Physical structures of SiO2 ultrathin films probed by grazing incidence x-ray reflectivity

Abstract

The effects of contamination overlayer and density as well as surface and interface roughnesses on the x-ray reflectivity of a SiO2 ultrathin film are discussed from viewpoints of experiment and theory. Grazing incidence x-ray reflectivity (GIXRR) is used to accurately measure physical structures of SiO2 ultrathin films grown on Si substrate by effectively resolving deviations caused by a contamination overlayer (i.e., H2O and carbonaceous compounds). GIXRR results indicate that only the thickness accuracy of a SiO2 film is largely affected by the overlayer. The thickness of a SiO2 film obtained from GIXRR peak extrema and theoretical fitting reveals that if a SiO2 film with the thickness of 2.64nm is considered as a single layer, the H2O overlayer with a thickness of 0.55nm makes the thickness of the SiO2 film increase to 3.09nm, and the deviation is about 17% from its corrected thickness. By evaluating the GIXRR results of three repeating measurements of a nominal 4-nm SiO2 film, its density, thickness, and surface and interface roughnesses are 2.43±0.01g∕cm3, 3.99±0.03nm, and 0.40±0.02nm and 0.25±0.02nm, respectively.

Ultra‐thin SiO2 on Si IX: absolute measurements of the amount of silicon oxide as a thickness of SiO2 on Si

“…An issue that also needs discussion is the zero origin of the relation for d. For clean Si, d should be zero but this will not be observed with zero uncertainty as a result of statistical noise in the background where the peak should be. That U E is 0.025 nm is confirmed for d > 0.2 nm in Refs [20,21] and for the range down to 0.022 nm in Ref. [26].…”

“…Measurements [17 -19] indicate that the physisorbed water thickness could be between 0.05 and 0.3 nm. Azuma et al [20] show, using XRR, by fitting with a two-layer model of water and oxide instead of a one-layer model of the oxide that the full extent of the water layer in air is 0.55-0.67 nm for thermal oxides of similar thicknesses to those measured here. Furthermore, they show using thermal desorption spectrometry that the water requires 800…”

Influence of contamination layer on thickness evaluation by X‐ray reflectometry

“…In general, XRR using Cu Kα is not sensitive to the composition of organic molecules consisting of light elements (e.g., H, C, N, and O). For this reason, similar calculated profiles were obtained even when the contamination layer was assumed to comprise H 2 O molecules . The densities of the two SiO 2 layers are also different in the model used later in the analysis process (as shown in Table , profiles (c) and (d)).…”

“…Surface changes caused by surface contamination and a growing surface oxide layer may be examined using XRR experiments and analysis. Attention is generally not paid to quantifying the properties of the contamination layer when conducting thickness determination, although there have been certain reports on layer thickness determination using XRR that considered the presence of a contamination layer …”

Nonlinearities in depth profiling nanometer layers

“…The material was of ͑100͒ and ͑111͒ wafers and, as a result of an extended study, the oxide thicknesses could be established with an accuracy better than 0.1 nm using XPS, with all of the wafers carefully interrelated using ellipsometry with a precision of ϳ0.002 nm. 36,37,39 The samples were sputter depth profiled in a Physical Electronics Adept 1010 quadrupole mass-spectrometer-based SIMS system using 600 eV Cs + primary ions set at 60°to the surface normal. Careful cleaning and storage ensured that the carbonaceous contamination was very low, 38 being typically 0.2 nm, but in the ambient environment, water could adsorb reversibly at the outermost surface.…”