Detection of defect states in low-k dielectrics using reflection electron energy loss spectroscopy

Abstract

A reverse Monte Carlo method for deriving optical constants of solids from reflection electron energy-loss spectroscopy spectra Reflection electron energy loss spectroscopy (REELS) has been utilized to measure the band gap (E g ) and energy position of sub-gap defect states for both non-porous and porous low dielectric constant (low-k) materials. We find the surface band gap for non-porous k ¼ 2.8-3.3 a-SiOC:H dielectrics to be ffi 8.2 eV and consistent with that measured for a-SiO 2 (E g ¼ 8.8 eV). Ar þ sputtering of the non-porous low-k materials was found to create sub-gap defect states at % 5.0 and 7.2 eV within the band gap. Based on comparisons to observations of similar defect states in crystalline and amorphous SiO 2 , we attribute these sub-gap defect states to surface oxygen vacancy centers. REELS measurements on a porous low-k a-SiOC:H dielectric with k ¼ 2.3 showed a slightly smaller band gap (E g ¼ 7.8 eV) and a broad distribution of defects states ranging from 2 to 6 eV. These defect states are attributed to a combination of both oxygen vacancy defects created by the UV curing process and carbon residues left in the film by incomplete removal of the sacrificial porogen. Plasma etching and ashing of the porous low-k dielectric were observed to remove the broad defect states attributed to carbon residues, but the oxygen vacancy defects remained.

Electrical Reliability Challenges of Advanced Low-k Dielectrics

“…This bandgap range of low-k dielectrics is similar to the one of amorphous SiO 2 , which is between 8.0 and 8.8 eV. The reflection electron energy loss spectroscopy results from King et al, 10 the ellipsometric data from Marsik et al, 13 the vacuum ultraviolet (VUV) spectroscopy measurements from Zheng et al 14 and the X-ray photoelectron spectroscopy analysis from Nichols et al 15 support that the bandgap of most porous SiOCH type low-k dielectrics (k = 2.0-3.3) are in the range between 7.5 to 10 eV. In addition, the barrier height at both the low-k/metal and the low-k/Si interfaces can be measured with internal photoemission experiments.…”

“…Knock-off of atoms from the low-k material network are believed to happen during the ion sputtering process resulting in the formation of Si vacancies like EX centers or dangling carbon bonds, where the carbon related defects contribute to a higher leakage. Furthermore, the results of King et al 10 show the generation of surface oxygen vacancies, probably due to the removal of terminal organic groups after Ar + sputtering, where two related sub-gap surface states are observed at 5.0 eV and 7.2 eV. Nichols et al 41 studied the effect of the ion energy in the plasma.…”

“…24 In general, depositing thinner low-k dielectrics is more favorable as thin films are cured by the UV treatment over the whole depth of the film. The electron-spin resonance (ESR) results from Afanas'ev et al 24 show that signals associated with a carbon dangling bond in a cluster are observed for 500 nm thick low-k dielectrics but are absent for 60 nm thick low-k dielectrics, where King et al 10 pointed out that the defect band between 2-6 eV in the bandgap of the low-k dielectric is attributed to these carbon dangling bonds from porogen residues. Also, the UV and VUV irradiated results from Lauer et al 25 show that the defect centers originating from porogen residues are able to capture and exchange electrons.…”

“…9 As a result, for porous low-k dielectrics, there is no real bandgap. But as King et al 10 argued, the concept of a bandgap is extendable to low-k dielectrics and has been successfully used to explain a number of empirical measurement results. To characterize bandgap in porous SiOCH type low-k dielectrics, various techniques are being used.…”

Measurement of bandgap energies in low-k organosilicates

“…23 Given the above considerations, it is possible that the claims of <6 eV band-gap energies stem from a possible misapplication of an otherwise physically sound model, despite the fact that it can still provide accurate thickness results even with a relatively poor estimate of the bandgap. Additional verification of the results presented here, based on the onset of electron energy loss, comes from very recent work 24,25 using the method of reflected electron-energy-loss spectroscopy (REELS) in a Auger electron-spectroscopy system. Their work confirms a bandgap of approximately 8.0 6 0.4 eV for UV-cured SiCOH (irrespective of dielectric constant).…”