Filmetrics F40-UV Microscope-Mounted

Filmetrics F40-UV Microscope-Mounted Tool Type Inspection, Test and Characterization Location Bay #4 Supervisor Ning Cao Supervisor Phone (805) 893-4689 Supervisor E-Mail ningcao@ece.ucsb.edu Description Microscope-mounted optical reflectometry Manufacturer Filmetrics Inc. Model F40-UV

About

The Filmetrics F40-UV is a microscope-mounted thin-film measurement system, allowing you to non-destructively measure thin-film thicknesses in small (patterned) areas on your sample. The microscope is a standard Olympus BHJML metallurgical trinocular microscope. It is an optical reflectometer, acquiring reflection spectra between 400-900nm optical wavelengths (Vis to Near-IR) with a regular halogen microscope light source. The Filmetrics software then performs curve-fitting to determine the thickness and/or refractive index of the measured films.

Capabilities

• Measure optically transparent thin-films down to ~30nm thickness.
• Microscope Objectives: 10x, 20x, 50x, 100x, 150x
• Acquire Optical Reflection Spectra from 400-900nm
• Spectrometer/Detector is capable of detecting down to UV ~190nm, but light source does not support this wavelength.
• Reflectivity curve-fitting for thin-film thickness analysis, supporting many common materials (Si₃N₄, SiO₂ dielectrics, Si, GaAs, InP semiconductors, metals, photoresists etc.)

Operating Procedures

• Quick Start Procedure

Examples

Good Curve-Fitting

Do not believe your thickness measurements unless the red/blue curves show a reasonably good match!

Screenshot showing 10x microscope view and optical spectrum/curve fitting of the SiO2 film thickness on top of Platinum contact metal.

A good fit should show the Calculated (red) curve overlay on top of the Measureed (blue) curve.

Checking whether a dry etch is complete

The F40-UV is very useful for measuring whether a thin-film has been completely removed during a dry etch. This is similar to using laser monitoring during the etch, except that the microscope enables you to measure inside small patterned areas that a laser monitor spot may not fit inside.

In the following example, we are trying to etch away the SiO₂ from on top of a Platinum contact metal. The 20x objective allows us to measure the SiO2 thickness on top of the Platinum as we continue to etch in 1 minute increments, measuring in between etches.

Finally, when the SiO2 has been fully removed, we see that we can't get a good fit between the measured (blue) and simulated (red) data.

Since the Platinum layer is thick enough to be opaque to visible light (>50nm), we just modelled this as SiO2 on top of Platinum, ignoring any other materials below the platinum.

Incomplete Dry-Etch, showing some residual SiO2 left. "Goodness of Fit" is close to 1.0 in the results pane, and the red+blue curves match closely so we believe the measurement.	After re-etching, measurement shows bad fit (red and blue curves), and "Goodness of Fit" = 0 in the results pane. The thickness value shown is not real, and this indicates that the SiO2 film has been fully removed since no parameter can give a good fit.