The jump-into-contact effect in biased AFM probes on dielectric films and its application to quantify the dielectric permittivity of thin layers.

The jump-into-contact (JIC) phenomenon in biased atomic force microscopy (AFM) probes on dielectric films is studied. The influence of the film thickness on the position at which the AFM tip collapses irreversibly into the sample surface was theoretically analyzed using a widely accepted analytical expression of the probe-sample electrostatic interaction force. It was demonstrated that for relatively high values of voltage (V > 10-20 V) applied between the probe and the substrate the cantilever deflection at the JIC is independent of the dielectric film thickness for thin-ultrathin layers (h < 10-50 nm). Under the same conditions the z-piezo distance at the JIC follows approximately a linear behavior with the film thickness. Based on this effect an empirical model was formulated to estimate the dielectric permittivity of thin/ultrathin dielectric films using the jump-into-contact distance. The procedure was successfully applied on thin PVD-SiO2 films, obtaining good agreement with a dielectric constant value previously reported for the same material.

Autocorrelation Analysis Combined with a Wavelet Transform Method to Detect and Remove Cosmic Rays in a Single Raman Spectrum.

Cosmic rays (CRs) occasionally affect charge-coupled device (CCD) detectors, introducing large spikes with very narrow bandwidth in the spectrum. These CR features can distort the chemical information expressed by the spectra. Consequently, we propose here an algorithm to identify and remove significant spikes in a single Raman spectrum. An autocorrelation analysis is first carried out to accentuate the CRs feature as outliers. Subsequently, with an adequate selection of the threshold, a discrete wavelet transform filter is used to identify CR spikes. Identified data points are then replaced by interpolated values using the weighted-average interpolation technique. This approach only modifies the data in a close vicinity of the CRs. Additionally, robust wavelet transform parameters are proposed (a desirable property for automation) after optimizing them with the application of the method in a great number of spectra. However, this algorithm, as well as all the single-spectrum analysis procedures, is limited to the cases in which CRs have much narrower bandwidth than the Raman bands. This might not be the case when low-resolution Raman instruments are used.