Round-robin experiments have been performed to correct binding energy shifts of XPS spectra for oxides by charge-up which results from irradiation of x-rays or ion sputtering and to determine criteria for the change of chemical state resulting from Ar ion sputtering. AI,O,, SO,, MgO, TiO, and NiO, in plate and powder form, were used for the experiments. We have revealed that the binding energy difference between the Al 2p (Si 2p or Mg 1s) line and the 0 1s line gives a better correction for the measurements of A1,0, (SiO, or MgO) than that between the A1 2p (Si 2p and Mg 1s) line and the C 1s line. After Ar ion sputtering, TiO, and NiO are reduced and new components appear in the XPS spectra. For AI,O, and SO,, the damage is induced by sputtering and the full width at half-maximum (FWHM) values of XPS spectra increase, while MgO is unchanged. We have shown that the changes of oxides caused by ion sputtering depend on the change of free formation enthalpy and iooicity. We have also shown that both plate and powder samples are useful as standards to determine the binding energy from XPS spectra. By using XPS analysis, species of elements at the surface are identified from XPS spectra and the chemical state of the elements is determined from the chemical shift of the binding energy of the electron concerned. The kinetic energy (E,) of the emitted photoelectrons is expressed as followswhere hv is the energy of the x-rays, EB is the binding energy of the electron concerned and is the work function. The accurate binding energy has to be measured in order to determine the chemical state from the chemical shift. For pure metals, it is possible to use the Fermi edge as a a reference for the binding energy. For semiconductors and insulators, however, it is impossible to use the Fermi edge as the reference because the Fermi level exists within the band gap.Moreover, the surface of the material charges up positively as a result of irradiation with x-rays, because