Measurement Anomaly of Step Width in Calibration Grating using Atomic Force Microscopy

TL;DR

This study investigates how AFM scan parameters affect the measurement of step width and height on a silicon grating, revealing a measurement anomaly linked to pixel pitch and proposing a minimum pixel count for accuracy.

Contribution

It identifies the pixel pitch-dependent measurement anomaly in AFM and establishes a minimum pixel count to improve measurement accuracy of step features.

Findings

AFM measurements vary with scan parameters, especially pixel pitch.

SEM measurements align with AFM when pixel pitch is sufficiently small.

RMS roughness measurements are stable across different pixel pitches.

Abstract

We imaged the topography of a silicon grating with atomic force microscopy (AFM) using different scan parameters to probe the effect of pixel pitch on resolution. We found variations in the measured step height and profile of the grating depending on scan parameters, with measured step width decreasing from 1300 to 108 nm and step height increasing from 172 to 184 nm when a pixel pitch in the scan axis decreased from 625 nm to 3.91 nm. In order to resolve the measurement anomaly of step width, we compared these values with step width and height of the same grating measured using scanning electron microscopy (SEM). The values obtained from SEM imaging were 187.3 nm +/- 6.2 nm and 116 nm +/- 10.4 nm, which were in good agreement with AFM data using a 3.91 nm of pixel pitch. We think that we need at least four pixels over the step width to avoid the measurement anomaly induced by the stick-slip or dragging of the tip. Our findings that RMS roughness varied less than 1 nm and converged at the value of 77.6 nm for any pixel pitch suggest that the RMS roughness is relatively insensitive to the pixel pitch.