Standard Article

Ellipsometry in Analysis of Surfaces and Thin Films

Surfaces

  1. Robert W. Collins

Published Online: 15 SEP 2006

DOI: 10.1002/9780470027318.a2507

Encyclopedia of Analytical Chemistry

Encyclopedia of Analytical Chemistry

How to Cite

Collins, R. W. 2006. Ellipsometry in Analysis of Surfaces and Thin Films. Encyclopedia of Analytical Chemistry. .

Author Information

  1. The Pennsylvania State University, USA

Publication History

  1. Published Online: 15 SEP 2006

Abstract

An ellipsometry measurement consists of five steps.

  1. A light beam is generated in a known polarization state.

  2. The beam is reflected from or transmitted through a sample having specularly reflecting plane-parallel surfaces, leading to a linear (frequency-conserving) interaction that changes the polarization state.

  3. The polarization state of the reflected or transmitted beam is measured.

  4. Parameters are determined that characterize the interaction in terms of the change in polarization state.

  5. From these parameters, information about the sample is deduced, including its optical properties, or the thickness and optical properties of a thin film on the surface of the sample.

Steps 1–3 involve set-up and operation of instrumentation whereas step 4, data reduction, is generally performed on-line by computer during measurement. Step 5 is usually performed after measurement, unless ellipsometry is being applied for real time process control.

Several variations of the ellipsometry experiment have been developed. The most common is reflection ellipsometry, in which the polarization state change is measured upon oblique reflection of the light beam from a specularly reflecting sample. For an isotropic sample, the parameters describing this interaction are (ψ, Δ), defined by tan ψ exp(iΔ) = rp/rs, where rp and rs are the complex amplitude reflection coefficients for linear p and s polarization states. In these states, the electric field vibrates parallel and perpendicular to the plane of incidence, respectively. In spectroscopic ellipsometry (SE), (ψ, Δ) are measured continuously versus wavelength over a spectral range designed for sensitivity to the physical and chemical features of interest. In real time ellipsometry (RTE), (ψ, Δ) are measured versus time at fixed wavelength using a time interval designed to follow the kinetic processes of interest. The latter two modes can be combined to yield real time spectroscopic ellipsometry (RTSE), utilizing parallel detection at many wavelengths simultaneously. In imaging ellipsometry (IE), (ψ, Δ) are measured over a two-dimensional area of a nonuniform sample surface using a resolution designed to detect feature sizes of interest.

Because the polarization change is determined in an ellipsometry measurement, rather than the irradiance change as in a reflectance measurement, ellipsometry is a very high precision technique for characterizing the optical properties of materials and thin films, as well as the thicknesses of thin films. The ps phase shift difference Δ provides exceptionally high sensitivity to changes in the thickness of a thin film on a specular surface. In fact, changes on the order of hundredths of a monolayer in surface coverage can be detected. A primary advantage of ellipsometry is its ability to obtain data noninvasively from samples immersed in any medium that is transparent to the light beam. Thus, numerous in situ and real time applications exist in disciplines such as physics and chemistry, materials and surface science, electrical and chemical engineering, electrochemistry, and biochemistry. In contrast, analytical techniques based on electron spectroscopies, for example, are only effective in high vacuum environments; thus, in vacuo instrumentation is required.

Limitations of ellipsometry arise from restrictions on the nature of the sample, as well as from the difficulties of data analysis, particularly for complex samples e.g. those exhibiting nonuniformities over the area of the probe beam. To avoid measurement difficulties as a result of light scattering and to simplify data analysis, reflecting surfaces should be specular and free of optical nonuniformities on an in-plane scale greater than about one-tenth of the wavelength of the incident beam. Nonuniformities on a scale greater than the wavelength can be detected by IE; in other measurement modes the sample should be uniform over the beam area for ease of data analysis. In addition, characterization of films that are thinner than the thickness of the roughness modulations on the substrate surface is difficult since data analysis relies on optical models consisting of multilayer stacks with plane-parallel interfaces. Recent progress has been made in ellipsometry research that addresses the analysis of complex samples.

The most widely used instruments for SE and RTSE span the spectral range from the ultraviolet (200–300 nm) to the near-infrared (800 nm). Over this spectral range, the optical properties deduced from SE provide information on the processes of absorption and dispersion originating from the valence electron transitions. Specialized SE equipment designed for the infrared range can be more effective for chemical identification of thin films, based on the absorption processes originating from the bond vibrational modes.