Comparison for modern microscopes at research laboratories (Nordic-nano2: Step height)

Project Description

MIKES metrology, the national metrology institute in Finland, and national metrology institutes in Sweden (SP Sveriges Tekniska Forskningsinstitut), Norway (Justervesenet) and Estonia (Metrosert) are supporting micro- and nanotechnology research and development in Nordic and Baltic.

We are organizing a comparison measurement for microscope users in Nordic and Baltic countries starting spring 2015. The samples in the comparison are two or three step height standards with nominal steps of h = 20 nm, h = 100 nm and h = 1 µm, and measurand is the average step height of the sample (ISO 5436). The samples are gratings with pitch of 3 nm or 10 µm. Also smaller or larger pitch is available. Participants can participate in the comparison with one or several different instruments suitable for this kind of measurements (AFM, profilometer, interference microscope...).

Reference value for the comparison will be provided by MIKES IT-MAFM. MIKES step height calibration service is internationally accepted. By comparing their own results and the reference value measured at MIKES participants will get information about accuracy of their instrument and reliability of their measurement method.

Final Report 2017-05-29

An intercomparison of step height measurements in research institutes in Northern Europe. Three step height samples were circulated among the participating laboratories. There were 12 participants in the comparison, 8 from Finland, 2 from Sweden, 1 from Norway and 1 from Estonia. 8 participants reported results. In total results were reported from 16 instruments: 3 optical profilometers, 4 stylus instruments, 8 AFMs and 1 SEM.
The step heights of the samples were nominally 20 nm, 100 nm and 1 µm. Maximum differences from reference values were 1.1 nm for 20 nm sample, 14 nm for 100 nm sample and 120 nm for 1 µm sample.
The agreement with the reference value was good for the smallest step; in a Birge ratio evaluation, only one value En > 1. Also for 100 nm sample the results were quite good; four En > 1. But for the 1 µm step eight out of sixteen En > 1. Quite surprisingly the largest deviations were in AFM measurements. The results show that pure scale calibration is not enough, but also the nonlinearity of the scale should be taken into account. There was no clear correlation between the results for different step heights.
More information about the traceability and instrument calibration is needed. Most partners were universities. In addition to research, the universities have crucial role in teaching new generations of researchers. Good calibration practices should be one of the thing students learn during their studies.
Most partners estimated their uncertainty only using statistical standard deviation. Systematical errors are more difficult to estimate, but as can be seen in the results systematic errors can be much larger than the standard deviation of the results. Systematic errors should include at least instrument calibration and errors caused by sample alignment.