Removal of inelastic scattering part from Ti2p XPS spectrum of TiO2 by deconvolution method using O1s as response function

Masaoki Oku, Hideyuki Matsuta, Kazuaki Wagatsuma, Yoshio Waseda, Shigemi Kohiki

Research output: Contribution to journalArticlepeer-review

42 Citations (Scopus)

Abstract

Ti2p and O1s XPS spectra of a clean surface of single crystal TiO2 fractured in situ were taken to study the removal of the inelastic scattering part from the Ti2p spectrum. The features of inelastic scattering peaks in EELS were more influenced by surface than O1s XPS. It indicates that O1s XPS is proper as the response function to deconvolute a Ti2p spectrum. FWHM of the non-energy loss peak of Ti2p3/2 is smaller than that of O1s. When the raw O1s spectrum is used as the response function, the deconvoluted spectrum has negative intensity values in some regions and beat waves. The replacement of the non-energy loss peak of O1s with narrower Gaussian peak takes off the beat waves. Peak separation of the deconvoluted spectrum reveals that it has six peaks. Two of them are main peaks of Ti2p3/2 and 2p1/2. The others are classified into two kinds of satellite peaks, whose energy separation from the main peaks are 3 and 13 eV. Although the latter satellites have been discussed by many authors, the former satellites are first reported here.

Original languageEnglish
Pages (from-to)211-218
Number of pages8
JournalJournal of Electron Spectroscopy and Related Phenomena
Volume105
Issue number2-3
DOIs
Publication statusPublished - 1999 Dec

Keywords

  • Deconvolution
  • Electron energy loss spectrum
  • In situ fractured surface
  • O1s XPS spectrum
  • Ti2p XPS spectrum
  • TiO

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Radiation
  • Atomic and Molecular Physics, and Optics
  • Condensed Matter Physics
  • Spectroscopy
  • Physical and Theoretical Chemistry

Fingerprint

Dive into the research topics of 'Removal of inelastic scattering part from Ti2p XPS spectrum of TiO<sub>2</sub> by deconvolution method using O1s as response function'. Together they form a unique fingerprint.

Cite this