Separation of bulk and surface-losses in low-loss EELS measurements in STEM

K. A. Mkhoyan, T. Babinec, S. E. Maccagnano, E. J. Kirkland, J. Silcox

Research output: Contribution to journalArticlepeer-review

39 Scopus citations


To identify major features in low electron energy loss spectra, the different excitations (bulk plasmons, interband transitions, surface plasmons, Cherenkov and surface guided modes) must be delineated from each other. In this paper, this process is achieved by noting the linear thickness dependence of bulk processes contrasted with the constant thickness behavior of surface excitations. An alternative approach of analyzing bulk plasmon-loss is also introduced. Using a new algorithm, the parameters of plasma generation-plasmon energy EP, 0, a damping parameter Δ EP and the coefficient of the dispersion relation γ were obtained from a single curve fitting on the example of Si. The ability to separate surface-losses from the rest of the data permitted identification of the fine structure of the surface-losses. The strong peak at 8.2 eV characteristic of non-radiative surface plasmon excitations was measured for Si. Analysis of surface excitations indicates that a 10 ÅSiO2 surface coating layer is still present despite careful cleaning the specimen. Dielectric functions deduced from the EELS data prove to be considerably affected by the presence of the surface-losses for samples as thick as 800 Å.

Original languageEnglish (US)
Pages (from-to)345-355
Number of pages11
Issue number4-5
StatePublished - Apr 2007

Bibliographical note

Funding Information:
This work is supported primarily by the Nanoscale Science and Engineering Initiative of the NSF EEC-0117770 and NYSTAR C020071. The sample preparation facilities and STEM are supported by NSF through the CCMR DMR 9632275. We would also like to acknowledge M. Thomas for technical support.


  • Bulk plasmon-loss
  • Dielectric function
  • EELS
  • STEM
  • Si
  • Surface-loss


Dive into the research topics of 'Separation of bulk and surface-losses in low-loss EELS measurements in STEM'. Together they form a unique fingerprint.

Cite this