Invited Paper AS-MoM1
Characterization of Nano-structures from Analysis of the XPS Background: Automation and 3D-imaging

Monday, October 31, 2011, 8:20 am, Room 102

The XPS peak intensity and its background of inelastically scattered electrons varies strongly with the depth distribution of atoms. This phenomenon is the basis for a widely used method for nano-structure analysis.¹ With the rapid increase in the application of XPS, there is a growing need for automated XPS analysis. To meet this demand, a modified simpler and less accurate algorithm, which however is suitable for automation was developed.² For each XPS peak, this algorithm determines just two numbers: the total amount of atoms within the outermost ~ 10 nm and the approximate depth of the atoms. The validity of the algorithm was demonstrated experimentally by comparison to more elaborate quantification methods.² The algorithm is thus of interest for automated XPS analysis. Another application of the algorithm is in XPS imaging, where thousands of spectra must be analyzed. Here automatic data-handling procedures are crucial. Software that can automatically analyze thousands of spectra corresponding to the situation in XPS imaging is now being developed. The method produces nondestructively a 3-D image of the surface with nanometer depth resolution. The practical applicability of this to XPS imaging was recently demonstrated.^3-4

In ref ³ the algorithms ability to produce images of Ag taken from a series of samples with increasing thicknesses of plasma patterned Octadiene (2, 4, 6 and 8nm) on Ag substrates was demonstrated. The obtained images of the amount of Ag atoms in the outermost few nano-meters were in good agreement with the nominal thicknesses. Produced images of atoms at different depths clearly proved the potentials of the method for quantitative and nondestructive 3-D characterization of nano-structures. Spectral noise is a major limitation in imaging because the time allowed to acquire the spectrum at each pixel is typically ~1 sec. Different procedures for noise reduction were studied in ref⁴ and principal component analysis (PCA) was found to significantly improve the 3D images of a thermally patterned oxidized silicon where detailed 3-D images of the Si, O, and C atoms were determined on the nano-meter depth scale.

In the talk we will summarize the technique and discuss its limitations and potentials for both automation of conventional XPS data analysis and for 3D XPS imaging.

¹ S. Tougaard, Surf. Interf. Anal. 26 (1998) 249

² S. Tougaard, J. Vac. Sci. Technol., A21 ( 2003) 1081; A23(2005) 741

³ S. Hajati, S. Coultas, C. Blomfield and S. Tougaard, Surf. Interf. Anal. 40 (2008) 688