< back to main site


Ultra thin SiO2 on Si IX: Absolute Measurements of the Amount of silicon oxide as a thickness of SiO2 on Si.

Seah, M P; Unger, W E S*; Hai Wang*,, ; Jordaan, W*; Gross, Th*; Dura, J A*; Dae Won Moon*,, ; Totarong, P*; Krumrey, M*; Hauert, R*; Mo Zhiqiang*, (2009) Ultra thin SiO2 on Si IX: Absolute Measurements of the Amount of silicon oxide as a thickness of SiO2 on Si. Surf. Interface Anal., 41 (5). pp. 430-439.

Full text not available from this repository.


Results are reported for a study between National Metrology Institutes of measurements of the absolute thicknesses of ultra-thin layers of SiO2 on Si. These results are from a key comparison (K32) and associated pilot study (P84) under the auspices of the Consultative Committee for Amount of Substance (CCQM) to measure the amount of silicon oxide on silicon wafers expressed as a thickness of SiO2. "Amount of substance" may be expressed in many ways and here the measurand is the thickness of the silicon oxide layers of nominal thicknesses in the range 1.5 to 8 nm on Si substrates, expressed as the thickness of SiO2. Separate samples were provided to each institute in containers that limited the carbonaceous contamination to below about 0.3 nm. The samples were of ultra-thin SiO2 on (100) and (111) orientated wafers of Si. The measurements from the participating laboratories were conducted using ellipsometry, neutron reflectivity (NR), X-ray photoelectron spectroscopy (XPS) or X-ray reflectivity (XRR) measurements, guided by the protocol developed in the earlier pilot study, P38. A very small correction is made for the different samples that each laboratory received. Where appropriate, method offset values attributed to the effects of contaminations from the pilot study, P38, were subtracted. Values for the Key Comparison Reference Values (KCRVs) and their associated uncertainties for these samples are then made from the weighted means and the expanded weighted standard deviations of the means of these data. These results show a dramatic improvement on previous comparisons, leading to 95% uncertainties in the range 0.09 to 0.27 nm, equivalent to 0.4 to 1.0 monolayers over the 1.5 to 8.0 nm nominal thickness range studied. If the sample-to-sample uncertainty is reduced from its maximum estimate to the most likely value, these uncertainties reduce to 0.05 to 0.25 nm or relative standard uncertainties ~1.4%. The best results achieve ~1% relative standard uncertainties. It is concluded that XPS has now been made fully traceable to the SI, for ultra-thin thermal SiO2 on Si layers, by calibration using wavelength methods in an approach that may be extended to other materials systems.

Item Type: Article
Keywords: attenuation length, calibration, silicon dioxide, thickness, ultra-thin oxide
Subjects: Nanoscience
Last Modified: 02 Feb 2018 13:15
URI: http://eprintspublications.npl.co.uk/id/eprint/4360

Actions (login required)

View Item View Item