Frequently Asked Questions

A: Yes. This can be accomplished using reflectometry. We have found that for this particular situation, a reflectance measurement covering the infra-red wavelength range (1000nm – 15,000nm) will have the necessary sensitivity. One caveat: the optics of the reflectometer needs to be such its noise level is not intertwined with percent reflectance due to surface roughness. That is this can only be done with an IR reflectometer that has an extremely high signal to noise ratio, such as the n&k Olympian.

A: Yes. If the measured UV-Vis reflectance spectrum of the film on Si-substrate surpasses the UV-Vis reflectance spectrum of bare-Si at any wavelength in the measured range, the film is inhomogeneous. All n&k tools n&k Olympian, n&k OptiPrime-CD, n&k LittleFoot-CD and TF Series, have the capability to determine inhomogeneity of a film. In which case, n&k tools will then report the n and k spectra at the top of the film, the bottom of the film and the average n and k spectra.

A: As can be demonstrated when using the n&k Olympian, a UV-Vis-IR scatterometer, it is necessary to extend the measurements to the infra-red wavelength range when initial measurements using DUV-Vis-NIR wavelengths does not reveal sufficient information to determine depths, CDs and profiles of 2D and 3D structures. The extension to the infrared spectral range is necessary in order to measure complex ultra-deep structures, or structures with high aspect ratios.

A: Yes, if accuracy and repeatability are important for such measurements. As can be demonstrated by the Scatterometer / Thin Film metrology tool, the n&k Gemini, very little reflected light reaches the detector for cases where the substrate is transparent – most of it is transmitted. Thus, by including measured transmittance, a better Signal-to-Noise (S/N), compared to reflectance, is obtained, since more transmitted photons reach the detector than reflected photons, and the (S/N) ratio is proportional to # of detected photons. Furthermore, the noise in the reflectance spectrum can overshadow variations in reflectance due to CD variations, whereas the high (S/N) of transmittance measurements means CD variations can be better resolved. Same arguments apply to variations in side wall angle.

A: Scatterometry depends on Rigorous Coupled Wave Analysis (RCWA) of the measured
raw data to determine depths, CDs and Profiles of 2D (trenches) and 2D (contact holes) structures. In turn, RCWA depends on the n and k spectra of the films comprising structure that is being measured. Thus, if the n and k values are incorrect, then the RCWA will yield incorrect results for Depths, CDs and Profiles. Scatterometers that combine physically valid dispersion equations for n and k with RCWA, e.g., the n&k Olympian or n&k OptiPrime will provide the most accurate measurement results.

A: The importance of “All-Reflective Optics” is that there are no beam-splitters, refractive lenses, or other light absorbing components contained within the optics. Such light absorbing components reduce the amount of light that is sensed by the detector, thus reducing the signal-to-noise, which in turn affects the accuracy and repeatability of the measurement. Basically, with an “All-Reflective” optical design, the signal-to-noise ratio is optimized, which in turn produces accurate results and excellent repeatability.

A: Yes. Although Epi-Si is opaque in the UV-Vis-NIR wavelength range, Epi-Si is transparent in the infrared wavelength regime. In fact, the n&k Olympian, with its wavelength range covering UV-Vis-IR wavelengths, is successfully used for measuring thickness and n and k spectra of Epi-Si.

A: Yes. The n&k 450-M, a manual load UV-Vis-NIR Scatterometer/Thin Film metrology tool, is available today for the larger 450mm wafers. The 450-M will provide the necessary OCD and Thin Film metrology for development of new technology for the 450mm transition in the most cost effective way.

A: Yes. n&k’s scatterometer, the n&k OptiPrime, is able to decouple FinFET parameters due to its innovative optics that produces the optimized signal-to-noise ratio of order 2%-5% per 2nm change in the parameters. The OptiPrime’s wide spectral range helps, because, in general, if there are changes in the measured raw data at all wavelengths, then the effect of coupling is not too significant.