ALD Al2O3 thickness profile in microscopic rectangular channel, TMA-water 300 degC 500 cycles, PillarHall(TM) LHAR3-1b, V0001



Thickness profile data measured for aluminium oxide thin film grown by atomic layer deposition (ALD) using trimethylaluminium and water as reactants at 300 degrees Celsius on an all-silicon lateral high-aspect-ratio (LHAR) test structure of type PillarHall(TM) LHAR3 (layout v1b), nominal gap height 500 nm. Part of this data has been used in the publication "Saturation profile based conformality analysis for atomic layer deposition: aluminum oxide in lateral high-aspect-ratio channels", where this sample was number 8 (unique manufacturing code V0001, see Table S3 of the supplementary info, which also contains details of the layouts:

Thickness profiles have been measured optically after removal of the top membrane by adhesive tape, by spectroscopic reflectometry in the visible wavelength range. In the optical thickness measurements, the refractive index is assumed constant (refractive index vs. wavelength this has been determined earlier for similar films also grown at 300 degrees Celsius) and thickness is allowed to vary. The optical model consists of smooth single-crystal silicon and ALD aluminium oxide films. Any silicon oxide below the ALD film is counted in the aluminium oxide thickness. Also roughness will impact the measurement result. For each measurement point, the location, the fitted thickness, and the goodness of fit is presented. For the 50x magnification, measurement spot diameter size is estimated as about 5-6 micrometers. The raw data have the centre of measurement as zero point of horizontal axis. The data needs to be shifted horizontally to have the zero point at the beginning of the LHAR cavity. Sometimes, the removal of the membrane has not worked in the beginning of the cavity (where the support pillars in the PillarHall LHAR3 design are more densely located) and the measurement has been made on a place that includes membrane remnants. In such cases, the fit of the optical model to the results has been poor and the data points should be rejected.
Date made available31 May 2021

Dataset Licences

  • CC-BY-4.0

Cite this