviation from the theoretical value can be attributed to the insufficient orientation of AlN as well as defects in the SAW electrode formation, as several anti-resonance peaks were observed. However, these can be improved by adding a substrate heating mechanism. The newly developed system exhibits satisfactory performance as a minimal RF sputtering device.