Surface metrology

Surface metrology is the measurement of small-scale features on surfaces, and is a branch of metrology. Surface primary form, surface fractality and surface roughness are the parameters most commonly associated with the field. It is important to many disciplines and is mostly known for the machining of precision parts and assemblies which contain mating surfaces or which must operate with high internal pressures.The effect of different form removal techniques on surface finish analysisPlots showing how filter cutoff frequency affects the separation between waviness and roughnessIllustration showing how the raw profile from a surface finish trace is decomposed into a primary profile, form, waviness and roughnessIllustration showing the effect of using different filters to separate a surface finish trace into waviness and roughness Surface metrology is the measurement of small-scale features on surfaces, and is a branch of metrology. Surface primary form, surface fractality and surface roughness are the parameters most commonly associated with the field. It is important to many disciplines and is mostly known for the machining of precision parts and assemblies which contain mating surfaces or which must operate with high internal pressures. Surface finish may be measured in two ways: contact and non-contact methods. Contact methods involve dragging a measurement stylus across the surface; these instruments are called profilometers. Non-contact methods include: interferometry, digital holography, confocal microscopy, focus variation, structured light, electrical capacitance, electron microscopy, and photogrammetry. The most common method is to use a diamond stylus profilometer. The stylus is run perpendicular to the lay of the surface. The probe usually traces along a straight line on a flat surface or in a circular arc around a cylindrical surface. The length of the path that it traces is called the measurement length. The wavelength of the lowest frequency filter that will be used to analyze the data is usually defined as the sampling length. Most standards recommend that the measurement length should be at least seven times longer than the sampling length, and according to the Nyquist–Shannon sampling theorem it should be at least two times longer than the wavelength of interesting features. The assessment length or evaluation length is the length of data that will be used for analysis. Commonly one sampling length is discarded from each end of the measurement length. 3D measurements can be made with a profilometer by scanning over a 2D area on the surface. The disadvantage of a profilometer is that it is not accurate when the size of the features of the surface are close to the same size as the stylus. Another disadvantage is that profilometers have difficulty detecting flaws of the same general size as the roughness of the surface. There are also limitations for non-contact instruments. For example, instruments that rely on optical interference cannot resolve features that are less than some fraction of the operating wavelength. This limitation can make it difficult to accurately measure roughness even on common objects, since the interesting features may be well below the wavelength of light. The wavelength of red light is about 650 nm, while the average roughness, (Ra) of a ground shaft might be 200 nm. The first step of analysis is to filter the raw data to remove very high frequency data (called 'micro-roughness') since it can often be attributed to vibrations or debris on the surface. Filtering out the micro-roughness at a given cut-off threshold also allows to bring closer the roughness assessment made using profilometers having different stylus ball radius e.g. 2 µm and 5 µm radii. Next, the data is separated into roughness, waviness and form. This can be accomplished using reference lines, envelope methods, digital filters, fractals or other techniques. Finally, the data is summarized using one or more roughness parameters, or a graph. In the past, surface finish was usually analyzed by hand. The roughness trace would be plotted on graph paper, and an experienced machinist decided what data to ignore and where to place the mean line. Today, the measured data is stored on a computer, and analyzed using methods from signal analysis and statistics.