XPS REFERENCE PROCEDURE FOR THE ACCURATE INTENSITY CALIBRATION OF ELECTRON SPECTROMETERS - RESULTS OF A BCR INTERCOMPARISON CO-SPONSORED BY THE VAMAS SCA TWA

A calibration procedure has been developed for XPS instrument intensity scales. The calibration is achieved by measuring the spectrum for reference foils of Ag, Au and Cu for which the true spectrum has been measured usng the metrology spectrometer. The ratio of the measured and reference spectra at 1 eV intervals from 200 eV kinetic energy to 1600 eV provides the calibration known as Q(E). The calibrations from all reference foils for a given instrument and setting should be the same and may be averaged to enhance accuracy. An interlaboratory study involving 58 different instruments, covering 25 different models from 9 manufacturers, has been completed with over 1200 individual spectra measured. The results show that the calibrations are excellent but that care is required or systematic errors may arise from instrumental problems such as x-ray anode contamination, sample contamination, contributions from the sample holder and internal scattering in the spectrometer. However, these problems may be readily diagnosed. The Cu data often show carbon contamination and so those data are omitted from the assessments of Q(E). The instruments provide unmonochromated x-rays from Al and Mg anodes as well as monochromated Al sources. The Q(E) curves for the unmonochromated sources always agree with each other whereas that for the monochromated source may agree with those for the unmonochromated sources in some instruments but not in others. The reason for the differences may lie in stray magnetic fields or lens magnification effects. The different reference materials give the same Q(E) on any given instrument for a given set of operating conditions with a typical scatter of 3% but actually ranging from 0.99% in the best case to 19.49% where problems, as noted above, occurred. Changing the slits or pass energy can change the relative intensities at low and high energy by more than a factor of 2. However, even larger differences may occur between instruments of the same model operated under identical settings in different laboratories. It is thus shown that each individual instrument needs a separate and regular calibration and that such a procedure would reduce the present intensity variations of a factor of 7 over the energy range 200-1400 eV to an average scatter of 3% or less. The individual calibrations for Q(E) for all of the 58 instruments are presented and it is shown that the methodology and reference data work effectively and consistently in all known situations where the reference procedure is adequately conducted.