Nanoscale Schottky barrier mapping of thermally evaporated and sputter deposited W/Si(001) diodes using ballistic electron emission microscopy

Abstract

Ballistic electron emission microscopy has been utilized to demonstrate differences in the interface electrostatics of tungsten-Si(001) Schottky diodes fabricated using two different deposition techniques: thermal evaporation using electron-beam heating and magnetron sputtering. A difference of 70 meV in the Schottky barrier heights is measured between the two techniques for both p- and n-type silicon even though the sum of n- and p-type Schottky barrier heights agrees with the band gap of silicon. Spatially resolved nanoscale maps of the Schottky barrier heights are uniform for the sputter film and are highly disordered for the e-beam film. Histograms of the barrier heights show a symmetric Gaussian like profile for the sputter film and a skewed lognormal distribution for e-beam film. A Monte-Carlo model is developed to simulate these histograms which give strong indication that localized elastic scattering is causing this skewing as forces the hot electrons to need a greater total energy to surmount the barrier. These differences are attributed to silicide formation from the unintentional substrate heating during the e-beam deposition, which is confirmed with transmission electron microscopy.

Detection of silicide formation in nanoscale visualization of interface electrostatics

“…Visualizing the nanoscale fluctuations in the electrostatic barrier at the interface between a metal and a semiconductor is possible with Schottky barrier mapping by ballistic electron emission microscopy (BEEM). [11][12][13] BEEM is an STM based technique, where the STM tip is utilized as a local emitter of hot electrons that tunnel into the metal, and the fraction of electrons that make it over the Schottky barrier and into the semiconductor is counted as the BEEM current. BEEM measures the threshold voltage of the unbiased diode, making it the most accurate measure of the barrier height.…”

Nanoscale Schottky barrier visualization utilizing computational modeling and ballistic electron emission microscopy

“…5.2. Typically, changes in barrier heights due to silicide formation are on the order of tens of meV as has been observed when comparing annealed and non annealed W/Si measurements [41,44,103,43].…”

“…5.2. This causes the Schottky barrier from the fit to the average spectra to be less than the mean threshold from the distribution as averaging of multiple spectra with different onsets causes the average spectra to deviate from the linearity near the threshold [44]. However, the fit to the averaged spectra gives the best agreement to the band gap of the substrate and is used as the measured Schottky barrier height [91,44].…”

“…The dramatic difference between the e-beam and sputter deposited W/Si samples arises from unintentional radiative heating of the sample during e-beam deposition, which does not occur during sputter deposition [44]. The heating promotes silicide formation which has been shown to shift the tungsten-silicon barrier height by roughly 70 meV [86,87].…”

“…Fig. 4.1 showed TEM and EDX profiling of a 10 nm wide cross section of the e-beam sample, showing a significantly intermixed tungsten-silicon region when compared to the sputter sample, which showed an abrupt tungsten to silicon transition [42,44]. The presence of a small amount of gold in the model suggests that either gold is in contact with the silicon or that a sub-nm-thick tungsten layer might not be able to "screen" the gold from affecting the barrier height in some regions.…”