Thursday, December 3, 2015

Choosing the Best Method for Measuring Various Roughness Ranges

The NanoCam Sq dynamic surface profiler can measure surface roughness on smooth to super-smooth surfaces. To reliably measure nanometer-scale surface texture, the height resolution of the measurement method must be less than 1⁄10 of the Sq (the Root Mean Square Roughness) of the surface. Three techniques are suggested, depending on the range of roughness to be measured.

A. RMS roughness greater than ~1nm 

Averaging multiple measurements improves the height resolution; the more measurements that you average, the finer the roughness that can be measured.

The figure below shows that a single measurement can be used to measure RMS roughness greater than ~5nm. As the roughness decreases, increase the number of averages to maintain reliable measurement. For example, to measure RMS roughness of ~1nm an average of 10 measurements or more should be used.

Note that surfaces with very low reflectivity may require a greater number of averaged measurements.

Number of measurements to average based on RMS roughness of surface. 

B. RMS roughness is between ~ 0.2nm and 1nm

When the RMS roughness is below 1nm averaging alone may not provide sufficient resolution. In this range is it recommended to subtract a reference measurement from the data to remove the residual noise due to the system and optics.

One way to do this is to measure the residual system errors using a precision optic with a surface roughness better than the desired accuracy. The precision optic must be free of defects such as scratches and digs. 4D Technology can supply a calibration mirror with a surface roughness less than 0.1nm. The reference measurement must be created using the interference objective, reflectivity setting, and aperture stop setting and resolution that will be used to measure the actual test piece.

More information on reference subtraction can be found in the 4D Technology 4Sight Software User Manual.

C. RMS roughness is less than ~ 0.2nm 

To measure samples with an Sq less than 0.2nm it is necessary to eliminate all effects of the reference path. In this case the Absolute Sq technique can be used. Two averaged measurements are made of the sample at positions laterally separated by a distance greater than the sample surface correlation length. The difference between the measurements is calculated, and the effects of the reference path are thereby removed from the resulting “difference.” See the figure above to determine the required number of averaged measurements.

The Sq of the sample is then estimated as:

Sq = [Difference]/√2 

This technique does not produce a measurement map as in methods A and B; instead, the Sq of the sample surface is reported as a single number. The two averaged measurements should be acquired as rapidly as possible to prevent environmental changes from influencing the results.

More information on surface roughness measurement can be found in the ASME B46.1 2009 standard on "Surface Texture (Surface Roughness, Waviness, and Lay)."