Multilayer approach to the quantitative analysis of x-ray photoelectron spectroscopy results: Applications to ultrathin SiO2 on Si and to self-assembled monolayers on gold

Abstract

X-ray photoelectron spectroscopy (XPS) is widely applied for the chemical characterization of surfaces and multilayers of thin films. In order to obtain quantitative results, XPS peak areas generally are divided by sensitivity factors and normalized to 100at.% to obtain so-called raw concentrations. For homogeneous materials, materials with randomly distributed atoms within the analyzed surface layer, these concentrations may be a useful quantity. Yet, for a material consisting of a substrate on top of which a number of chemically different layers are present, the raw concentrations depend on measuring details like the takeoff angle during the XPS analyses and clearly are not a satisfactory way to describe the sample. The main purpose of this article is to present a calculation method that converts raw concentrations into more meaningful quantities. The method is applicable to a restricted but technologically relevant class of samples: substrates on top of which one or more homogeneous layers are present. Examples are: gate dielectrics on Si or GaAs, self-assembling monolayers on a metallic substrate, thin oxide films on metals with an organic contamination on top. The method is based upon standard exponential attenuation of the photoelectron intensity as a function of traveled distance. For each element or chemical state in the system it has to be known to which layer(s) it belongs. Sensitivity factors are corrected for matrix effects and for intrinsic excitations. Starting from the raw concentrations, the method calculates in a self-consistent way the composition of all layers in the system and the thickness of each layer. Only one measurement at one measuring angle is required to obtain these results. To obtain insight into the accuracy of the calculation method, XPS results obtained on ultrathin SiO2 layers on Si that were slightly contaminated with hydrocarbons have been analyzed with the method. The obtained thicknesses were in good agreement with values for the thickness of the SiO2 layer and the organic surface contamination as obtained by other methods. Consistent values were also obtained for the concentration ratio O∕Si in the SiO2 layers. The calculation method has also been verified for three types of self-assembled monolayers (SAM layers) on gold. Layers of C18 (octadecane-thiol) and of EG4 (a mercaptoalkyloligo-ethyleneglycol) deposited from solutions with different concentrations were examined. Also, SAM layers deposited from mixtures with molecules with different chain lengths, mercapto-undecanol (MUO), and a biotinylated oligo-ethyleneglycol-alkyl thiol (BAT), were investigated. The model analysis provided the thickness of the organic layers, the concentrations of the components in the layers, and the coverage of the gold with sulphur (in atoms∕cm2). Rutherford backscattering spectrometry (RBS) was applied to determine (in an independent way) the amount of sulphur at the gold surface. The RBS results correlated well with the XPS data. The obtained values for the concentration ratios of the SAM layers were in agreement with the theoretically expected values. It is shown in the article that it is essential to model the mixtures of MUO and BAT as a three-layer system (gold substrate, aliphatic interlayer, and top layer containing the ethylene oxide groups) in order to obtain agreement.