Measurement of areal roughness parameters

Project Description

The measurement of roughness parameters of technical surface in industry and research institutes is mainly related to 2D-profiles as described in ISO 4287, ISO 5436, etc. However in ISO 25178 a large number of functional surface parameters described are based on areal measurements.The conditions for measurement and evaluation are more flexible to allow the adaption to a larger variety of instruments to be used and types of surfaces to be evaluated. There are already some national ( e.g. Optassyst) as well as European projects (SURFSTAND; AUTOSURF, …) which have investigated the measurement of areal parameters. However this has not been done at NMI level to achieve traceability and comparability. The calibration standards suggested in section 8 reflect the necessity, to use standards for roughness representation as well standards to find out metrological characteristics of the instruments, that may influence the measurement and evaluation rules that must be agreed. Therefore we propose to use in a first step a set of available standards to set-up and to fix rules how to measure simple areal surface parameters and to achieve comparable values to 2D profile roughness data.

The methods and rules investigated successfully should be a good base to perform traceable measurements of areal roughness parameters and transferred to calibration rules to be developed in ISO/TC213.

Potential calibration standards: Optical flat for noise verification, 2d-gratings for lateral scaling and resolution test, depth standard for vertical verification, good natured profile roughness standard (SFRN 300 nm), fine ground roughness standard (Halle, 1,5 µm), areal roughness standard (Simetrics).

Progress Report 2023-09-28

Roughness measurements were performed on four samples. Three samples, which demand a rather high spatial resolution but a rather small FOV, are two lapped Si chips (ARS-F1 and ARS-F2) from Simetrics and one ultrafine roughness standard (UFRS) produced by focused ion beam (FIB) from point electronic. The fourth sample (B40-VP04) from Rubert & Co. in England with very longwave-features, thus a much larger measurement field is required to reliably determine the roughness.

Furthermore, measurements were performed on a resolution standard of type RS-N from Simetrics, which has rectangular profiles with the height (240 nm) but different grating periods in the range of 300 nm to 6 µm, to get some rough information about the instrument transfer function.

The comparison revealed that especially the finer roughness standards with comparatively short wavelength roughness cause unexpectedly great difficulties in optical roughness measurement. The values reported by the individual institutes, e.g. for the roughness Sa, deviated from the reference values by up to 60 %, and the uncertainties reported by experienced metrologists proved to be too small. In addition, the measurement results show a large dependence on the selected measurement principle and the objectives used. The preliminary analysis shows that the spatial frequency spectrum, the flank angles of the surface texture have a significantly greater influence on the roughness results than previously assumed.

The comparison also underlines that if only rectangular gratings with different periods are used, the transfer behavior of an instrument for different slopes cannot be characterized. Therefore, sinusoidal structures (see the short report on determining topographic spatial resolution) must be used in future. It is suspected that the artifacts obtained by optical measurements lead to strong deviations of the measured roughness characteristics. However, more experimental investigations are needed.

The challenge for the very long wavelength standard (B40-VP04) was to validate stitching methods implemented in the instruments. Here, the selected stitching algorithm, leveling strategy, and preprocessing methods chosen have influence on the poor consistency of the roughness parameters determined by the participants. However, a re-evaluation of the raw data provided by the participants with an evaluation software developed by PTB showed that the measurement results of the participants agreed very well.

A follow-up project “TracOptic” has been therefore applied for within the framework of EMPIR call 2020 “Metrology for Industry” and was funded to investigate the influence factors arising from the measuring principles, hardware setup and feature properties by closely linking experimental investigations and numerical modelling together with other research partners. The overall objective of the project is to enable traceable 3D roughness and dimensional measurements using optical 3D microscopy and optical distance sensors, with a special focus on the development of practical instructions (Good Practice Guides) for the selection of most suitable measuring devices and their settings for the respective measurement purpose.
The results from VNIIMS are temporarily excluded. Guided by EURAMET’s standpoint, the participants have decided to conclude the comparison without VNIIMS’s contribution unless there is a change with respect to Russia’s invasion of Ukraine. VNIIMS data remain confidential to the pilot and may be used to amend the report when the political situation allows.

Draft B has been sent to TC-L on July 12, 2023.

Progress Report 2018-10-31

The progress report can be downloaded here>>.